Pension Research Council Papers on Decumulation and Robust DC Pension Plans
The Boettner Center for Pensions and Retirement Research at UPen’s Wharton provides access to readers interested in pension to current research papers from high calibre sources. They have just posted 15 new research papers. I found a couple of these papers particularly interesting, one focused at decumulation strategies in retirement and another on how to strengthen DC pension plans by using a feedback mechanism which both informs and triggers action for DC plan participants.
In “Comparing Spending Approaches in Retirement” Ameriks, Hess and Ren compare decumulation strategies- specifically 9 income generating methods: (1) endowments style systematic payout strategy (annual 5% of previous 12 quarterly NAV of fund), (2) time horizon systematic payout funds (start at 5% exhaust all funds over 30 years), (3) 5% inflation adjusted from a balanced mutual fund, (4) Variable immediate annuity (65 year old male with AIR=3.5%), (5) Variable annuity with Guaranteed Lifetime Income Benefit (65 year old male 5% guaranteed minimum income base with potential step-up and residual value can be withdrawn any time or part of estate , essentially GMWB described before), (6) Required Minimum Distribution (RMD) withdrawal plan, (8) combination strategies, (7) 50% real fixed annuity and 50% RMD, (8) 50% Variable Annuity 3.5% AIR and 50% RMD. (9) 65 year old male real fixed annuity. See paper for assumption on asset allocations, cost, expected returns and standard deviations, etc. The only problem with the analysis is that the assumption on fees are overly generous to variable annuities and GLWBs because while buying an insurance product to protect the downside might have some merit, but the fees associated with the underlying funds are much higher than those available outside of the context of the insurance product. Also, it would have been interesting to see the performance of the combination of an endowment style withdrawal payout strategy on 90% of the funds and a longevity insurance product (joint delayed payout annuity of payments starting at 85) for the remaining 10% of the funds. Actually GMWBs as a product (i.e. buying downside protection for 1%) might even make sense to many, so long that you could buy them together with 0.25% MER index funds, but certainly would not be recommended with a 2.5% mutual fund see my GMWB II blog (but then the 2.5% MER mutual fund doesn’t make sense with or without GMWB).
One of the problems of DC plans (as opposed to DB plans) is that the employer does not guarantee a certain pension on retirement; instead employer (and employee) contributes over the period of employment to the DC plan, and the employee’s retirement income is a function of many controllable and uncontrollable variables (e.g. market returns). So during the extended asset accumulation phase, the employee is really flying blind not knowing his/her ultimate retirement income. Hammond and Richardson tackle this problem in an interesting paper entitled Retirement Saving Adequacy and Individual Investment Risk Management Using Asset/Salary Ratio” in which they propose the Asset-Salary Ratio (ASR) as feedback mechanism to the DC plan contributor to indicate whether (s)he is on track to accumulate by retirement time sufficient assets to secure a certain income Replacement Ratio (RR). The Replacement Ratio (RR) is the “proportion of pre-retirement income an individual is able to replace through purchasing a guaranteed annuity at time of retirement”. The required ASR to achieve a target income Replacement Ratio is dynamic over time as one is approaching (and saving) for retirement, typically starts near 0 and for Replacement Rates between 50-70% reaches the range of 6-8 near retirement. This is based on certain assumptions pertaining to: asset returns, annuity rates, nominal salary growth and contribution rate. One can then compare at any time one’s current ASR to the required ASR given the years to retirement and target Replacement Rate, and thus have an indication of being on-track, ahead or behind plan and if necessary taking corrective actions (e.g. higher contributions, changed investment strategies, delayed retirement date or reduced expected income in retirement). The authors believe that “it is possible to design robust DC plans which considerably increase the likelihood of achieving sufficient savings for generating adequate retirement income”. (Very interesting and I am sure a feedback process such as this (or similar) is essential for credible next generation retirement income systems.)
The Boettner Center for Pensions and Retirement Research at UPen’s Wharton provides access to readers interested in pension to current research papers from high calibre sources. They have just posted 15 new research papers. I found a couple of these papers particularly interesting, one focused at decumulation strategies in retirement and another on how to strengthen DC pension plans by using a feedback mechanism which both informs and triggers action for DC plan participants.
In “Comparing Spending Approaches in Retirement” Ameriks, Hess and Ren compare decumulation strategies- specifically 9 income generating methods: (1) endowments style systematic payout strategy (annual 5% of previous 12 quarterly NAV of fund), (2) time horizon systematic payout funds (start at 5% exhaust all funds over 30 years), (3) 5% inflation adjusted from a balanced mutual fund, (4) Variable immediate annuity (65 year old male with AIR=3.5%), (5) Variable annuity with Guaranteed Lifetime Income Benefit (65 year old male 5% guaranteed minimum income base with potential step-up and residual value can be withdrawn any time or part of estate , essentially GMWB described before), (6) Required Minimum Distribution (RMD) withdrawal plan, (8) combination strategies, (7) 50% real fixed annuity and 50% RMD, (8) 50% Variable Annuity 3.5% AIR and 50% RMD. (9) 65 year old male real fixed annuity. See paper for assumption on asset allocations, cost, expected returns and standard deviations, etc. The only problem with the analysis is that the assumption on fees are overly generous to variable annuities and GLWBs because while buying an insurance product to protect the downside might have some merit, but the fees associated with the underlying funds are much higher than those available outside of the context of the insurance product. Also, it would have been interesting to see the performance of the combination of an endowment style withdrawal payout strategy on 90% of the funds and a longevity insurance product (joint delayed payout annuity of payments starting at 85) for the remaining 10% of the funds. Actually GMWBs as a product (i.e. buying downside protection for 1%) might even make sense to many, so long that you could buy them together with 0.25% MER index funds, but certainly would not be recommended with a 2.5% mutual fund see my GMWB II blog (but then the 2.5% MER mutual fund doesn’t make sense with or without GMWB).
One of the problems of DC plans (as opposed to DB plans) is that the employer does not guarantee a certain pension on retirement; instead employer (and employee) contributes over the period of employment to the DC plan, and the employee’s retirement income is a function of many controllable and uncontrollable variables (e.g. market returns). So during the extended asset accumulation phase, the employee is really flying blind not knowing his/her ultimate retirement income. Hammond and Richardson tackle this problem in an interesting paper entitled Retirement Saving Adequacy and Individual Investment Risk Management Using Asset/Salary Ratio” in which they propose the Asset-Salary Ratio (ASR) as feedback mechanism to the DC plan contributor to indicate whether (s)he is on track to accumulate by retirement time sufficient assets to secure a certain income Replacement Ratio (RR). The Replacement Ratio (RR) is the “proportion of pre-retirement income an individual is able to replace through purchasing a guaranteed annuity at time of retirement”. The required ASR to achieve a target income Replacement Ratio is dynamic over time as one is approaching (and saving) for retirement, typically starts near 0 and for Replacement Rates between 50-70% reaches the range of 6-8 near retirement. This is based on certain assumptions pertaining to: asset returns, annuity rates, nominal salary growth and contribution rate. One can then compare at any time one’s current ASR to the required ASR given the years to retirement and target Replacement Rate, and thus have an indication of being on-track, ahead or behind plan and if necessary taking corrective actions (e.g. higher contributions, changed investment strategies, delayed retirement date or reduced expected income in retirement). The authors believe that “it is possible to design robust DC plans which considerably increase the likelihood of achieving sufficient savings for generating adequate retirement income”. (Very interesting and I am sure a feedback process such as this (or similar) is essential for credible next generation retirement income systems.)
RetirementAction.com 