The third issue of The Journal of Retirement once again showcases articles that address a broad range of issues in retirement security. The first article, by Sylvester Schieber and Betty Miller, deals with an extremely important statistical issue: namely, the measurement of retirement income. It is standard practice to rely on data from the Current Population Survey (CPS) to measure income from pensions, both defined-benefit (DB) and defined-contribution (DC) and annuities. However, pension income is defined to include only regular payments (like the monthly income received by DB pensioners and annuitants) and not the generally irregular withdrawals from DC plans. What treatment is correct really turns on how income should be defined. Is a drawdown of both principal and interest correctly treated as income?
Billie Jean Miller and Sylvester Schieber make a strong case that the CPS data substantially underestimate private pension income, especially that enjoyed by middle and upper middle income groups. They argue that the use of the CPS data leads many observers to overstate a perceived crisis in retirement security, although they do acknowledge that there are good reasons to believe that some workers are not preparing adequately for retirement.
Meir Statman’s article is a stimulating think piece that offers a fresh look at the whole issue of retirement security. He divides the population into four groups: the wealthy, the steady-middle, the precarious-middle, and the poor. The wealthy have enough money that retirement in security is not really an issue. The steady-middle save enough of their adequate income that they too have enough income in retirement, although they may not actually feel secure. The precarious-middle suffer from problems of self-control, which means that they do not save enough for retirement. The poor may or may not have problems of self-control, but their working income is simply insufficient to support them, however modestly, even in retirement. Statman relies on brief portraits of members of the middle and poor classes to illustrate his points. He is skeptical of the utility of so-called “nudge” solutions, like automatic enrollment in 401(k) plans, to the problems of the precarious-middle and poor and instead proposes a mandatory defined-contribution plan to complement Social Security to help the precarious-middle, as well as increased redistribution to help the poor, who simply cannot save enough, voluntarily or otherwise. He makes a strong case that the financial problems of the poor do not simply reflect fecklessness.
Students of retirement finance like to divide the process into the two phases of accumulation and decumulation. The accumulation stage coincides with a person’s working lifetime, and the decumulation phase begins as a general rule upon retirement. A great deal has been written over the years on the accumulation phase—on how much to save and how to invest savings. Until quite recently, however, the decumulation phase was overlooked. That is in some respects puzzling, because the decumulation phase presents challenges both to specialists and the ordinary citizen approaching retirement. This issue of the JOR helps make up for the neglect of the decumulation phase by devoting four articles to the subject, as well as an article to the accumulation phase of a relatively new type of variable annuity.
Paul Newfield’s article will introduce many of the JOR’s readers to a close cousin of the life annuity, which was originally called the tontine, and which he names the pooled survival fund, or PSF. I’ll leave it to the reader to explore the details of his proposal, but the basic idea is that a group of older people invests part of the balance of its DC plan in the PSF, and when any member of the pool dies, his or her assets are reallocated to the pool’s surviving members. The longer a member lives, the greater his share of the assets of members who do not survive. As a result, a PSF provides a sort of longevity insurance, although the income it generates is not fixed, like that of the immediate life annuity. Newfield argues that despite its variability, the income from a PSF should, on average, exceed that of a standard life annuity—possibly substantially. He provides the reader with a case study of Australia, in which he considers how the PSF might be grafted onto the existing pension system, and argues that PSFs would not require a major change to the regulatory framework of that country.
The next two articles address the merits of life annuities and similar instruments. David Blanchett’s article is particularly timely, in that he considers how the current low-interest-rate environment should affect the asset allocation of the portfolios of older Americans. He uses a complex simulation model to consider how different allocations affect the tradeoff between the risk of a shortfall in income—due to disappointing asset returns or an unusually long post-retirement life—and the variability of a bequest. His model has five asset classes: stocks, cash, bonds, a life annuity, and a guaranteed lifetime withdrawal benefit (GLWB) annuity. His measure of income is utility based, which effectively penalizes an investment strategy that generates highly variable income. He further assumes that the investor is a 65-year-old couple and takes into account the couple’s uncertain lifespan. Blanchett comes to the surprising conclusion that annuities and GLWB annuities, which are similar to a life annuity, tend to dominate bonds and cash. This asset allocation is very different from the asset allocation typically recommended to older Americans, which places a large share of assets in bonds.
Steve Vernon also uses a simulation exercise to explore how annuities and other retirement income generators (RIGs) compare with one another. He reports on the behavior of six RIGs: a fixed-income life annuity, an inflation-adjusted annuity, a guaranteed minimum withdrawal benefit (GMWB), a constant real withdrawal rule, a constant percentage of remaining assets rule (4%), and a withdrawal rule determined by remaining life expectancy. The simulation exercises assume that the whole of the retirement nest egg is invested in one or another of the RIGs and that all of the income that each RIG generates is spent. With a nest egg of $100,000, initial income ranges from about $5,500 for the fixed annuity to $3,500 for the inflation-adjusted annuity.
The results of three different scenarios are compared: an unfavorable market scenario, a median scenario, and a favorable scenario. In the median strategy, the constant percentage withdrawals strategy and the GMWB are clear losers as income generators, because the market’s performance is not good enough to offset the decline in the real value of assets in the case of the former strategy and because it does not offset the extra fees that a GMWB charges. Under unfavorable conditions, the constant real withdrawals strategy and the life-expectancy-based strategy are also losers. In the favorable scenario, however, all the withdrawal strategies that are driven by the value of accumulated assets perform much better. As might be expected, well-performing RIGs do not perform well if the criterion is switched to remaining wealth.
Vernon’s article is an excellent pedagogical piece on the basic properties of annuities and competing instruments. It also includes a section on how financial advisors can answer some of the questions that clients unfamiliar with annuities could raise and another section on the circumstances in which investing part of the nest egg in an annuity is a good idea. For some investors, a combination of systematic withdrawals and an annuity could work well.
The conventional economic model of the consumer assumes that he or she “rationally” makes choices that best suit his/her preferences (or that maximize utility). The homo economicus choosing to invest money in a tax-favored investment like an IRA should be concerned with the total tax relief he can expect to receive in present value terms—and not with the form that the relief takes. A tax credit of the same value as a tax deduction should not have any effect on his decision.
In practice, however, as the article by John Scott and Jeffrey Diebold explains, the form that the relief takes can make a difference. An upfront rebate to encourage saving can appeal more than a series of uncertain tax deductions of equal actuarial value (i.e., the same present value taking mortality into account). The same could be true of schemes to promote the purchase of annuities. The authors conduct an experiment with a group of 301 subjects, better educated than the population at large, who are divided into groups, with each group being presented with an investment decision in which an annuity is one possible choice. The experiment simulated the relative attractiveness of a credit versus a deduction (which takes the form of a reduction in the cost of living). In one group, a credit encouraged the purchase of an annuity; in two others, the choice was encouraged by tax deductions of different values. A control group did not enjoy either tax incentive. The subjects stayed in the game for a number of periods until they ran out of money or “died,” with the later event being determined randomly.
The authors find that a credit is more effective in encouraging annuity purchases than a deduction of the same actuarial value—and even more effective than a deduction of greater value. Drawing conclusions for policy from this type of behavioral experiment is difficult, but its results do suggest that if annuity purchases were to be encouraged by making them tax favored, a tax credit would be more successful than a deduction.
A new type of variable annuity, the structured product–based variable annuity (spVA), has developed a market of close to $2 billion a year. This new instrument is more complex than traditional variable annuities, because its payout is based on a complex formula applied to the returns of a reference asset—similar to structured products. In their article, Geng Deng, Tim Dulaney, Tim Husson, and Craig McCann show how the value of these products can be replicated by a combination of put and call options and valued using the Black–Scholes model.
The authors note that the level of the cap on returns is not disclosed until the investment is made, making an ex ante assessment difficult from the investor’s perspective. The authors calculate what they call fair cap levels, which are the levels high enough to make the estimated value of the product equal to the premium paid, less reasonable fees. The complexity of the product, which the article explains with great clarity, means that investors may not be able to make accurate comparisons with other investments, like mutual funds or the traditional variable annuity. The authors conclude that the complexity of the spVA raises questions of its suitability for many older Americans.
The final article, by John Turner, David McCarthy, and Norman Stein, addresses the question of the proper role of tax incentives for retirement saving. The holdings of some individuals in 401(k) plans in particular are very large. Having examined data from the 5500 returns that pension plans are obligated to file, the authors find that one employer-provided plan had two participants with average account balances of over $253 million. To gain a better idea of the holdings of individuals, not just in 401(k) plans, but in IRAs, Roth IRAs and other tax-favored forms of retirement saving, the study analyzes data from the 2010 Survey of Consumer Finances, which reports the combined holdings of all these instruments by the surveyed households. The authors estimate that there are more than 175,000 households with combined retirement saving assets of over $3 million and more than 3,000 households with assets of $10 million or more.
The authors argue that Congress intended the retirement saving tax incentives to promote genuine retirement saving—not to create a tax shelter. They argue that the holdings of some households go well beyond that needed for a reasonable standard of living in retirement. After surveying the international practice—Australia, Ireland, and the U.K. have limits on tax-favored contributions and on either the balance or the income from DC plans—they propose a cap of $5 million, to apply to the total holdings of an individual, and briefly discuss the proposal’s administrative implications.
I am confident that readers will find, as I do, that these studies provide great deal of food for thought for all professionals concerned with retirement security.
TOPICS: Retirement, pension funds, derivatives
George A. (Sandy) Mackenzie
Editor
Footnotes
Publisher’s Note
Institutional Investor, the publisher of The Journal of Retirement, wants to extend a special thank you to Bank of America Merrill Lynch for its support of The Journal of Retirement. Please note that no sponsor has influence on the editorial content found in The Journal of Retirement. Representatives from any firm are encouraged to submit an article to our independent Editor, Sandy Mackenzie, for review and prospective acceptance into the publication. All editorial submissions, acceptance, and revisions are the sole decision of Mr. Mackenzie. The editorial submission guidelines are found on the last page of the publication and online at www.iijor.com. Thank you, and I hope you enjoy this and future issues of The Journal of Retirement.
Allison Adams
Group Publisher, Institutional Investor Journals, aadams{at}iijournals.com
- © 2014 Pageant Media Ltd