This quarter’s edition of The Journal of Retirement features articles by Michael Finke and Jason Fichtner, Alicia Munnell and Anthony Webb, Mark Warshawsky, and Srinidhi Kanuri. The edition also features three book reviews and a guest column by Warren Cormier of the Defined Contribution Institutional Investment Association (DCIIA), the official association partner of JOR. We thank Warren and the authors of the book reviews for their invaluable contributions.
The article by Michael Finke and Jason Fichtner reports results of a survey on employer attitudes toward partial annuitization. The authors find that almost half of respondents expressed a preference for a mixture of lifetime income and investments, with only a quarter preferring just a monthly pension and another quarter preferring just an investment account. As Anthony Webb and others have shown, annuities are a highly effective means of insuring the tail risk of living to very advanced ages. Consistent with this finding, over 90% of respondents reported that they would choose to invest at least some of their wealth in an advanced life deferred annuity, priced at commercial rates, with payments starting at age 80.
Take-up of advanced life deferred annuities is, of course, vanishingly small. What is causing the disconnect between, on the one hand, both theoretical models and stated preferences, and on the other hand, real-world behavior. Is it simply a lack of awareness? We suspect that procrastination may play a role. The cost of postponing the purchase of an advanced life deferred annuity by a year is not so large—the rate for a 66-year-old is not so different from that for a 65-year-old. But the perceived cost of making an irreversible mistake may loom large. So, why not postpone the decision? We view procrastination in relation to annuity purchase as a fertile area for research.
The article by Alicia Munnell and Anthony Webb examines the impact of leakages on retirement account balances. In theory, workers contributing to defined-contribution plans for a lifetime should accumulate more than enough for a secure retirement. In practice, plan balances fall far short of projections. Many factors contribute, including the immaturity of the system (workers now retiring would have entered the labor force in the early 1980s, when coverage was far lower than today), spotty eligibility and participation, high fees, and leakages. The authors estimate that leakages reduce plan balances by an average of 20%, although that average hides a wide variation.
So, what should policymakers do? One response would be to simply prohibit leakages, many of which occur on job-change. Few would object to prohibiting leakages used to buy a boat or to pay for a vacation. But many households withdraw from their accounts to alleviate economic hardship. Prohibiting all leakages would exacerbate that hardship and might discourage participation in the first place. Some, noting that many American households would struggle to raise even $500 in an emergency, have proposed that employers should add the option to contribute to a “rainy day fund” through payroll deductions. But households already face an alphabet soup of accounts. Do we really need to add more complexity? And although rainy day funds may come in handy if the car needs a repair, they are hardly going to solve the problems caused by extended periods of unemployment.
We believe much can be achieved through the automatic rollover of small account balances. Years ago, when Anthony Webb was at the Center for Retirement Research, he met Spencer Williams and Tom Johnson, who had an idea for automating the transfer of 401(k) plan balances from one plan to another. They really were trying to do good. Anthony wished them luck, but was certain they would never overcome the myriad technical and regulatory barriers to implementation. They did, and Vanguard adopted their solution during 2021. This is a big win for small and financially unsophisticated savers who would otherwise cash out or forget about their savings.
The article by Mark Warshawsky is entitled “The Trouble with State Government Employee Pension Plans: The Case of Connecticut.” The article is worth reading for two reasons. First, it contains a fair and balanced explanation of differing viewpoints as to how defined-benefit retirement plan liabilities should be valued, written in nontechnical language. Second, it projects the financial status of the Connecticut plans in 2030, using stochastic modeling. If the returns earned by state and local retirement plans equal those assumed in their actuarial models (and if required contributions are made, a big assumption), most plans will be in tolerable shape. But as investors are constantly reminded, past performance is no guarantee of future returns, and it is the taxpayers who will be mostly on the hook if returns fall short. The Warshawsky article shows that plausible variations in investment returns can have a big impact on the funded ratio. We believe that, if anything, the Warshawsky article understates the risk because the model assumes draws from historical returns, whereas most economists believe that investment returns will be lower going forward.
The article by Srinidhi Kanuri investigates the performance of lifecycle funds and finds that after deduction of fees, they have delivered negative alpha. This is a significant finding, but it leaves open two important questions. First, to know whether lifecycle funds add value, we need to know what households would have done in their absence; we also need to have a metric to rank outcomes. In theory, 401(k) plans enable households to tailor their portfolios to match their ability and willingness to bear risk. In practice, given generally low levels of financial literacy, it is far from clear that many households are able to make informed choices or that the effects of financial education are so powerful as to enable households to make appropriate use of the choices available to them. Lifecycle funds may deliver better outcomes than the alternative, and those better outcomes may exceed the fees charged. Second, the one-size-fits-all approach of lifecycle funds is clearly suboptimal. The question arises—can employers use the information in their possession about their employees to achieve more appropriate defaults? We encourage submissions that address these questions.
Brett Hammond
Co-Editor
Anthony Webb
Co-Editor
- © 2022 With Intelligence Ltd