Editor’s Letter

FAST NEWS AND TIMELY RESEARCH

Rightly concerned about the knock-on effects of the COVID-19 pandemic, the US retirement research and policy community has begun to examine both immediate and longer-term impacts on retirement savings, plan solvency, and individual financial security. For example, based on its defined-contribution Retirement Security Projection Model, a recent Employee Benefit Research Institute (EBRI) report tells us that the aggregate US retirement savings deficit, already large, will increase about 4.5% using intermediate assumptions and about 11% in the worst case (Vanderhei 2020). Unsurprisingly, some of this is due to declines in asset values, which will fall more heavily on those nearer to retirement, and some to plan terminations and contribution suspensions, which will fall more heavily on younger workers. Although not driven by the pandemic, a different EBRI model simulates individual savings required to cover retiree health care costs, which, while remaining substantial, have fallen recently for the first time since 2013 (Fronstin and Vanderhei 2020). This is good news for retirees and those saving for retirement as it can offset a fraction of total retirement spending, but perhaps little comfort for millions who have lost current health insurance coverage along with their jobs.

Other research organizations and individuals are now working on COVID-19 impacts and implications across a wide range of areas, including the economy; financial markets; social institutions; governments and, yes, retirement programs; individual retirement-related behavior; and financial security. We can expect to see more to appear, and The Journal of Retirement will be looking for high-quality work.

Beyond rapid research followed later by deeper analyses, retirement-oriented organizations and industry groups have sponsored informational research-based seminars and other resources on the effects of the pandemic. These include Defined Contribution Institutional Investment Association (DCIIA),¹ National Association of Defined Contribution Administrators (NAGDA),² Life Insurance and Market Research Association (LIMRA),³ and many other groups.⁴ In addition, legal, financial, and consulting firms with retirement-related practices and products are now in rapid-response mode.

Much of our ability to do all this and do it with speed rests on advances in modeling techniques, data collection, and computing power in the last three decades. Although not focused on the virus, I would point especially to three of the articles in this issue. “Participation and Pre-Retirement Withdrawals in Oregon’s Auto-IRA,” by Laura D. Quinby, Alicia H. Munnell, Wenliang Hou, Anek Belbase, and Geoffrey T. Sanzenbacher, presents early experience with an important state-level public policy initiative to extend retirement plan coverage to approximately 50% of previously uncovered private-sector employees. These results are likely to be of intense interest to policymakers in states such as Illinois and California who will be offering similar programs, other states that may be considering such programs, and all who are looking for ways to increase individual financial security. While admittedly not definitive, this study relies on improvements in data collection and laptop-level analytical tools to measure the increase in participation rates as well as the level and rate of withdrawals (a form of leakage) among participants.

The article “Financial Literacy and Wellness among African-Americans: New Insights from the Personal Finance (P-Fin) Index,” by Paul J. Yakoboski, Annamaria Lusardi, and Andrea Hasler, also relies on a new data set to provide much-needed illumination regarding 13% of the US population. Patterns of personal financial knowledge among African-Americans are similar to those among white Americans, (e.g., better among men, more highly educated, older and higher income). Unfortunately, the overall level of knowledge among African-Americans is lower, suggesting that education and other programs should emphasize raising financial literacy within this important group.

Another article that illustrates advances in the field, in this case in the area of methodology, is the fascinating piece by Gordan Irlam, “Machine Learning for Retirement Planning.” He acknowledges the contributions of modern portfolio theory, Merton’s continuous time portfolio approach, stochastic dynamic programming, Monte Carlo and Bayesian methods to improving our ability to solve the “portfolio problem” (i.e., figure out how much to consume now, how much and where to invest, and what to expect to consume later). Machine learning (artificial intelligence) has been used in a wide variety of fields, but this is its first use to assist with retirement modeling. Irlam offers a simple application to portfolio solution challenges, going beyond the usual modeling issues to incorporate effects such as income taxes, mean reverting asset classes, time-varying bond yield curves, and guaranteed income, all of which complicate and sometimes stump traditional approaches.

In the area of guaranteed income, David Blanchett returns to these pages with “The Value of Allocating to Annuities.” The value of this article is that he uses an understandable model to estimate the total additional benefit generated from allocating to annuities based on the percentage of assets actually used to purchase the annuity. Using conservative assumptions and several different scenarios, he shows that the optimal allocation to annuities is about 30% of total assets for the client of a financial advisor. He also shows that this mixed approach (marked-to-market assets plus annuities) can achieve a long-term superior benefit to the client in comparison with 100% allocation to nonannuity assets.

Also returning are William Reichenstein and William Meyer, in “Investment Implications of the Rising and Falling Pattern of Marginal Tax Rates for Retirees.” They look at the nonlinear interactions among Medicare benefits, taxation of Social Security benefits, and changing retirement income. They go on to propose how a rising and falling pattern of effective marginal tax rates due to these interactions should affect optimal retirement plan withdrawal strategies.

“Measuring Sequence of Returns Risk,” by Andrew Clare, Simon Glover, James Seaton, Peter N. Smith, and Stephen Thomas, is closely related to two articles in a recent issue of this journal (see the articles by Fox and Kintzel, respectively, in Vol. 7, No. 3). All of these address a critical issue in achieving financial security through investment results, namely, that an identical long-term return (arithmetic or geometric) can be successful or unsuccessful in providing adequate retirement income, depending on when bad or good subperiod returns occur. In this article the authors demonstrate the problem and propose three measures investors can use to gauge their sequence of returns risk and, therefore, improve portfolio management.

Brett Hammond

Editor