A consensus is emerging among researchers focused on the topic of retirement income planning that simplistic rules of thumb, such as the traditional 4% safe withdrawal rule, are woefully inadequate to guide investors to optimal outcomes in retirement.
At the same time, researchers are also calling into question the usefulness of more sophisticated but potentially equally problematic approaches to managing retirement income — particularly those that rely heavily on binary failure metrics generated by poorly contextualized Monte Carlo simulations, as well as those which rely on solving esoteric utility functions that are more likely to confuse than inform the typical investor.
Against this backdrop, a new analysis published in the Journal of Financial Planning by Javier Estrada of the IESE Business School in Barcelona, Spain, proposes an alternative framework for modeling retirement income that strives to cut something of a middle ground, using a concept known as the "risk-adjusted coverage ratio."
According to Estrada, when selecting an optimal retirement strategy, a retiree may aim to maximize the coverage ratio, which he calls "a novel metric superior to the failure rate."
Estrada's article suggests retirement savers and their advisors focus on the risk-adjusted distribution of coverage ratios while setting up an income plan. Although such an approach may not be as neat as making decisions based on optimizing a single variable, Estrada proposes, it does enable the consideration of the relevant trade-offs a retiree needs to evaluate in order to find an ideal retirement strategy.
Understanding Coverage Ratios
As Estrada writes, there are generally two critical variables that retirees and their financial planners need to consider when deciding an optimal retirement income strategy. These are the initial withdrawal rate and the portfolio's asset allocation.
"The standard methodology to make these choices properly is to first select a target variable that needs to be maximized or minimized, and then to choose the optimal initial withdrawal rate and asset allocation that solve the optimization problem," Estrada explains. "Although there is a substantial literature on target variables to be considered, the most typical choice is the failure rate; that is, the proportion of retirement periods in which a strategy failed to sustain a retiree's planned withdrawals."
However, as Estrada points out, this approach has two major flaws. It neither distinguishes between a failure early or late in a retirement period, nor does it account for the size and value of any bequest left.
In order to overcome both flaws, Estrada previously introduced the concept of coverage ratios in a 2019 paper (co-authored with MIT's Mark Kritzman). Put simply, the coverage ratio represents the number of years of withdrawals supported by a strategy relative to the length of the retirement period considered.
Thus, the coverage ratio variable directly takes into account how early or late a strategy fails when it does — and it speaks to the size of the bequest when one is left.
In the earlier paper, Estrada also proposes a utility function that penalizes failures more than it rewards bequests, and he suggests that the optimal strategy is the one that maximizes the expected utility of the coverage ratios across all the retirement periods considered.
As the new paper points out, while selecting an optimal retirement strategy following this approach clearly improves upon selecting the strategy that merely minimizes the failure rate, a potential shortcoming is that retirees are typically not familiar with utility functions.
Hence, Estrada warns, they would not be able to implement the methodology themselves, nor would they likely welcome or even understand their financial planners' explanation of the underlying approach.
A Better Way
As Estrada writes, one potential solution may be to skip the utility function and simply select the strategy that yields the highest average coverage ratio across all the retirement periods considered, but this alternative may suffer from a typical problem with averages.
"Just like the [proverbial] individual that drowned crossing a river four feet deep on average, a high average coverage ratio may be hiding many periods in which a strategy failed, which are compensated by a few periods in which the strategy left very large bequests," Estrada explains.
As such, the approach proposed in the new article, besides avoiding utility functions, also avoids focusing on just one average number. In fact, Estrada's suggestion is to focus on the whole distribution of coverage ratios, or at least on some relevant percentiles of such a distribution.
"Doing so would enable a retiree to carefully consider not just the mean or median coverage ratio but also the coverage ratios that may happen with a low probability, particularly on the left tail of the distribution," Estrada argues, referring to scenarios in which the strategy fails. "Clearly, this approach is not as neat as selecting a strategy that maximizes or minimizes the value of a target variable, but it does enable retirees to consider likely (average) scenarios and unlikely (positive or negative) scenarios, as well as to weigh them according to their individual preferences."
