Executive Summary
As safe withdrawal rate research gains in popularity, it is both increasing used - and misused - by financial planners and the press. Although the research does have its limitations - which I discuss frequently in my presentations at various financial planning conferences throughout the year - and is built on many assumptions that deserve to be challenged, a rising number of safe withdrawal rate critics appear to criticize the approach based on inaccurate statements. So let's clear up a few points of confusion about safe withdrawal rate assumptions.
The inspiration for today's blog post was a recent article by Michael Sivy on Moneyland's TIME column, entitled "The 4% Myth: Where the Wisdom on Retirement Goes Wrong" that discusses what Sivy claims are three incorrect assumptions about safe withdrawal rates. Unfortunately, though, it appears as though the assumptions that the article makes ABOUT the safe withdrawal rate research are even less accurate than the assumptions IN the safe withdrawal rate research!
First of all, the article claims that safe withdrawal rates force clients to save more money than is really necessary, because it assumes that your savings and investments have to provide all of your retirement income and ignores other income sources like Social Security or pensions. Unfortunately, this is absolutely false; the research does NOT assume that all of your retirement income comes from your savings and investments. It simply connects a sustainable withdrawal amount TO an asset base. If your goal is $5k/month and you'll be getting $2k/month of Social Security, you would use the withdrawal rate framework to identify the assets needed to produce the remaining $3k. NOTHING in the research says it would need to be or has to be based on the gross $5k/month spending goal and ignoring other income sources.
Second, the article suggests that once a safe withdrawal rate spending level is set, it can never be changed, while people in the real world are flexible. Yet this is not accurate either. See, for example, the work of Guyton (2004), Guyton & Klinger (2006), other work in the Journal of Financial Planning by Klinger in 2007 and 2010, as well as Guyton's more recent exploration of adopting a Withdrawal Policy Statement for clients. These studies SPECIFICALLY measured the impact of adjusting spending upwards and downwards according to various rule-based frameworks, in the exact manner the Moneyland article criticizes, and show how withdrawal rates can be adjusted at the start in light of anticipated adjustments to market events down the road (as well as when/how to make those adjustments in real time).
Third, the Moneyland article suggests that safe withdrawal rate research is inappropriate because it assumes stocks are sold in down markets, when according to the author retirees could simply own high-dividend bond portfolios that generate sufficient cash flows. However, again, the safe withdrawal rate research does NOT implicitly assume you sell stocks in a down market. In fact, it (generally) assumes that you rebalance annually, which means it will still have you BUY stocks in a bear market, in whatever amounts of excess bonds are remaining after withdrawals. In the classic example (just to use round numbers), if the portfolio has $1M in a 60/40 portfolio, there is $600,000 in stocks and $400,000 in bonds. If the market crashes 40%; the stocks are down to $360k and the bonds are at $400k, for a total balance of $760k. We need to take a $40,000 withdrawal, which we can do from bonds, further reducing the total portfolio to $720k. But we still have to rebalance; to get back to a 60/40 portfolio, we sell MORE bonds to reduce the bond exposure to only $288k and the stock exposure back up to $432k to restore the 60/40 mix. Thus, at no point did we sell stocks in a down market; in fact, we BOUGHT them.
There are some challenges and valid criticisms of safe withdrawal rates. But the "assumptions problems" claimed in this article are unfortunately just a series of either misstatements of the research, misapplications of the research, or just outright fallacies about the research. It seemed like it was worth taking a few minutes to set the record straight.
So what do you think? Do you use safe withdrawal rate discussions in your retirement recommendations with clients? Do you find it an effective approach? What are your criticisms? Have you come across these "assumptions problems" before?
