Valuation-Informed Indexing #258

by Rob Bennett

Michael Kitces has written an important and helpful and yet also highly frustrating and perhaps dangerous article on safe withdrawal rates. It is titled How Has the 4 Percent Rule Held Up Since the Tech Bubble and the 2008 Financial Crisis?

The article focuses on an aspect of the safe-withdrawal-rate matter than many people seem to miss. The safe withdrawal rate as it is defined in the literature is a very conservative number. Many investors (and even some of the experts advising them) seem to think of the safe withdrawal rate as the recommended withdrawal rate. It is not that. The safe withdrawal rate is the withdrawal rate that is virtually certain to work in a worst-case scenario. Investors who want to be super safe might well decide to use the safe withdrawal rate as their own withdrawal rate. But many others might reasonably conclude that they can afford to assume that a worst-case scenario will not pop up for them.

When I developed The Retirement Risk Evaluator (the first SWR calculator that adjusts for the valuation level that applies on the day the retirement begins), I had it identify not only the safe withdrawal rate (the rate that would work in 95 percent of the possible return scenarios) but also the reasonably safe withdrawal rate (the rate that would work in 80 percent of the possible return scenarios). You pay a price for wanting to be super safe. Kitces is doing us a service by showing that the 4 percent rule is in a general sense a very conservative rule.

He argues that the retirements of retirees who retired in 2000 or in 2008 are not looking to be in terrible shape today. This is a valid and important and underreported point. So I am happy that Kitces makes this point. However, I believe that the arguments advanced in the article may be giving those retirees a false confidence in how their retirement portfolios will perform from this point forward.

Kitces makes a critical (and common) mistake, in my assessment. He fails to distinguish the risk that comes from retiring at a time of high valuations and the risk that comes from experiencing a poor returns sequence. He refers to the time-periods that produced the 4 percent rule as time-periods in which we saw poor returns sequences. But it was NOT the returns sequences that caused those retirement failures (for investors taking withdrawals of more than 4 percent).

Consider the retirees who retired at the top of the bubble that brought on the Great Depression. The returns experienced in the years following 1929 were poor. So there is a widespread inclination to say that those we retired in 1929 were “unlucky.” But this is not so!

What the data actually shows is that the returns sequence that we saw from 1929 forward was neither a lucky one nor a lucky one — it was an average returns sequence. The returns were poor. But that was because the valuation level that applied in 1929 was insanely high. Given the valuation level that applied in 1929, the returns experienced in the following years were about what you would expect. It is not possible to make an informed statement as to whether a particular return was a lucky one or an unlucky one without first taking into consideration the valuation level that applied at the beginning of the time-period under examination.

Kitces makes a valid point when he observes that it is not just the size of a price crash that determines whether retirements that began soon before that crash will survive or not but that the length of time before those prices recover is also a significant factor. That’s so. But he makes a mistake when he suggests to those who retired immediately before the 2008 crash are out of the woods because prices recovered so quickly.

Those retirees would have good cause for confidence in their retirement plans if we were at low or even moderate valuation levels today. We are not. We are at insanely high price levels. That means that those retirement plans will soon likely be taking a second big hit. That’s the worst of all worlds. Multiple hits are more likely to cause retirement failure than single sustained hits. For one thing, retirees adjust their spending following single sustained hits while price crashes that are followed by quick recoveries usually do not inspire permanent spending reductions.

The two big factors affecting retirement safety are the valuation level that applies on the day the retirement begins and the nature of the returns sequence by which that valuation level moves in the direction of the fair-value valuation level. It is because the returns sequence (which cannot be predicted) has a significant effect that we cannot make precise long-term return predictions by making reference to valuations. But it is a terrible mistake to believe that it is only the returns sequence that matters and that valuations need not be taken into consideration. The valuations factor is the more important factor by a wide margin.

Rob Bennett’s bio is here.