by Rob Bennett
I gained my fame (infamy?) on the internet by being the person who discovered the errors in the Old School safe withdrawal rate (SWR) studies, studies that millions of middle-class people have used to plan their retirements. The flaw that I discovered was easy to spot and easy to understand and yet the implications of the discovery have proven to be profound and far-reaching.
The studies don’t adjust for valuations!
Prescience Partners returned 6.75% for the second quarter, underperforming the S&P 500's 8.55% return but coming out ahead of the Barclay Equity Long/ Short Index's 2.62% return. However, for the first six months of the year, Prescience is up 30.66%, doubling the S&P's 15.25% return and smashing the Barclay Equity Long/ Short Index's 9.27% return. Read More
It’s so obvious, isn’t it? Robert Shiller showed in 1981 that valuations affect long-term returns. The long-term return obtained by a retirement portfolio obviously affects whether the retirement plan is likely to survive or not. And so it is obviously not possible to determine whether a specified withdrawal rate is safe or not without taking into account the valuation level that applies on the day the retirement begins.
The safe withdrawal rate is not a single number, as was once believed to be the case (the Old School studies say that the SWR is always 4 percent) but a number that varies. The SWR can be 3 percent. The SWR can be 5 percent. The SWR can be 7 percent. It depends.
On what? On whether stocks are undervalued, fairly valued or overvalued on the day the retirement begins. The valuation level that applies on the retirement day tells us how much of the starting-point portfolio value is real, lasting wealth and how much is cotton-candy nothingness fated to be blown away in the wind over the course of the next 10 years or so.
It sounds like such a simple thing that many people have told me that they are sure that someone must have discovered the error before I did. Perhaps (if you know of anyone who did so, I would be grateful if you would let me know). I doubt it, however. The reality is that, while I have known since the late 1990s that the Old School SWR studies get all the numbers wildly wrong, I am myself still in the process of coming to terms with the implications of the idea that valuations affect SWRs (and everything else relating to stock investing, to be sure).
I remember the reaction of a fellow I got to know on discussion boards who posted under the screen-name “BenSolar.” I got a lot of heat from Buy-and-Holders for being the person who discovered the error in the SWR studies (the cause of the error was a belief by the people who developed the Old School methodology in the Efficient Market concept and in the Buy-and-Hold Model in general). Ben thought that I was on the right track and defended me numerous times, causing a lot of bricks to be thrown in his direction as well.
A group of us founded a board (“The SWR Research Group”) with the purpose of ascertaining the accurately calculated SWR. Our lead researcher (John Walter Russell) determined through use of a regression analysis that, at the tip of the bull, the SWR for an 80 percent stock portfolio had dropped to 1.6 percent. That means that a retiree with a $1 million portfolio would not be safe taking out of his portfolio more than an inflation-adjusted $16,000 in each year of his retirement. This was a big change from the claim made by the Old School studies, that a $40,000 annual withdrawal was perfectly safe.
Ben’s reaction was telling. He refused to believe the numbers! He agreed that a valuations adjustment was required to compute the SWR accurately. And he had full confidence in Russell’s work, having worked with him over a long period of time and having learned that he was a careful straight-shooter of a guy. Still, Ben could not accept a 1.6 SWR. “That just cannot be right!” he exclaimed.
I related to the feeling. Having been the person who advocated the effort to calculate the retirement numbers correctly, I was in no position to deny the mathematical reality when it was reported to me. Still, I did acknowledge that the number shocked me. I encouraged any community members who could find any flaws in Russell’s methodology to please speak up. No one did. So over time I came to acceptance of the new number. But it was a hard thing to do so, even for someone who believes strongly that Shiller is right about valuations.
I faced an even more shocking reality a short time later, when Russell identified the SWR that applied for those who retired in the early 1980s. It’s one thing to tell retirees that they should not place confidence in claims that it is safe to withdrawal 4 percent even at times of insanely inflated stock prices. At least in those circumstances the person arguing for the New School methodology is arguing for greater caution, generally a good thing when it comes to retirement planning.
But valuations do not only pull the SWR down at times of insanely high valuations, They also pull the SWR up at times of insanely low valuations. If we believe that the numbers really do reveal anything of value, we need to be willing to let retirees know when the Old School claims are causing them to be excessively cautious as well as when the Old School claims are causing them to take on crazy risks. The same historical data that shows that the SWR can drop to as low as 1.6 percent also shows that it can rise to a number a lot higher than 4 percent.
That’s the number.
I had a BenSolar-type reactions when I learned the result of the calculation. “Sure, that’s what the numbers say,” I thought, “but I’ll be darned if I am going to tell retirees that there are circumstances when it is safe for someone who has put away $1 million to take out $90,000 every year to cover living expenses.” It’s irresponsible! It’s dangerous! It’s loco!
Or is it?
Stocks were priced at one-half fair value in the early 1980s. They were priced at three times fair value at the top of the bubble. That’s a difference of 600 percent. Take the low number (1.6 percent), round it off to 1.5 percent to make the math easy, and multiply by 6 to see what the SWR should be when stocks are priced where they were in the early 1980s and the answer you get is — 9 percent!
The numbers aren’t shocking at all. They are perfectly logical.
They seem shocking because we all are affected by the psychology of a mad bull, even those of us who like to entertain the idea that we are above that sort of thing.
None of us can think clearly about valuations. Even Rob Bennett, the fellow who never shuts up about the importance of taking valuations into consideration when forming any stock investment strategy imaginable. It’s not just my critics who have a hard time taking this stuff in. I do too. The realities of overvaluation are intellectually easy to identify but emotionally hard to accept.
I try to keep that in mind when I reflect on how hard a time the Buy-and-Holders are having coming to terms with the amazing things we have learned in recent years about how stock investing works in the real world.
Rob Bennett enjoys learning about the best ways to save money. His bio is here.