Valuation-Informed Indexing #307
by Rob Bennett
I wrote last week about a recent article in the Wall Street Journal (What Is the Sustainable Spending Rate for Retirees in 2016?) by Wade Pfau that adopts my suggestion (from 14 years ago) that a valuation adjustment be included in safe withdrawal rate calculations while also tweaking the concept a bit so that the new calculations cannot be compared directly with the old, discredited ones (Wade calls his new version of the concept “Sustainable Spending Rates”).
Wade writes: “It is not a conservative guess about a safe withdrawal rate, but rather it is the best guess based upon the historical relationship between withdrawal rates, market valuations, and interest rates. It could end up being more or less (we won’t know which for another thirty years). To be conservative, a lower withdrawal rate is required to account for the additional random fluctuations from outside the model. This analysis further confirms the idea that the 4 percent withdrawal rate cannot be treated as safe for retires in today’s market environment.”
A comment was posted to the article that offers a concise description of the problem created by these words. The poster wrote: “Too vague; tell me what you think is reasonable based on the data. 4+% is probably excessive.” There’s a reason why the safe withdrawal rate concept has long been defined as the withdrawal rate that will work in a worst-case scenario; that’s the number that aspiring retirees want to identify. People don’t want to leave their jobs with a belief that it is more likely than not that their plans will work; they want to be sure.
So why does Wade identify the “best guess” number rather than the safe one? I can tell a story from my personal experience that I think hits the nail on the head re what is going on here.
John Walter Russell and I published the first valuation-adjusted retirement calculator in July 2005. John had posted research in June 2003 showing that the safe withdrawal rate for retirees who started their retirements at the top of the bubble (January 2000) was 1.6 percent. I learned a few days before the calculator was to be published that he intended to report the safe withdrawal rate for those retirees as 3.2 percent; that’s the withdrawal rate that gives a retirement a 50 percent chance of surviving.
I was shocked. The core purpose of our work together was to report the safe withdrawal rate accurately. The 3.2 percent number wasn’t accurate, not in a context in which people were seeking to identify the withdrawal rate that would work in a worst-case scenario. I knew John to be a detail-oriented person and an exceedingly honest person. I was mystified as to why he was reluctant to put his name to a calculator that reported the accurate number and okay with putting it to one that reported an inaccurate one.
It turns out that John was one of those darned humans. We are social creatures. We don’t like to disappoint our friends too much. We all know lots of people whose hopes for their futures are riding on continued upward movement in stock prices. An asset class that produced zero gains for 30 years would provide a safe withdrawal rate of 3.3 percent (because 30 years of withdrawals would not reduce the portfolio value quite to zero in that amount of time). If our calculator reported the numbers accurately, it would be showing that stocks were so overpriced at the the top of the bubble that the safe withdrawal rate was one-half of the number that would apply for an asset class with a long-term return of zero. We were delivering a slap in the face to the retirees taking a look at our work.
We ended up delivering that slap in the face. I persuaded John that we had to report accurately what the historical return data revealed regardless of how unpopular it made us. Doing so did indeed make us even more unpopular. I had already been banned at numerous web sites at that time. I have been banned at many more in the years that have followed.
Shiller’s “revolutionary” (his word) finding of 1981 was that it is emotions, not economic developments, that are the primary drivers of stock price changes. It follows that there is virtually no limit to how high we can push stock prices in a bull market. And that there is virtually no limit to the pain we can cause in the economic crisis that inevitably follows as prices fall back to earth and trillions of dollars of spending power are sucked out of the economy. Reporting the safe withdrawal rate accurately when stock prices are insanely high is like telling a fat man who hasn’t gotten on a scale for years because he doesn’t want to know the realities his true weight. It’s a painful hit.
Wade does not pretend to be telling us the safe withdrawal rate. He is not being dishonest. He is being diplomatic.
But those who read between the lines of his new analysis should be more worried than comforted by his adoption of terminology that breaks with the accepted understanding of the safe withdrawal rate concept that has been widely accepted for many years. It was probably my journalism training that compelled me to report the numbers straight no matter how painful they were to hear; it is a reporter’s job to upset people, to be the bearer of bad tidings, to throw cold water on out-of-control optimism. Most of the rest of the good people who work in this field feel at least somewhat constrained by marketing considerations; they don’t just want to be right, they want to be liked.
There’s a sense in which that reality renders Wade’s report more scary than it would had he told the story as bluntly as John Walter Russell and I did a number of years back. He tells us the sustainable spending rate, not the safe withdrawal rate. Not because the safe withdrawal rate cannot be calculated accurately. Because he doesn’t believe that we could bear to hear the accurate numbers.
The most scary reality of all is that the 11 years of reactions that I have seen to the publication of my own retirement calculator shows that he is right.
Rob Bennett’s bio is here.