Executive Summary
As readers of my newsletter know, in May I published research that challenges the safe withdrawal rate as potentially being TOO safe in some environments, where market valuation is not at unfavorable extremes. However, in some feedback I've received from readers, another important point is being made - in some cases, the safe withdrawal rate may also still be too aggressive!
In a compelling blog entry, Rob Bennett of PassionSaving.com makes a good point - that while my research accounts for valuations in one direction, maybe we still haven't given enough weight and attention to what happens when valuations move to the MOST significantly unfavorable extremes.
As my research noted, all of the least favorable "safe" withdrawal rates occur, not surprisingly, when market valuation is unfavorable. However, there's perhaps a difference between unfavorable valuations, and UNFAVORABLE valuations.
For instance, the valuations at the start of the current decade reached extremes far beyond anything we have ever seen in history. Which begs the question - if high market valuations compress returns and thereby compress safe withdrawal rates, is it possible that the record high valuations from year 2000 could produce new record low safe withdrawal rates?
Offhand, I have to admit that I think this is a genuine possibility. To the extent that high valuations compress returns and interest rates, it is certainly reasonable to expect that the record valuations of 2000 will produce record-low long-term returns in the decades that follow - which in turn, could produce new record-low safe withdrawal rates. In point of fact, research by John Walter Russell implies that a safe withdrawal rate from year-2000-level valuations could be under 2%!
It's worth noting that Russell's work is still only hypothetical, although there is a fascinating retirement calculator based on his work that projects the safe withdrawal rate in any particular valuation environment. It remains true that the worst safe withdrawal rate period we've ever ACTUALLY seen ran from 1966 to 1995, and produced our traditional "4%" rule for safe withdrawal rates. Will 2000-2029 ultimately produce a worse result? Yes, it's quite possible, and even likely, given the incredibly distorted valuations, but only time will tell.
So what should we do from here? Well, the "good" news is that today's valuations are no longer anywhere near the year 2000 extremes. Instead, we're merely in the "high" valuation environment that produced the kind of 4% safe withdrawal rates we've seen at similar high valuation points in history - which means we should be comfortable with the 4% safe withdrawal rates being applied today.
On the other hand, for those who already retired at the start of the decade following the 4% rule, it may be time to sit down and update the plan, and face some hard facts given that the portfolio will almost assuredly be underperforming any original projections at this point. If Russell's projections of a sub-2% withdrawal rate prove to be true, or are anywhere close, those who retired back in 2000 may ultimately find that the "4% safe withdrawal rate" was still far too aggressive, making the point once again about how critical it is to incorporate market valuation into retirement projections!
Michael,
Very interesting.
And if I could add another wrinke here, during this time of horrible stock market returns since 2000, we have had rising inflation as well! This combo makes it extremely tough. Even Bengen talked about this combo back in 1994. We know that stocks underperform during inflationary times.
I know that some will argue that we are not in an inflationary period, but I ask: can you really trust government headline inflation?
Apart from that, if you take a look at the type of inflation retirees face, there is no question that perhaps we overestimate “real” returns for this cohort. For example, seniors typically do not buy the types of products and services that experience “deflation”, like consumer electronics etc. Instead, they spend their money on products and services that tend to experience high inflation, like medical care, etc.
To account for this, as you know, most retirees just spend less in other areas. In other words, the assumption is that retirees just rearrange their basket of consumption. I think it is more realistic to treat these high inflationary items more as a “tax” that people have to pay in addition to maintaining their regular consumption.
I’m just talking off the top of my head, but it seems to me that if you are experiencing high inflation in food and energy (the staples) while at the same time experiencing high inflation on other items, the reusult will be a reduction in purchasing power over time and that the compounding long-term effect of that will be devastating, particularly during a period of low stock market returns (and I may add, low fixed income returns!).
Doug
.
.
.
Hi Mike,
what’s your take on the two recent articles in the SEP 2008 FPA Journal:
“When Should Retirees Retrench?” By Gordon Pye
Link:
http://www.fpajournal.org/CurrentIssue/TableofContents/WhenShouldRetireesRetrenchLaterThanYouThink/
And
“Data Dependence and Sustainable Real Withdrawal Rates”, by the Blanchettts
Link:
http://www.fpajournal.org/CurrentIssue/TableofContents/DataDependenceandSustainableRealWithdrawalRates/
Also, why your discussions so far, relate to SWRs from a portfolio without references to the investors’ withdrawal horizons such as per Moshe Milevsky’s Gamma Distribution “Retirement Ruin formula?”
I continue from my post above [mentioning Moshe Milevsky’s Retirement Ruin Formula (RRF)]
Mike wrote above:
“It’s worth noting that Russell’s work is still only hypothetical, although there is a FASCINATING RETIREMENT CALCULATOR based on his work that projects the safe withdrawal rate in any particular valuation environment.”
The calculator can be seen at this site:
http://www.passionsaving.com/stock-valuation.html
Question: if we could use the calculator to enter a probability distribution for the expected portfolio return, into Moshe Milevsky’s RRF, wouldn’t we achieve a more accurate retirement ruin prediction?
Explanation:
In the formula for the retirement ruin probability, the parameters are:
Portfolio’s expected return (take from the calculator)
Investment volatility (SD) [doesn’t change much over time for same allocation]
Mortality rate (from probability distribution)
Initial nest egg $s per dollar of desired spending
For those unfamiliar with the formula please check this report:
http://www.qwema.ca/pdf_research/2007JULY_RRQ.pdf
I authored a spreadsheet, TIP$TER, which I’ve hosted at http://www.prospercuity.com, that takes mortality rates into account in estimating shortfall risk. TIP$TER is a valuation-based monte carlo simulator. It asks the user to enter the “expected risk premium” for a broad stock market index. Logically, the more richly stocks are valued, the less the expected risk premium.
TIP$TER also compares the retirement expenditures that a simulated equity portfolio would be likely to sustain with the retirement expenditures a 100%-TIPS portfolio (in tax-sheltered vehicles like IRAs or 401Ks) would sustain. This, I believe, provides some very helpful insight. Often, retirees are advised to invest aggressively, but spend defensively (e.g., a 4% or less SWR), even though it would require the retiree to live a more frugal lifestyle than a 100%-TIPS portfolio could sustain.
I read your report “Resolving the Paradox–Is the Safe Withdrawal Rate Sometimes Too Safe?” and I was also utterly bewildered that you would conclude that withdrawal rate could only increase with P/E10 and not decrease.
Looking at Figure 6 Starting P/E10 vs. SWR from 1881 to 1972. It shows the highest P/E10 was around 30 back in 1930 and the SWR was a little over 4%.
But we all know in 1999 P/E10 peaked at 44 which is about 50% higher than the highest in the chart.
Using the same logic that leads you to conclude that higher SWF are possible at lower PE/10, would also lead to the conclusion that at higher the SWF should be lower.
Michael, you have developed and deserve a reputation for fair, honest, open minded analysis and writing. So I ask you to depend on that and let this comment through, because it seems only fair to note:
Rob Bennett’s record as a “financial analyst” speaks for itself. Two places he posted frequently as “Hocus” are:
http://www.s152957355.onlinehome.us/cgi-bin/yabb2/YaBB.pl?board=HOCO
as well as
http://s162532268.onlinehome.us/Sewer/viewforum.php?f=1
where as moderator, he deleted posts very frequently for disagreeing with him.
Of late, he frequently ascribes words and contexts to you, that I cannot find substantiated in your own material. This includes some of his 33 podcasts (to date).
If I were you, I would run not walk to distance myself from the methods and views of this layman hobbyist who admits he cannot even balance his own checkbook, yet espouses ‘grave analytical errors’ existing in studies he simply can’t manage to understand in context.
Because of his own shortcomings in understanding, the man has held a vendetta for ten years against John Greany, a respected and valued contributor to much material of use to early retirees. When Bennett is asked to provide any objective argument for his various claims, he says death threats or other factors prevent him from doing so.
In a word, he is a “kook.”
Dear Michael,
I enjoyed reading your May 2008 newsletter and also your blog post. I have cited them both in an article I just finished on this topic.
My paper can be found on SSRN:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1726225
Also, I’d otherwise like to introduce myself to you as I hope to become more involved in FPA events and joint the FPA annual conferences.
I have an article in the newest Journal of Financial Planning, and I have been really impressed by that journal. In the last couple of weeks I think more people have looked at that article than the history of all the other articles I ever published. It is this one:
http://www.fpanet.org/journal/CurrentIssue/TableofContents/AnInternationalPerspectiveonSafeWithdrawalRates/
Thanks, Wade