Is the 4% withdrawal rate foolproof?
Shortly after I published an analysis of William Bengen’s groundbreaking retirement sustainability work–“Where’d the 4% rule come from, anyway? (Hint: it wasn’t the Trinity Study)”–a reader named Dave offered several excellent insights that unfortunately got buried in the article’s comments. I got in touch with him to see if he minded me splitting his throughts into a new post, and he was fine with that. Thanks, Dave: I’m grateful. I think you’re doing the FI/ER community a solid by sharing your perspective.
Dave’s theme was essentially, “Retirees age 65 (forget 40-year old FI/RE acolytes) are tragically naive if they follow Bengen’s SAFEMAX of 4.15% withdrawal or Trinity’s 4.00% withdrawal and believe them to be sustainable.”
I’d qualify Dave’s statement by saying, “…if they BLINDLY follow Bengen’s SAFEMAX of 4.15% withdrawal…,” because I’ve long thought that many aspirants to FI/ER, especially those new to the concept, use the 4% rule as a surrogate for developing their own model. This is a mistake.
For instance, even Bengen said in his groundbreaking 1994 paper “Determining Withdrawal Rates Using Historical Data” that:
“…the client who retired in 1929 with $500,000 in a retirement fund saw that fund dwindle to less than $200,000 by the end of 1932. Although his withdrawals have also declined from $20,000 in 1929 to $15,300 in 1932, owing to deflation, those withdrawals now equal about 7.6 percent of his portfolio, whereas he began by withdrawing only 4 percent.”
And so while the 4% rule can be a useful benchmark, it’s demonstrably fallible. Again, don’t let it do your thinking for you.
In short, retiring at the wrong time can blow up your plan. This is an obvious point in general, but know specifically that any retiree faces two kinds of risk that a time-series metric like the 4% rule can’t protect against:
- Bull markets in bonds–whether governmental or corporate–can negate the value of a portfolio’s bond component and even render it harmful to total-portfolio returns, and
- Sequence of returns risk, in which unfortunate timing obliterates one’s savings even at the “safe withdrawal rate.”
It’s necessary for those seeking FI/ER to have a good handle on these two risks
So I now quote from Dave’s commentary. He begins with a few simple principles and then develops his argument.
Dave on bond risk:
With regard to bonds, bond prices/values are inversely related to bond yields. When bond yields decline, bond prices/values increase and vice versa. And the longer the duration of a given bond, the more severe the inverse relationship is.
For a bond mutual fund, the one metric that captures this “interest rate risk” is duration. A bond fund with a duration of 5 years (about right for an intermediate bond fund) generally means that if interest rates go up 1%, the value of the bond fund will decline by about…5%. A bond fund with a duration of 15 (typical of a long term bond fund) would decline by about 15%.
“Bull markets” in bonds are periods of time when bond yields decline. We’ve had a 30 to 35 year bond bull market as yields on bonds have fallen fairly steadily over the past 30 or so years (and they’ve been historically low for several years now). Mortgage rates in the 1980s were 15% or so; today they’re 3-4%. Same is true (trend-wise) with 10-year Treasuries, 30-year Treasuries, and corporate bonds as well.
We can’t possibly be on the edge of a bull market for bonds because there is NO WAY bond yields will drop from plus 3-5% today to minus 10-12% in the future.
Bengen/Trinity bond allocations picked up higher yielding bonds which helped provide a “safe” withdrawal rate (a bond yielding 8% is nice when you want to sustain a 4% withdrawal rate.) And…as the yields dropped, bonds increased in value. That helped Bengen/Trinity too.
What are retirees/FI-RE people using for their bond allocation today? What is the duration of their bond portfolio? What is their bond portfolio yielding? Beats me. I wonder if they know or if they are thinking it’s OK to be 100% in stocks?
Bottom line with bonds is this: going forward they are not going to perform like the intermediate or long bonds Bengen/Trinity models captured.
Dave on stocks and sequence of return risk:
Sequence of return risk is what happens when early in my “retirement” my portfolio suffers a bear market (a drop of 20% or more.) See this for historic returns: http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/histretSP.html
[Author’s note: bookmark that page. It’s a useful presentation of facts you’ll be referring to a lot, and on it is a link to an Excel spreadsheet of the same data.]
Notice that the S&P 500 did the following:
2000 = -9.03%; 2001 = -11.89%; 2002 =-21.97%
Wait. 2008 = -36.55%.
Retiring with $1,000,000 on January 1, 2000, and taking out $40,000 a year would be pretty tough to do.
So, let’s say I’ve got $1,000,000 and I take out $40,000 on day one of retirement. I’ve got $960,000 invested (75% in stocks = $720k, and 25% in bonds = $240k).
And the stock market, where I have 75% of my assets allocated, drops by the end of the year by 20% I’ll be down to $576,000. We’ll assume bonds yielded 2% for the year so the $240,000 in bonds grew to $244,800 — so there’s bear market in bonds.
Now I’m starting year two with not $1,000,000, but $820,800.
Forget inflation. I want/need $40,000 to cover my expenses. I need to draw down 4.873% ($40,000/$820,800). I take out $40,000 and start year two with $780,800 (with $585,600 in stocks and $195,200 in bonds).
This year stocks drop 25% and bonds drop 3%. At the end of year two, my stock portfolio is worth $439,200 and my bonds are worth $189,344.
I start year three of retirement with $628,544 ($471,408 in stock, $157,136 in bonds).
Forget inflation, I need $40,000 (which is 6.36%) and now my nest egg is $588,544.
So if you have a couple bad years at the beginning of “retirement,” the sequence” of returns can wipe you out. Forget that most humans can’t handle those kinds of losses. What would you or I do if we need $40,000 a year but three years into retirement, our nest egg is less than 60% of where it was when we started? Probably panic.
If we bailed out of stocks then, we’re human but we’re out of the game. The next five years the S&P 500 gains 25%, 18%, 3%, 12% and 10%. The “average annual return” for stocks is pretty decent (I’m not going to do the math here with -20, -25, 25, 18, 3, 12, and 10, but there is a sequence of return that over 10 years is 8%). But if all the good years are late in the sequence and we’ve bailed out of stock, we’re done. And if we withdraw an increasing amount, we’re at best pretty darned panicked.. If the good returns come early, no problem.
Average portfolio returns (stocks average 8% a year for the past 80 years) are of absolutely no value to retirees and FI-RE when you are trying for sustainable withdrawals. It’s the sequence (the order) of returns that make all the difference.
Sequence of return for a 65-year old is one thing. For a 35-year old who is in FI-RE, it’s an entirely different kettle of fish.
This is essentially longevity risk…the risk of living too long.
Bengen/Trinity were assuming a 30-year period for retirement. That’s probably too short for 65-year olds today. Bengen and Trinity didn’t study 40 or 50 years of sustainable withdrawal rates.
As for inflation, if one needs 2, 3, or 4% increase compounded annual in their withdrawal, the $40,000 becomes:
year 2 = $40,800
year 3 = $41,616
year 4 = $42,448
year 5 = $43,297
year 6 = $44,163
year 7 = $45,046
year 8 = $45,947
year 9 = $46,865
year 10 = $47,803.
In 20 years at 2%, the initial $40,000 withdrawal would be $59,437; 30 years it’s $72,454; in 30 years and 40 years on it’s $88,321.
Today’s 40-year old person on FI-RE needs to think about 30 years on at their age 70. Anyone think a million-dollar portfolio will sustain a 2% inflation withdrawal starting with $40,000 (4%) withdrawal?
I do not.
We’ve had 8 years of a bull stock market with very, very little price inflation. These conditions will not last forever. Stocks (and bonds) have bear markets. Inflation will likely return, at some point.
In a second comment, Dave put these insights into an up-to-the-minute context.
The last five years in the 1990s generated historic returns:
$300,000 on January 1, 1995 invested in the S&P 500 with no additional investments, would end 1999 being worth $1,042,069.
Someone hoping to “retire” on January 1, 2000 with $1,000,000 withdrawing $40,000 on that date and thinking they could increase their withdrawal on subsequent January 1st by 2% was in for a surprise.
The first three years of the ’00s were negative returns of -9.03%, -11.85%, and -21.97%.
By the end of 2002, the million-dollar nest egg (after withdrawals of $40,000; $40,800; and $41,616) stood at $540,157. It would take superhuman efforts to continue taking ever-increasing (by 2%) withdrawals from a nest egg that was just about cut in half in three years.
The next five years were all positive for the S&P 500 and by the end of 2007, even taking 2% inflation-adjusted withdrawals) the nest egg was up to $683,253. Not great but perhaps encouraging.
Then came the Great Recession and the S&P 500 declined by 36.55%, Withdrawing $46,866 at the beginning of that year, the bear market resulted in the nest egg shrinking to $403,788 by the end of the year.
The next year’s withdrawal would have been $47,804 on a nest egg worth a bit more than $400k.
Fast forward to 2016.
The 2% inflation-adjusted withdrawal on January 1, 2016 was $54,911 and by December 2016, the nest egg was $444,457. Eight years of positive returns did wonders.
[Author’s note: a personal anecdote in Dave’s favor is this. My portfolio survived the Great Recession intact, in large part because I’d harvested some tax losses shortly before the crash, luckily raising enough cash to where I didn’t have to sell any of our investments for a couple of years. This kept our losses on paper through the worst of things. Note that I used the word “luckily.” And although I came close, I didn’t panic-sell.]
Two bear markets — the first protracted over three years, the other rather more quickly — coupled with increasing the withdrawal rate by 2% would have been problematic for “retirees”. Withdrawing $55k on a nest egg of $450k is about a 12% withdrawal. Is a 12% withdrawal rate sustainable?
Withdrawing a level $40k (instead of taking a 2% inflation adjusted withdrawal) helps quite a bit even with this extremely adverse stock market environment as the ending balance would be slightly more than $665k at the end of 2016.
Here is another “rear-view mirror” look at sequence of return risk withdrawing a level $40k from a million dollar nest egg starting ten years ago on January 1, 2007.
The second year of this “retirement” has the nest egg getting hit by a bear market (down 36.55%). The eight subsequent years have ALL been positive and the next egg is valued at $1,217,188 at the end of 2016.
Flipping the returns around (so the bear hits in the ninth year of the ten-year period instead of the second)–so this lucky retiree has 8 years of positive stock market gains–ends better with the nest egg valued at $1,442,583. Of course, and the end of the eighth year it had ballooned to $2,258,497.
Bengen and Trinity were important academic works that were insightful GENERALLY, but very, very dangerous for individuals to implement PERSONALLY.
The safe/sustainable withdrawal concept is reasonable in theory but by now, everyone should know that they used historical data that no longer applicable. Said another way, they don’t work today. At all.
A flexible approach in terms of income needed as well as time segmentation and other strategies help. But it is sheer fantasy to believe that a 4% withdrawal rate (forget a 4% withdrawal that increases x% a year) is sustainable.
It is extremely difficult to “keep calm and carry on” when bear markets hit. Which they always have and always will. No one, knows when, why, how big, or how long.
So there you have it.
Again, I’d like to thank Dave for taking the time to share his insights. Any thoughts? Do you agree or disagree? Please leave a comment below.