It looks like you're new here. If you want to get involved, click one of these buttons!
Bengen says based on the current environment he thinks a new retiree should be safe if they start with a withdrawal rate of…no more than 5%.
“That’s what I use myself,” Bengen told me when we spoke by phone.
....retirees right now have one saving grace: Very low inflation.
© 2015 Mutual Fund Observer. All rights reserved.
© 2015 Mutual Fund Observer. All rights reserved. Powered by Vanilla
Comments
From 1992 to 2005-6 inflation(link) was around 2.5-3%. In the last several years it's about 2%. When the long term performance of the SP500 was about 10%, Bengen used 4%. Now he says SP500 performance will be around 7% and inflation is lower (at 2%), the rule should be under 4%.
In the past it was 10-3 = 7% performance(after inflation) and now it's lower 7 - 2 = only 5%. Lower performance means lower withdrawal rate.
Our portfolio withdrawal rate would be under 2% long term. This will keep our current standard of living and our portfolio for the next 4-5 decades similar to today.
Michael Kitces is my favorite writer: a better choice is to start with lower % in stocks in early retirement years and increase the % with age.
BTW, The withdrawal rule has nothing to do with distributions. Higher distributions isn't a guarantee for better performance or volatility but it's being promoted for years as the best solution.
As I've posted before, this work by Pfau and Kitces work breaks down when rates are low. Dr. Pfau acknowledged this, writing that Kitces, incorporating CAPE P/E 10 data, concluded that the safe withdrawal rate is never less than 4.5%, and can be increased if the ratio at the start of retirement is under 20.
The only enhancement that Bengen made to Kitces' work was to incorporate inflation, i.e. part of what you are concerned about. Now he says SP500 performance will be around 7%.
You may have misread Marketwatch's writing: "Historically, he says, the average safe withdrawal rate has turned out to be about 7%." Bengen doesn't make market predictions. Also on point regarding predictions, he writes: "if you have strong feelings that the inflation regime will change in the near future, you can choose another [presumably more conservative] chart".
Thanks to @bee for having posted Bengen's article yesterday, so that one could read what he actually wrote.
https://mutualfundobserver.com/discuss/discussion/57156/william-bengen-revisits-the-safe-withdrawal-rate-at-retirement
I wonder if he still feels this way. I could check. The link is from spring of 2008, a wonderful time to be writing about anything financial, and the p/e he cites has not been <20 since like ~1993 except for that sharp 08-09 dip, and much if not most of the time it's been way >25 if not >30.
So like most (esp those of us out of equities) I am hoping this time, meaning since the 1980s, it's different.
https://www.multpl.com/shiller-pe
No need to wonder. Just check whether any of the following additional periods would have failed with a 4.5% withdrawal rate, inflation adjusted:
1979-2008, 1980-2009, 1981-2010, 1982-2011, 1983-2012, 1984-2013,
1985-2014, 1986-2015, 1987-2016, 1988-2017, 1989-2018, 1990-2019.
Portfolio Visualizer can give you a good sense on the second row (its data starts with 1985). I don't believe there was a bond index fund around then, but FBNDX can serve as a proxy.
Here's 1985-2014 to start with. As you can see, the mid 80s were a good time to start, even with the 1987 crash.
As far as the starting in one of the previous six years (1979-1984) is concerned, you can find the figures for those years here.
For 1979, $5K stocks rise to $5,926. $5K in bonds, worst case (using the worst performing bonds in the table) fall to $4,899.50. Total is $10,826. WIthdraw $450, leaving $10,376. Rebalance into stocks and bonds: $5,188 each. Adjust next year's withdrawal for inflation: $500.63.
For 1980, stocks rise to $6,834. Bonds, worst case, drop 3.32%, to $5,016. Total is $11,850. Withdraw $500, leaving $11,350. Rebalance into stocks and bonds: $5,675 each. Adjust next year's withdrawal for inflation: $568.47.
For 1981, stocks fall to $5,408. Bonds, worst case, rise to $6,140. Total is $11,548. Withdraw $568, leaving $10,980. Rebalance into stocks and bonds: $5,490. Adjust next year's withdrawal for inflation: $627.19.
For 1982, stocks rise to $6,611. Bonds, worst case (3 mo. T-bills), rise to $6,072. (T bonds and Baa bonds rose around 30%) Total is $12,683. Withdraw $627, leaving $12,056. Rebalance into stocks and bonds: $6,028. Adjust next year's withdrawal for inflation: $665.64.
For 1983, stocks rise to $7,375. Bonds, worst case, rise to $6,221. Total is $13,596. WIthdraw $666, leaving $12,930. Rebalance into stocks and bonds: $6,465. Adjust next year's withdrawal for inflation: $687.01.
For 1984, stocks rise to $6,863. Bonds worst case (3 mo T-bills), rise to $7,080 (T bonds and Baa bonds rose around 14%) Total is $13,943. Withdraw $687, leaving $13,260. Adjust next year's withdrawal for inflation: $716.55.
At this point, we can use PV, with a starting balance of $13,260 and a starting withdrawal of $717. It goes from Jan 1985 through Dec 2008. The early 2000s are not good, and 2008 is not good, but by then the mountain of cash has risen so high that it hardly matters.
That takes care of showing that 4.5% works for 1979-2008. The five other 30 year periods are left as an exercise for the reader. Also, make sure to check my arithmetic, I wasn't too diligent here.
of course; whoever has said otherwise?
>> the mid 80s were a good time to start
of course, the best ! Thank goodness.
My curiosity, as I said, is about what the scenario might look like ...
>> [4.5%] can be increased if the ratio at the start of retirement is under 20.
... when ratios are ~25%-50% and more above 20. So I will monitor PV going forward to get a sense. Are you thinking Kitces et alia would maintain their 4.5% SWR view starting now?
Obviously anyone who thought otherwise in March has been shown the wisdom of staying the course and the foolishness of doing the opposite.
One's "feeling" about immutable historical numbers does not change. Though one can add more historical data as time passes. What I illustrated is that today, all the historical data, including the additional dozen data points since Kitces' piece, (probably) does not change the Kitces' result. Of course all the additional data since Bengen's original work does not change his conclusions, as he reiterated them (with refinements) in his current work.
>>My curiosity, as I said, is about what the scenario might look like
That's a different question. You're asking what they would speculate about future data. I addressed that in writing "Bengen doesn't make market predictions."
The curiosity is understandable. The closest you're going to come to an answer is Bengen's observation: "Unfortunately, as Michael observed in his 2008 article, the “CAPE needle” has been jammed against the upper valuation stops for almost all of the last 25 years. As a result, almost the only choice for safe withdrawal rates has been the highest CAPE value in each table."
That means that there are now a few 30 year spans that started with high CAPE ratios. Obviously not enough for Bengen to break out into a separate (higher CAPE) bucket, else he would have done so in his current paper. You can hope that he revisits his partitioning of CAPE ranges in a few years when he has more high CAPE data points to work with. Though as I've tried to show, the 4.5% withdrawal rate still works with the first few periods that have rolled in since Kitces' paper.
I expect the 4.5% withdrawal rate to succeed with the next data point (1991-2020) as well. PV shows that after 29 years (1991-2019) one would be left with 4.2x one's starting value.. For the annual inflation-adjusted withdrawal at year end (Dec 31, 2020) to exhaust that portfolio would require an incredible market swoon in the last two months of the year.
>> You can hope that he revisits his partitioning of CAPE ranges in a few years when he has more high CAPE data points to work with.
Yup. I wonder what the wisest advisers are telling clients now and for the last several years of CAPE needle jamming. 'Sure, go ahead and plan w 5%? 4.5%' -- ?
>> One's "feeling" about immutable historical numbers does not change.
Har, not the case, or not invariably, with historians I read, including science historians.
On the other hand, yesterday in the history course I'm taking, we discussed how various reports and analyses on the Rwandan refugee crisis of the mid 90s offer not only different (often diametrically opposed) perspectives, but radically different figures. These numbers are not immutable.
For my portfolio sustainability I always add inflation. The last time CAPE was over 30 was in 01/1998. I'm trying to be fair and not start at much higher CAPE such as 01/2000.
PV(link) shows that 4.5%(withdrawal)+ 2.5%(inflation) = 7% withdrawal in PV isn't good enough. I know, it's not 30 years but almost 23 years is still a good one.
It's worse now because bonds future returns will be worse in the next 30 years.
Also, the idea is to reproduce Bengen's scheme, not introduce one of your own design. Bengen starts with a fixed amount (4.5% of the initial pot), and adjust that dollar amount annually by the actual rate of inflation. What you did instead was withdraw 7% of the portfolio value at the end of each year.
From Bengen's original paper:https://www.retailinvestor.org/pdf/Bengen1.pdf
Here's Bengen's model in PV applied to the time frame you selected.
At the end of the 22+ years, it shows a remaining portfolio value of $1.366M, or $847K in inflation adjusted dollars (check the inflation-adjusted box at the bottom of the graph). It's hard to see this investor's portfolio going down to zero in the next seven years, given that it's only dropped 15% in real value over the first 22 years.
FWIW, the annualized inflation rate for the years 1998-2019 (based on the NYU/Stern figures I previously cited) is 2.16%, and the current year's inflation rate is even lower.
Remaining portfolio:...Mine=$876K...yours=$847K.
From this point it may get much harder not easier. FBNDX(bonds) past performance since 1998 was 5% annually, it's going to be about 2% lower. This means that the same portfolio could make about 2% lower than the above and further erosion.
My portfolio like many others must be able to pay for LTC too.
I know, that's beyond the scope, but in my case I don't want my portfolio to lose purchasing power and the ability to handle unexpected health issues and why Bengen (and other) papers are just a starting point for me.
Remaining portfolio:...Mine=$876K...yours=$847K.
I wrote: "At the end of the 22+ years, it [Bengen's model, not mine] shows a remaining portfolio value of $1.366M, or $847K in inflation adjusted dollars (check the inflation-adjusted box at the bottom of the graph).
Your scheme: $543K in inflation adjusted dollars (check the inflation adjusted box at the bottom of the graph).
From this point it may get much harder not easier. FBNDX(bonds) past performance since 1998 was 5% annually, it's going to be about 2% lower. This means that the same portfolio could make about 2% lower than the above and further erosion
All talk, no numbers. Here's a start.
Take 50% US Stocks, 50% US Bonds. Start with a nominal value of $1.366M, i.e. Bengen's value as of now. Use your 3% nominal return figure for bonds. Use your 7% nominal return figure ("now it's lower 7 - 2 = only 5%") for stocks.
As a starting withdrawal, take Bengen's 4.5% figure ($45K on $1M), adjusted upward for inflation over the past 22 years (about 60%, as I've already shown), i.e. start with a withdrawal amount of $7200. Follow Bengen's scheme, not yours - adjust that withdrawal amount for inflation annually (based on historical inflation, which is likely too high, but that just makes this model more conservative).
Run a simulation based on these parameters for 10 years - a higher bar than the 8+years needed to complete the 30 year span. The survival rate is 1000/1000.
I'm done. You're writing about, to use @davidrmoran's term, what you "feel". I'm writing about numbers.
Inflation Adjusted (Final Balance) VFINX / FBNDX is $846,764 and VWINX is $1,098,373.
Comparing Withdrawals & Performance - VWINX to a 50/50 allocation
Unlike a portfolio that one invests in for a decade or more without touching, these calculations incorporate periodic withdrawals. So for this type of investing, volatility does matter. Much of the outperformance comes from Wellesley's extra stability. Remove the withdrawals and it still does better, but not by nearly as much.
My assumptions of 2.5% and why I used 7% are close to yours of 4.5% withdrawal and then using adjusted for inflation. Both are close...Mine=$876K...yours=$847K.
I ONLY CARE about numbers adjusted for inflation.
Basically in the last 22 years this portfolio lost a purchasing power at about 0.5% annually.
In the next 30 years, I think it will get even worse. There is a good chance to lose at least 1.5-2% annually after inflation. I used Savings Withdrawal Calculator, starting with 1million, taking out $15K annually for 30 years left me with $550K.
Still OK according to Bengen.
Unlike a portfolio that one invests in for a decade or more without touching, these calculations incorporate periodic withdrawals. So for this type of investing, volatility does matter. Much of the outperformance comes from Wellesley's extra stability. Remove the withdrawals and it still does better, but not by nearly as much.
Try this: Run a portfolio of pure cash (CASHX), no withdrawals. I've even set it up for you. I hope you'll agree that cash lost value to inflation over the past 22 years. Prove it. What's the inflation adjusted ending value of a $1M starting portfolio? How did you adjust for inflation, and did you do the same thing with your $876K amount?
Alternatively, you could just take your $876K and divide by 1.60, which is roughly the cumulative inflation between 1998 and now. That comes out to $548K. (The difference between this and $543K is likely due to the fact that 1.6 represents inflation until 2020. Add another 1% inflation for the first three quarters of the year and you're down to around $543K. Though the figures are close enough I really don't care about the cause of the 1% discrepancy.)
I didn't know 2 things:
1) that PV has Inflation adjusted box under the chart.
2) When you select Withdrawal fix Amount then another window open and you can select Inflation adjusted
So instead, I selected a fix Withdrawal Percentage of 4.5% + added an estimate 2.5% for inflation. There is no need to adjust again for inflation.
My results, which are not accurate as yours came up pretty close after 22 years.
That is Bengen's word, only about inflation. (From your above quote from his nice article [p3], entire thing starting here:
https://www.fa-mag.com/news/choosing-the-highest--safe--withdrawal-rate-at-retirement-57731.html?section=308&page=1)
I was wondering only whether his conclusion from the 2008 article might be different in such a high-CAPE era, whether he would so conjecture even though as you note he does not do conjecture.
The ending size of your portfolio is $876K nominal, $543K real. The size of the portfolio resulting from Bengen's scheme is $1.366M nominal, $847K real. These two portfolios are not close in value.
You designed a scheme radically different from Bengen's. Bengen assumed that one would withdraw a constant amount of money each year, in real dollars. Your scheme withdraws an amount each year that fluctuates based on the value of the portfolio.
From the way you describe your design, 4.5% + 2% for inflation, it seems that you still think that Bengen's scheme is to withdraw 4.5% of the value at the end of each year after adjusting for inflation. This is the miscommunication.
Bengen wrote (same quote as before): "After the first year, the withdrawal rate is no longer used for computing the amount withdrawn; that will be computed instead from last year's withdrawal, plus an inflation factor." In contrast, each year you use a withdrawal rate (7%) to compute the amount withdrawn.
You direct PV to withdraw 7% of the portfolio each year, leaving 93% to grow (or shrink) over the following year. So the withdrawals can never exhaust the portfolio. However, as the portfolio shrinks in size, so will the size of the withdrawal.
----------
For example, suppose that the portfolio is invested in something that loses 10% of its value, no more, no less, each year. Then under your scheme, a portfolio that started with $1M would progress as follows:
Start: $1,000,000.
Year 1 end: $900,000.
Withdraw $63,000 (7%).
Year 1 after withdrawal: $837,000.
Year 2 end: $753,300.
WIthdraw $52,731 (7%) - notice that the size of the withdrawal is shrinking.
Year 2 after withdrawal: $700,569.
...
Year 22 end: $21,452.46
Year 22 withdrawal $1,501.67
Year22 after withdrawal: $19,950.79
I find it instructive to do calculations by hand, but in case you don't believe me, here's PV's calculation. If you mouse over the graph to year 22, you'll see that the value is $19,951 (before adjusting for inflation).
----------------------
Bengen's scheme (assume 0 inflation for simplicity, higher inflation would merely make the results worse):
Start: $1,000,000
Year 1 end: $900,000
Withdraw: $45,000
Year 1 after withdrawal: $855,000
Year 2 end: $759,500.
Withdraw $45,000 - or more if there is inflation
Year 2 after withdrawal: $724,500.
...
Year 11 end: $50,025.36
Withdraw $45,000
Year 11 after withdrawal: $5,025.36
Year 12 end: $4,522.83
Not enough to withdraw $45,000. Portfolio is exhausted.
Again, I think that the figures above are more instructive than blindly using a calculator. But here's PV's calculation. Mouse over year 11, and you'll find the value $5,025 (before adjusting for inflation). Obviously PV reports that the portfolio is exhausted in year 12.
https://www.morningstar.com/articles/1004651/the-retirement-income-puzzle
Here's his table of a million dollar portfolio over five years where the nominal growth rate is 4.1% (I used -10%), and inflation rate of 2% (I used 0%). Note that even though he says that the withdrawal "rate" is 4%, what he is actually doing is what Bengen described: starting with 4% ($40K on $1M), he adjusts this fixed amount for 2% inflation annually. He is not withdrawing 4% of the balance annually. Or worse, 4% of the balance plus an additional 2% of the balance purportedly to "adjust" for inflation.
Most of us will be withdrawing from tax deferred accounts so keep taxes in mind. Paying taxes will also be part of the withdrawal calculation...withdrawal need + taxes.
After age 70.5 RMD required mandatory distributions kick in. These distributions don't necessarily need to be spent, but will form a required part of the distribution strategy.
Derf
One withdraws 4% of the original portfolio (here 4% of $1M, or $40K), and then adjusts that fixed amount for inflation. What percentage of the portfolio balance that withdrawal represents each year is not fixed, but depends on the real return of the portfolio.
For example, if the portfolio doubles in the first year, i.e. grows by 104% (100% after adjusting for 2% inflation), then the $40,800 would constitute only 2% ($40,800/$2,040,000) of the year end balance.
As to what return rate is needed assuming a constant rate of return and a constant inflation rate, Rekenthaler wrote: So a nominal rate of return of 6% or better means that the portfolio will last forever.
A nominal return rate of less than 6%, e.g. 4.1%, would lead to the portfolio ultimately being exhausted. But for this portfolio, that would still take 36 years. Here's a PV calculation illustrating this, where all figures are in terms of real dollars. One starts with $1M and $40K withdrawal (both real), the portfolio grows at a 2.1% real rate of return (4.1% - 2% inflation), and the withdrawal amount is a constant $40K (real).
The assumptions are simplistic. Rates aren't constant. So matters are more complex. Still, Bengen has not found any historical 30 year period, including ones with low inflation and richly priced markets, where a starting withdrawal rate of 4.5% did not survive 30 years.
People can always say "this time is different". Perhaps that is the only thing that isn't different. People are always saying that "this time is different."
ah
>> Bengen has not found any historical 30 year period, including ones with low inflation and richly priced markets, where a starting withdrawal rate of 4.5% did not survive 30 years.
amazing to contemplate with big drops in a given year
Someone born between Jan and June in 1951 would have been required to start their RMDs in 2021 (age 70.5). Now they can start in 2023 (age 72). Two years grace. Same two year grace for anyone born between Jan and June in years after 1951.
Someone born between July and Dec in 1950 would have been required to start their RMDs in 2021 (age 71). Now they can start in 2022 (age 72). That's only one year's grace. Same one year grace for anyone born between July and Dec in years after 1950.
Someone born between Jan and June 1950 would have been required to start their RMDs in 2020. Now they can start in 2022 (age 72). That's still two extra years, though one of those is coming from the fact that no one has to take RMDs this year.
Someone born between July and Dec 1949 would have been required to start their RMDs in 2020 (age 71). Now they can start in 2021 (age 72). So they're also getting one year of grace, though that is coming from the 2020 waiver, even without the age extension.
Someone born before July 1949 gets a year of grace (2020) not because of the change in RMD age but because everyone is excused this year. So they're also getting one year of grace.
In short, these "oldsters" get a one year break. That's the same amount as those born between July and December of any year from 1949 on, and just one year less than the two years anyone born between Jan and June from 1950 on get.
People born late 1949 are the ones who should be complaining. They're covered by the extension to age 72, but that doesn't get them even a single year extension.
https://humbledollar.com/2020/11/rate-debate/
cape ratio this weekend just under 34