The Current Annuity Factor & Dangers of the 4% "Rule"

[kevink]

I thought this article on Bloomberg by one of the architects of DFA's retirement income funds was well worth the read:

https://www.bloomberg.com/opinion/articles/2021-05-14/retirement-savings-may-not-go-as-far-as-you-think

For those contemplating (or already in) ER it can be easy to miss the fact that the recovery in equities since last March's deep dive has been more than offset by declining interest rates on the bond side for any remotely balanced portfolio - and safe withdrawal rates going forward will have to be much lower (and/or one has to accumulate a much larger nest egg).

I learned of this piece in a long (and worthwhile IMHO) thread on Bogleheads started by a guy who knows the world of pensions and annuities inside and out. If nothing else, this particular post by him midway through the first page of the discussion is a tour-de-force deconstruction of the 4% "rule" (or guideline, which is of course all the Trinity study authors intended it to be) as a tool for use in the real world - along with great links to much better tools:

https://www.bogleheads.org/forum/viewtopic.php?p=6039678#p6039678

You know that things have gotten truly weird when even the Bogleheads crowd is talking openly about options like buying single-premium immediate annuities and iBonds and dabbling in gold ETFs. Such is the new negative real return bond world we live in, and withdrawal methods based on very different assumptions (like the old 50:50 used by Bengen et al) could quite possibly end up being dangerous relics of the past.

 

[VanWinkle]

It's not different this time, bonds will be back in the future and the 4% rule also realized some low bond yields in the past. Sorry for the optimism.

 

[RetireeRobert]

It's not different this time, bonds will be back in the future ....

+1 With all the printing press trillions being passed out, with all the trillions of dollars spending plans in the works, with inflation in the real world showing disturbing signs of resurgence, it seems only a matter of time before the Fed must relent and "let" interest raise rise. Perhaps the Fed will have no choice, and no tools left to keep rates artificially low. Bonds will eventually have to pay more.

 

[Midpack]

It's not different this time, bonds will be back in the future and the 4% rule also realized some low bond yields in the past. Sorry for the optimism.

+1. When I read this in the article, it suggests the author had no clue what underlies the 4% rule. So in essence she’s saying 3% is the new 4%, that’s not a new idea. OTOH, there’s some good stuff in the bogleheads thread as usual thanks.

First of all, you need to account for risk. People nearing retirement typically have about 40% of their retirement savings in the stock market. That’s great when the market rises, but you’ll lose more when it falls. Odds are the stock market will have a significant decline at some point during the 20 or 30 years you are retired — maybe more than once.

Duh. Like the market hasn’t endured “significant declines” several times during the period the 4% rule was studied.

 

[ERD50]

...
For those contemplating (or already in) ER it can be easy to miss the fact that the recovery in equities since last March's deep dive has been more than offset by declining interest rates on the bond side for any remotely balanced portfolio - ...

Unless I made a mistake in my data entry here (so check it), I call BS.

https://bit.ly/3z7rtWc

That shows that a 60/40 portfolio, since Jan 2020 (so we see the march dip) with a 3.6% annual withdraw (taken monthly since we have a short time period here) is up a healthy 15% *after* withdrawals. A 100% bond fund was down a hair in total value, but that's not balanced.

-ERD50

 

[CardsFan]

Unless I made a mistake in my data entry here (so check it), I call BS.

https://bit.ly/3z7rtWc

That shows that a 60/40 portfolio, since Jan 2020 (so we see the march dip) with a 3.6% annual withdraw (taken monthly since we have a short time period here) is up a healthy 15% *after* withdrawals. A 100% bond fund was down a hair in total value, but that's not balanced.

-ERD50

Based on my first hand experience, this makes way more sense.

 

[kevink]

Unless I made a mistake in my data entry here (so check it), I call BS.

https://bit.ly/3z7rtWc

That shows that a 60/40 portfolio, since Jan 2020 (so we see the march dip) with a 3.6% annual withdraw (taken monthly since we have a short time period here) is up a healthy 15% *after* withdrawals. A 100% bond fund was down a hair in total value, but that's not balanced.

-ERD50

Yeah, mea culpa on misreading that part of the Bogleheads posts. Here's the applicable part:

"You might have encountered a guideline that you can spend 4% of your savings each year. But really the amount you can spend is dependent on interest rates and inflation, and at the current low rates that protect you from inflation, which have even dipped negative, your money doesn’t go as far. ...

There are few hard and fast rules when it comes to retirement finance, but one is this: if you want to see how much you can spend each year, divide your wealth by 32. That’s the current annuity factor, a number based on inflation-adjusted interest rates that helps you figure out how much $1 of spending will cost over a prospective 30 years of retirement.

Real interest rates, which protect you from inflation, have fallen in recent years, making it more expensive to finance future spending — at today’s negative rates, ensuring you have $1 to spend in 30 years costs $1.01 today. That change in real interest rates has also driven up the annuity factor by 13% since January 2020.

Think about it his way: Suppose you had $500,000 in January 2020 when the annuity factor was 28. That meant you could afford to spend $17,500 a year in retirement. By May 2021, if you invested all your money in the stock market, you’d have $646,000, a 30% increase. But your spending capacity only increased 14%, to $20,000 because the annuity factor jumped to 32."

So the real point is that falling interest rates have muted the positive effects of the recovery in equity, but things are still up, as you said. Sorry for the error.

 

[kevink]

As a P.S. this excerpt from one of the most recent posts by the OP of that thread spells things out quite clearly:

"Returns

If 10-year bond yields rise in the next 10 years the returns will remain extremely low because the bond prices will be falling. To put this in perspective, data from “Triumph of the Optimists” (and updates to the book’s data) have LT US Treasure real returns (they use 20 year maturities) for the 50 years from 1960-2010 of roughly 3% per year on average – not zero.

From my 2015 SBBI Yearbook I have 20-year Treasury returns from 1926-2014. For those LT US Treasuries the nominal average return from 1926-2014 was 5.7%. For 5-year US Treasuries the nominal average return from 1926-2014 was 5.3%. From 1926-2014 the average inflation rate was 2.9%. That makes the average real return from 1926-2014 for 20-year Treasuries approximately 2.8% and for 5-year Treasuries 2.4%. The current situation is far removed from those long-term averages.

Typical historical bond returns and today's returns and how they affect the retirees' portfolio

Below are some fairly typical historical returns.
Real return 10-year Treasuries 2.7%
Real return US stocks 6.0%
Real return international stocks (global ex US) 5.0%

Let’s give stocks the higher 6% US stock returns.

The 70 year old retiree’s portfolio is 1/3 equity and 2/3’s bonds
Portfolio return from stocks 2%
Portfolio return from bonds 1.8%
Portfolio return 3.8%

Now under current conditions
Portfolio return from stocks 2%
Portfolio return from bonds 0%
Portfolio return 2%"

Admittedly 1/3 equity, 2/3 bonds is a very conservative (Wellesley, Vanguard Target Retirement Income) allocation but upping the equities to 50 or 60% doesn't move the needle all that much - while it does introduce substantial drawdown and sequence-of-returns risks.

 

[GravitySucks]

Gloom and Doom.
Do you believe now is worse than any point over the last 100+ years? If you do cut your spending.

I don't think we're in a bad time for investors.

 

[Dtail]

As been mentioned before, the 4% "rule" is used rarely by anyone as a retirement withdrawal methodology. It is more useful as a guidance as to whether one is reasonably in good shape to retire.
The assumption that bond yields will always be low, is not a good assumption.
How many pundits have been calling for 2 or 3% WR for the last 5 years?

 

[ERD50]

Yeah, mea culpa on misreading that part of the Bogleheads posts. Here's the applicable part:

"You might have encountered a guideline that you can spend 4% of your savings each year. But really ....

Think about it his way: Suppose you had $500,000 in January 2020 when the annuity factor was 28. That meant you could afford to spend $17,500 a year in retirement. By May 2021, if you invested all your money in the stock market, you’d have $646,000, a 30% increase. But your spending capacity only increased 14%, to $20,000 because the annuity factor jumped to 32."

So the real point is that falling interest rates have muted the positive effects of the recovery in equity, but things are still up, as you said. Sorry for the error.

No problem, it's a useful discussion.

Now I can see there may be some real value in looking at annuity factors as a metric. But the example given seems to be off-base to me:

So I'll accept that metric and the statement that the spending capacity only increased 14% from Jan 2020 to May 2021. But, with the traditional "SWR", you only increase your spending by inflation. And that was just 3.52% (according to this calculator - latest data for April 2021).

https://www.bls.gov/data/inflation_calculator.htm

So it seems the traditional SWR'er is still in good shape, their spending is not outpacing their portfolio growth at all.

All this talk has me wanting a really good cup of coffee, but alas, we've moved and I have not replenished my stock of whole roasted beans! (For those unaware, kevink is the forum resident coffee expert, with credentials and everything, and has been a wealth of knowledge in the coffee arena! )

-ERD50

 

[kevink]

No problem, it's a useful discussion.

Now I can see there may be some real value in looking at annuity factors as a metric. But the example given seems to be off-base to me:

So I'll accept that metric and the statement that the spending capacity only increased 14% from Jan 2020 to May 2021. But, with the traditional "SWR", you only increase your spending by inflation. And that was just 3.52% (according to this calculator - latest data for April 2021).

https://www.bls.gov/data/inflation_calculator.htm

So it seems the traditional SWR'er is still in good shape, their spending is not outpacing their portfolio growth at all.

All this talk has me wanting a really good cup of coffee, but alas, we've moved and I have not replenished my stock of whole roasted beans! (For those unaware, kevink is the forum resident coffee expert, with credentials and everything, and has been a wealth of knowledge in the coffee arena! )

-ERD50

Yeah, and I should probably stick to coffee-related posts.

I don't buy the sky-is-falling/interest-rates-will-stay-low-forever mentality either but I did get a lot out of the Bogleheads discussion of the Bloomberg article (much more so than the article itself). And while you're of course correct that many who post here are far too savvy to use some version of the 4% rule, I run into a lot of younger folks aiming for super-early retirement (30's to maybe early 40's) who still use it for planning purposes. Surely it's better for planning purposes to use admittedly pessimistic projections like ones I shared (e.g. 2-3% real from equities, nothing from bonds) rather than relying on backtesting of data that's largely driven by bull markets in both bonds and equities that seem unlikely to recur.

 

[SecondCor521]

Surely it's better for planning purposes to use admittedly pessimistic projections like ones I shared (e.g. 2-3% real from equities, nothing from bonds) rather than relying on backtesting of data that's largely driven by bull markets in both bonds and equities that seem unlikely to recur.

Emphasis added.

Better depends on the relative value or cost of making two different kinds of errors: FIREing too early and FIREing too late. The first is the one that most people focus on "Gosh I don't want to run out of money and have to go back to work", but the latter is a real risk as well, and something that a young person may choose to weigh heavily as well: "Man, I'd hate to waste my youth at a job I dislike when I don't have to".

 

[Dtail]

Yeah, and I should probably stick to coffee-related posts.

I don't buy the sky-is-falling/interest-rates-will-stay-low-forever mentality either but I did get a lot out of the Bogleheads discussion of the Bloomberg article (much more so than the article itself). And while you're of course correct that many who post here are far too savvy to use some version of the 4% rule, I run into a lot of younger folks aiming for super-early retirement (30's to maybe early 40's) who still use it for planning purposes. Surely it's better for planning purposes to use admittedly pessimistic projections like ones I shared (e.g. 2-3% real from equities, nothing from bonds) rather than relying on backtesting of data that's largely driven by bull markets in both bonds and equities that seem unlikely to recur.

The back testing of data using historical sequencing of data is driven by bull and bear markets.
The 2000 retiree who has endured 3 bear markets already is still in good shape if they were using the 4% rule as a withdrawal methodology in theory.
The 4% guidance represents the worst sequencing results in history, not an average of results as one example.
The 1929 retiree also survived in theory.

 

[kevink]

Actually the 2000 retiree isn't doing all that well, in most cases. Scroll down to the "if you retired in 2000" section in this post by John Greaney:

https://retireearlyhomepage.com/reallife20.html

"What if you retired in January 2000?

If you happened to retire in January 2000, the last nineteen years haven't been pleasant. Only the Warren Buffett portfolio has a value appreciably exceeding its $100,000 starting balance. The 100% fixed income portfolio is underwater while the MPT portfolio, Larry Swedroe Portfolio, Harry Browne Portfolio, and Harry Dent Portfolio are all 16% to 34% in the black. The other two portfolios both show losses. The worst performer was the 75% S&P500/25% fixed income portfolio which is now closing in on one-half of its starting value . In hindsight, buying the 30-year TIPS in 2000 wouldn't have been a bad idea. You'd have a 4% inflation-adjusted yield and the $100,000 bond would be worth over $130,000 as of year end 2019 -- only Berkshire Hathaway beat that by a significant margin. The chart below illustrates this performance."

I'll also point out that the Trinity study that's the origin of the 4% rule used Ibbotson data from 1925-1995 only. Having ER'd myself in 2002 I'm just glad I wasn't all-in on the kind of portfolio they used as the basis for that study.

 

[Dtail]

Actually the 2000 retiree isn't doing all that well, in most cases. Scroll down to the "if you retired in 2000" section in this post by John Greaney:

https://retireearlyhomepage.com/reallife20.html

"What if you retired in January 2000?

If you happened to retire in January 2000, the last nineteen years haven't been pleasant. Only the Warren Buffett portfolio has a value appreciably exceeding its $100,000 starting balance. The 100% fixed income portfolio is underwater while the MPT portfolio, Larry Swedroe Portfolio, Harry Browne Portfolio, and Harry Dent Portfolio are all 16% to 34% in the black. The other two portfolios both show losses. The worst performer was the 75% S&P500/25% fixed income portfolio which is now closing in on one-half of its starting value . In hindsight, buying the 30-year TIPS in 2000 wouldn't have been a bad idea. You'd have a 4% inflation-adjusted yield and the $100,000 bond would be worth over $130,000 as of year end 2019 -- only Berkshire Hathaway beat that by a significant margin. The chart below illustrates this performance."

I'll also point out that the Trinity study that's the origin of the 4% rule used Ibbotson data from 1925-1995 only. Having ER'd myself in 2002 I'm just glad I wasn't all-in on the kind of portfolio they used as the basis for that study.

It appears the article was updated on Apr 1, 2020 and thus completely ignores the large gains in the market from that point until the current date.
Thus a misleading reference when looking at it currently.

 

[kevink]

It appears the article was updated on Apr 1, 2020 and thus completely ignores the large gains in the market from that point until the current date.
Thus a misleading reference when looking at it currently.

I don't see anything "misleading" about preferring 20 years of data to 1. The S & P gained 16% in 2020. The additional year doesn't move the needle all that much - or detract from the value of looking at actual returns for the two decades a year 2000 retiree (I ER'd myself in 2002 so I've lived most of this) actually experienced.

 

[SecondCor521]

I don't see anything "misleading" about preferring 20 years of data to 1. The S & P gained 16% in 2020. The additional year doesn't move the needle all that much - or detract from the value of looking at actual returns for the two decades a year 2000 retiree (I ER'd myself in 2002 so I've lived most of this) actually experienced.

It's not that it's 20 years of data to 1. It's that it's outdated.

The S&P total return from 4/1/20 to 6/4/21 is 71%. If your original portfolio had dropped by 50% over those 20 years as of 4/1/20, you'd be back up to 85% of your original portfolio (ignoring any withdrawals in the interim). That's still not wonderful, but it is a different situation, and I would say it does move the needle quite a bit. (If you were 75% S&P, the S&P portion of your portfolio would have recovered to almost 77% of that portion of your original portfolio. I don't know what the bond portion did during that time frame - I think it was about level.)

Data from https://finance.yahoo.com/quote/^GS...istory&frequency=1d&includeAdjustedClose=true using the adjusted close column. Percentage calculations by me.

 

[ncbill]

It's not that it's 20 years of data to 1. It's that it's outdated.

The S&P total return from 4/1/20 to 6/4/21 is 71%. If your original portfolio had dropped by 50% over those 20 years as of 4/1/20, you'd be back up to 85% of your original portfolio (ignoring any withdrawals in the interim). That's still not wonderful, but it is a different situation, and I would say it does move the needle quite a bit. (If you were 75% S&P, the S&P portion of your portfolio would have recovered to almost 77% of that portion of your original portfolio. I don't know what the bond portion did during that time frame - I think it was about level.)

Data from https://finance.yahoo.com/quote/^GS...istory&frequency=1d&includeAdjustedClose=true using the adjusted close column. Percentage calculations by me.

Yep, a 75% S&P 500 / 25% fixed income portfolio is doing just fine:

https://retireearlyhomepage.com/reallife21.html

 

[Dtail]

Yep, a 75% S&P 500 / 25% fixed income portfolio is doing just fine:

https://retireearlyhomepage.com/reallife21.html

It's not that it's 20 years of data to 1. It's that it's outdated.

The S&P total return from 4/1/20 to 6/4/21 is 71%. If your original portfolio had dropped by 50% over those 20 years as of 4/1/20, you'd be back up to 85% of your original portfolio (ignoring any withdrawals in the interim). That's still not wonderful, but it is a different situation, and I would say it does move the needle quite a bit. (If you were 75% S&P, the S&P portion of your portfolio would have recovered to almost 77% of that portion of your original portfolio. I don't know what the bond portion did during that time frame - I think it was about level.)

Data from https://finance.yahoo.com/quote/^GS...istory&frequency=1d&includeAdjustedClose=true using the adjusted close column. Percentage calculations by me.

Exactly. The S&P return for all of 2020 is not relevant, when the data cuts off right in the bear market of 2020.
Simple math.

 

[kevink]

Yep, a 75% S&P 500 / 25% fixed income portfolio is doing just fine:

https://retireearlyhomepage.com/reallife21.html

You missed the fact that the updated 2021 return article on Greaney's site begin in 1994, which is the year he retired. That skews the returns much more than using just the 2000-2020 numbers - which was what I did since the claim was that folks who retired in 2020 did just fine.

Anyway we're kind of beyond beating a dead horse here while perhaps missing the useful take-aways from both the Bloomberg article and the excellent thread it inspired on Bogleheads. These include the importance of liability matching rather than using either fixed or flexible withdrawal rates in retirement, and making sure to take into account what it truly unprecedented about the current situation: negative real returns on safe bonds of all durations.

 

[Midpack]

You missed the fact that the updated 2021 return article on Greaney's site begin in 1994, which is the year he retired. That skews the returns much more than using just the 2000-2020 numbers - which was what I did since the claim was that folks who retired in 2020 did just fine.

Anyway we're kind of beyond beating a dead horse here while perhaps missing the useful take-aways from both the Bloomberg article and the excellent thread it inspired on Bogleheads. These include the importance of liability matching rather than using either fixed or flexible withdrawal rates in retirement, and making sure to take into account what it truly unprecedented about the current situation: negative real returns on safe bonds of all durations.

There have been negative real and even nominal bond returns before, we'll see how persistent the current trend is. People have been saying bonds are bad since 2008, they may finally be right, we'll see.

https://novelinvestor.com/historical-returns/#:~:text=Asset%20Class%20Annual%20Returns%201926%20%E2%80%93%20Current%20,%20%2018.69%25%20%2068%20more%20rows%20

 

[daylatedollarshort]

Anyway we're kind of beyond beating a dead horse here while perhaps missing the useful take-aways from both the Bloomberg article and the excellent thread it inspired on Bogleheads. These include the importance of liability matching rather than using either fixed or flexible withdrawal rates in retirement, and making sure to take into account what it truly unprecedented about the current situation: negative real returns on safe bonds of all durations.

I was a fan of the liability matching poster, Bobcat2, since we FIREd and have followed his advice. We loaded up on TIPS early on, though I haven't been buying at the current low rates, which have been below the inflation rate. My main solution to negative real returns on bonds is to get our personal inflation rate well below the CPI rate by continually optimizing our expenses. Like this year we refinanced our mortgage when rates went down; we're shutting down a small business we haven't really worked on that had some lingering overhead costs; we found a free app channel that provides local programming which save $25 a month over the cable fees; DH switched to a hair stylist that is half the price of what he used to pay; and a bunch of other stuff that didn't lower our quality of life but combined shaved over $5K off our annual run rate.

 

[Dtail]

You missed the fact that the updated 2021 return article on Greaney's site begin in 1994, which is the year he retired. That skews the returns much more than using just the 2000-2020 numbers - which was what I did since the claim was that folks who retired in 2020 did just fine.

Anyway we're kind of beyond beating a dead horse here while perhaps missing the useful take-aways from both the Bloomberg article and the excellent thread it inspired on Bogleheads. These include the importance of liability matching rather than using either fixed or flexible withdrawal rates in retirement, and making sure to take into account what it truly unprecedented about the current situation: negative real returns on safe bonds of all durations.

Fair point on the 2nd paragraph.

From the article, even though it begins with the year 1994, one can easily extract the year 2000 starting point and can see that at the end of 2020, the 2000 retiree is still doing well.

 

[ncbill]

^
Yep, using the "4% rule" the 75/25 portfolio has still more than doubled (nominal) from 2000-2020.

So I don't see anything wrong with using it, primarily as a check to see if you've saved enough, i.e. 25x expenses for a 30-year retirement.

Using the annuity factor remains too conservative, IMHO...ignores equity returns.