Fallacy of the 4% Rule

[jazz4cash]

We just reviewed our Fidelity Retirement Planner results. They propose 4-5%.

So what is Bengen’s new formula? I couldn’t get through paywall.

 

[audreyh1]

Many of us use the current portfolio value to determine our withdrawals. This has the advantage of ratcheting up if the portfolio growth outpaces inflation, but also means income drops after bad market years. Still it’s better than underspending or flat running out of money.

 

[reneeh63]

Seems I remember that my spinster great-aunt (born around 1895) had her own rule which was to never delve into the principle and to keep spending to 5% or less. A bit generous compared to the 4% 'rule' but most of her living was before the terrible inflation of the 1970s as she died in the early '80s.

So maybe the earlier advice was a bit rounded off to make the rule even simpler (at 5%) and I don't know how widely followed that advice was - I suppose many who'd saved less needed to spend more than 5% out of necessity to meet their expenses.

 

[CalBird]

Only if you equate money with life.

Money is not equal to life. Money does help, but only if you spend it.

 

[Gumby]

Money is not equal to life. Money does help, but only if you spend it.

Look up the concept of "Fleet in Being". Having extra money in the bank improves my life whether I spend it or not.

 

[engineernerd]

Ha, that article gives no solid information about how Mr. Bengen defines "safe".

I will stick with FIRECalc, which lets me decide for myself how conservative to be. And lets me do all kinds of "what-if" / sensitivity analyses. And includes data covering a lot of history, making no claim to see into the future or discount past eras.

 

[pacergal]

A 30 year plan may not be long enough if you retire early in your 40-50's. We retired at age 60, so. not real early, but I have one grandparent who lived to 100. It is amazing to me the difference in my potential life span when I run longevity calculators with that 100 years included or not.
Without it, it's mid to late 80s, with it it is 97+. That is a whole decade I need to include in my planning, just in case.....

 

[Sojourner]

It seems to me that the 4% "rule" is best used as a very rough guideline for discussing retirement spending with the average, everyday, non-FIRE crowd. Us FIREd folks are much better off utilizing tools such as FIRECalc, which allows me to precisely specify the parameters of my situation (number of years retired, expected SS income, asset allocation, etc.). When I plug all of my numbers in, FIRECalc gives me a 100% historically safe WR of 3.3%, which I feel is far more reliable than any one-size-fits-all 4% rule.

 

[marko]

This is not a new topic. Bengen is looking at returns over the past 2 decades and saying “Oh, look, equities have done really well and we could have spent more”.

I agree with pb4uski, the 4% rule of thumb has held up well. It’s certainly no fallacy. Keep in mind, the primary objective is not to maximize withdrawals, it’s to determine a high level of withdrawals while maximizing survival odds of the portfolio.

The most important condition is portfolio survivalability, and sustsinsble withdrawal % follows.

The idea is not to die rich, but to avoid dying poor.

 

[Chuckanut]

My gut tells me that most folks starting out - and actually "using" the 4% rule - will figure out in a few years that they can spend more. At THAT time, they will likely begin to spend more. How much better an approach is that than starting at say 5% spend and find out it's too much. Always easier to spend more than to spend less IMHO. No rule is perfect. Let's call the 4% rule a "good" rule.

You have obviously become one of the dangerous radicals who infest this site. You make too much sense.

 

[engineernerd]

We just reviewed our Fidelity Retirement Planner results. They propose 4-5%.

So what is Bengen’s new formula? I couldn’t get through paywall.

According to the article, 4.7% (for a 30-year retirement).

Two paragraphs taken from the Marketwatch article:

"Under the original rule, he used a simple portfolio of two asset classes: U.S. large-company stocks and U.S. 5-year bonds. Over time, he built a more sophisticated and balanced portfolio by using U.S. large-cap stocks, U.S. midcap stocks, U.S. small-cap stocks, U.S. microcap stocks, international stocks, intermediate-term U.S. government bonds and U.S. Treasury bills.

That diversification lifted the 4% rule to 4.7%. He fiddled some more and found that adding even more asset classes — such as gold, commodities, real estate, and emerging-markets equities — didn’t make a big difference. In addition to diversifying, Bengen urges investors to rebalance their portfolios each year."

 

[Dtail]

There are six 30 year periods that would have failed, over the last 125 years, according to FireCalc, at 4%. I think 4% was designed to give you roughly a 95% chance of success, not 100%. To get to 100%, you need to cut back a bit, to a 3.55% withdrawal. At least, at 3.5% I get 100%. At 3.6% I get 99.2% (1 cycle failed). So I split the middle at 3.55% and got 100%, but didn't feel like fine tuning it any further than that.

I think it came to 3.58% when I tried it once. lol

 

[Dtail]

IF you think about it, 4% is just 1/25th, so over 30 years you are just hoping to make up 5 years in growth to get to your thirty years. I believe you can achieve that with less than a 2% annual growth on average.

I know the market doesn't grow consistently and there are SORR and such, but 4% is a WORST CASE SCENARIO.

I just asked Chat GPT to see what 1million would be today if I retired 30 years ago to the day in 1995, assuming actual SP500 returns and using 4% withdraws with actual CPI increases, and I would have 968,000 after 30 years of withdraws.

For me 4% is just a general thought exercise, and now Bengen says 4.7 up to 5.

Check this out >>> Is the 4% Rule Obsolete? - TheWeFIRE

Yes IIRC, the real rate of return for 30 years was between 1.0 and 1.5% to have the portfolio survive.

 

[Dtail]

I believe that in FIRECalc, 1966 was the absolute worst year to retire.

I believe that is correct. It is also currently the last starting 30 year period date to fail. So no failures in over 50 years.
Yes the jury is still out on the 2000 year beginning retirement year.

 

[MichaelB]

The idea it not to die rich, but to avoid dying poor.

Yes, and there’s really no good alternative. Even withdrawing 4%, FIRECalc shows 5% of portfolios failing, which is pretty high for a low risk plan. A withdrawal rate of 6%-7% as suggested in the linked article by Bengen completely ignores sequence of returns risk and assumes that the “bad times in the 70’s” won’t repeat.

 

[Dtail]

We just reviewed our Fidelity Retirement Planner results. They propose 4-5%.

So what is Bengen’s new formula? I couldn’t get through paywall.

Kind of interesting that Fidelity uses a 4-5% SWR, but their results on their most conservative output is more conservative than Firecalc's which is conceptually based the 4% historical sequencing.
This is most likely due to the Monte Carlo type of sequencing.

 

[Dtail]

Many of us use the current portfolio value to determine our withdrawals. This has the advantage of ratcheting up if the portfolio growth outpaces inflation, but also means income drops after bad market years. Still it’s better than underspending or flat running out of money.

Yup that's us, but will switch to the Clyatt 95/5 version within 2026 or 2027. As we discussed before, this is just a slight tweak.

 

[COcheesehead]

Kind of interesting that Fidelity uses a 4-5% SWR, but their results on their most conservative output is more conservative than Firecalc's which is conceptually based the 4% historical sequencing.
This is most likely due to the Monte Carlo type of sequencing.

I *think* Fidelity assumes a 15% haircut or about that in year 1 of the most conservative output. I’d have to look at mine again, but it used to be that way. It would mimic a bad SORR start.

 

[Dtail]

I *think* Fidelity assumes a 15% haircut or about that in year 1 of the most conservative output. I’d have to look at mine again, but it used to be that way. It would mimic a bad SORR start.

IIRC it used to be a 15% haircut in Year 1, but I believe it is less than that now, but still a haircut of sorts.
Additionally conceptually a Monte Carlo concept could put the starting years as 2008,1929,2000 etc which is not how the historical markets worked.

 

[Exit 2024]

On Bogleheads.org there are interesting thread where some senior members constructed Variable Percentage Withdrawal model, there are also spreadsheet to plug your numbers and separate thread where every month one of the authors shows how that works - search for VPW Forward test.

It gives you ability to take out larger % but also gives boundaries which you need to keep in mind if market will go sour. It take in the account expected SS( you can discount it) and any other pensions or other incomes. I believe it uses life expectancy tables and smoother out your income before and after SS/Pension start.

Example for my numbers if I retired in 2025:
$3M portfolio invested as 70/30 stock/bonds age 56 (today)
I can take out this year $171,102 (~5.7%) but should have flexibility is my budget
to scale down to $121,457 (~4%) if needed.

There are a lot of thoughts and discussions was put into the building that tool, it is free to anyone and could be used as additional modeling in case you really need more than 4% but afraid of consequences.

 

[Dtail]

On Bogleheads.org there are interesting thread where some senior members constructed Variable Percentage Withdrawal model, there are also spreadsheet to plug your numbers and separate thread where every month one of the authors shows how that works - search for VPW Forward test.

It gives you ability to take out larger % but also gives boundaries which you need to keep in mind if market will go sour. It take in the account expected SS( you can discount it) and any other pensions or other incomes. I believe it uses life expectancy tables and smoother out your income before and after SS/Pension start.

Example for my numbers if I retired in 2025:
$3M portfolio invested as 70/30 stock/bonds age 56 (today)
I can take out this year $171,102 (~5.7%) but should have flexibility is my budget
to scale down to $121,457 (~4%) if needed.

There are a lot of thoughts and discussions was put into the building that tool, it is free to anyone and could be used as additional modeling in case you really need more than 4% but afraid of consequences.

It appears to be a good model, but in the end it is effectively stating 5%+WR are fine but then using the large caveat of potentially large adjustments if the situation is not going well.
Not that any model is blindly followed.

 

[Al18]

It’s a rule of thumb, not an in-depth retirement plan. I hope anyone considering early retirement relies on an individually tailored retirement plan. It’s also advisable to revisit your plan yearly. I don’t think it’s out of line to start at 6%, and review your investment performance and adjust regularly.

 

[Koolau]

Look up the concept of "Fleet in Being". Having extra money in the bank improves my life whether I spend it or not.

Indeed! Life without money stress is SO much better!

 

[Koolau]

You have obviously become one of the dangerous radicals who infest this site. You make too much sense.

 

[SnowballCamper]

“Although you can call it a 4.7% rule for ultraconservative people — if they wanted to be the safest that’s ever been in history — but for most people they’ll end up with a lot of money and probably a lot of regrets at the end of retirement and wishing they’d spent more earlier,” Bengen said.

“You have to look at the circumstances at when you retired,” Bengen said.

Yes, and there’s really no good alternative. Even withdrawing 4%, FIRECalc shows 5% of portfolios failing, which is pretty high for a low risk plan. A withdrawal rate of 6%-7% as suggested in the linked article by Bengen completely ignores sequence of returns risk and assumes that the “bad times in the 70’s” won’t repeat.

The 4% rule is just a talking point...to me. The optimal result is that I spend my last nickel on the day I die. A planned withdrawal rate decided upon thirty years before I think I will die is just a guessing game. The fallacy is not that the originator revised the number from 4% to 4.7%. The fallacy is believing any simplistic understanding is sufficient to make a decision today that will yield reliable, market based, results in thirty years.