You often see the “3% rule” quoted in retirement literature. Many times it is applied incorrectly, so I’d like to revisit the rule, clear up some misconceptions, and let Dick and Jane show you how it works in their retirements.

Refresher Course

Financial planner William Bengen is responsible for the 3% rule, although that’s not what he called it. Actually it’s the 4.5% rule, but he didn’t call it that either. He called it “Safemax,” and it goes like this:

“*If you spend 4.5% of the assets in your tax-deferred accounts in your first year of retirement and increase that annual dollar amount each year by the inflation rate, your nest egg will last at least 30 years under all the historical scenarios tested*.”* *

Bengen realized that relying on the last 80 years’ averages for stocks, bonds, and inflation only helps the hypothetical retiree that actually experiences those exact averages. But what about a retiree in a real 30-year time chunk during that historical period? How would he have fared in the best of times and in the worst of times?

His research showed that a retiree in the worst 30-year time stretch over the 80 years that he studied (which included some whopper recessions and even the Great Depression), would not have run out of money if he withdrew 4.5% of his nest egg the first year of retirement, and then adjusted that dollar amount each year for inflation.

Caveats

-This only held true for retirees that could stomach at least 50% of their portfolios in equities. If you want to do it all on treasuries, the rule does not apply to you. Your money would not have lasted in the worst of times if you were not willing to invest at least 50% in equities.

-In many of the time periods studied, your money would have lasted even longer than 30 years (or put a different way, you could have withdrawn more than 4.5% over those 30-year stretches). The Safemax is purely to figure out what would have been the worst outcome. In the case of the 4.5% rule, the worst outcome would have been that you would be left with no money at the end of 30 years.

-Thirty years isn’t long enough for a healthy retiree wanting to retire in her 40’s or 50’s. That’s why I often quote the 3% rule. It’s just the 4% rule adjusted for the research showing the Safemax for 50 years instead of 30 years.

-The rule does NOT say that you withdraw 3% of your assets each year. That is a common misconception. Read this again: You withdraw 3% of your assets only the first year and then adjust that dollar amount each year for inflation. After the first year, you never again multiply your ending assets by 3% to come up with the withdrawal rate.

Meet Dick and Jane

Dick and Jane are both 50 years old. The year is Year 1, and the S&P index is at 900. Dick and Jane each have identical $1 million portfolios of which 50% is invested a well diversified mix of U. S. large-, mid- and small-cap stock funds as well as international developed- and emerging-market funds. The other 50% is diversified among short-, long- and mid-term bond funds and includes a 3-year supply of living expenses in an FDIC-insured savings account. (As the author of this story, I will not let Dick or Jane invest 100% of their portfolios in the 30 stocks that make up the DJIA. That would be foolish, and as their advisor, I would not let them do something foolish.)

Dick decides to retire right now because his retirement budget is $30,000 and that represents exactly 3% of his next egg. I failed to mention that Dick and Jane live in Podunk, where the cost of living is very low. I also I failed to mention that Dick really hates his job. (And of course Dick doesn’t know this now, but he will live exactly 50 years in retirement, unfortunately dying one day short of his 100^{th} birthday in a tragic bungee jumping accident.)

Jane likes her job ok and wants to stick it out at least for another five years, to hopefully let her portfolio recover from a Big Recession. She retires in Year 5 when the S&P is at 1,400. Now, of course her portfolio is more than the $1 million it was in Year 1, but it is NOT 56% higher as would be indicated by the measure of the S&P index. Only half of her nest egg was even exposed to the stock market, and some of that was in international funds, so her portfolio is 20% higher than it was at Year 1, now at $1.2 million. Dick’s is of course smaller since he has been taking withdrawals from his identical portfolio for five years.

(It might be a good time to tell you that Jane also lives for 50 years in her retirement, dying of natural causes on her 105^{th} birthday.)

Between Year 1 and Year 5 inflation has been 3.5% per year, so while her budget was identical to Dick’s five years ago, $30,000 doesn’t get you what it used to, even in Podunk. Dick is now withdrawing $34.5k (his inflated base amount). Three percent of Jane’s $1.2 million portfolio puts her first year withdrawal Safemax at $36k, which by the way, she really doesn’t need because she already lives on $34.5k—the inflated $30,000. But she has expensive taste in shoes, so she decides to take out the whole $36k maximum.

Now some of you might be thinking, that’s so unfair, why does Jane get to take out more for the rest of her life than Dick does? Just because of the arbitrary place on the S&P (and the affect on her respective portfolio)?

**The 3% rule does not say that Dick can’t take out more. The** **rule just says that if Dick had been unlucky enough to retire during a 50-year period that was as bad as the worst 50 years in recent history, he wouldn’t run out of money.** In fact, it’s likely that he didn’t retire in the worst 50-year stretch of history given that the market went up 56% in the first 5 years of his retirement.

In fact, it’s likely that Jane’s 50-year stretch is going to be worse than Dicks, given that Dick already experienced a great 5 years in the beginning of his retirement. The next 45 years, they experience the exact same market together, and then even if Jane experiences a stellar market in her last 5 years of living, it would be on a much smaller asset base by that time. **But the rule says that even if she experiences a 50-year stretch as bad as the worst in recent history, she won’t run out of money either.**

So what does this rule mean for Dick and Jane? Since Dick and Jane’s retirement overlapped 45 of the same years, and Dick had a stellar first five years in the market, if either of them experienced the worst 50 years, it would have been Jane. If Jane did, indeed experience the worst 50 years, based on Bengen’s research, this means that Jane dies with no money, and Dick dies with some money left. That’s all the rule says.

Raise your hand if you have any questions.

**Related Posts:**

** **How Much Money do You Need to Retire?

How to Set a Retirement Savings Goal

A Smarter Approach to the 4% Rule

Can’t keep track of my non-existent posting schedule? Subscribe—it’s free!

@New: No, Bengen's rule doesn't say Guy 2 can't retire, it says that if he wants to be totally safe (in case the 50 years FROM HIS RETIREMENT DATE are the worst 50 years) he can only withdraw 3% of his assets (the lower number).

Guy #1 and Guy #2 can't both possibly have the worst 50 years (based on history) from their start dates because they didn't start at the same place. One of them has to do worse than the other when you look back over the 50 years. Since 49.5 of them are the same, it's likely Guy #1 had the worst 50 years, historically speaking. So he dies with nothing, Guy #2 dies with some left.

That's how they reconcile. Now if you are assuming Guy #2 had the worst 50 years, then Guy #1 had an even worse 50 years than history showed capable, and the rule doesn't cover that. It presumes you will do at least as well as the worst 50 year chunk.

The rule doesn't say that everyone that follows the 3% rule dies with nothing. Only the ones with the worst 50 years from their start date die with nothing. The others die with money left over (or could have taken out more than 3% like your Guy #2 above--in other words, he COULD have retired with a higher than 3% rate, just not if he wanted to feel totally safe at the front end of those 50 years. At the front end, you're not sure which guy you're going to be.

The rule also doesn't say that you can only retire if you have enough to meet the 3% rule. It just says if you want to bet on the fact that you will have the worst possible (historically speaking) luck, you shouldn't pull the trigger unless you have met that 3% rule.

With that, haven't we beaten this thing to death yet?

Posted by: Retired Syd | March 26, 2012 at 06:22 PM

Your answer is akin to saying if I just flipped a coin and it came up heads, then if I flip it again it should come up tails. But as we all know, the events are completely independent of each other, just like guy1 and guy2 in our example. I really wish I could get you to understand this. But I'm willing t give up if u r.

Posted by: New at this | March 26, 2012 at 07:52 PM

@New: Yes, if only you could get me to understand this. But alas, you cannot. I think I'm beyond help. It's time for you to give up on me. Sorry.

Posted by: Retired Syd | March 26, 2012 at 09:17 PM

So the takeaway for anyone else following this thread is:

It wouldn't be advisable for Guy2 to retire using a 5% withdrawal rate. Therefore, since his assets are identical to Guy1's, Guy1 should be VERY nervous about continuing to withdraw at his originally calculated amount (which now represents 5% of his Assets).

Syd doesn't agree with the above statement, but if you're basing your own financial future on this, I'd advise putting some type of over-ride into your plan. Independent of where you started, you should always ask whether someone else starting out from scratch would agree to the plan, that you are already in the middle of.

Best of luck to all, and thanks for engaging Syd.

Posted by: New at this | March 27, 2012 at 06:03 AM

Syd, Does the 3% rule assume a 50% stock, 50% bond/cash investment portfolio until you die?

Posted by: Mark j | March 27, 2012 at 08:21 AM

@New: You are correct, I do not agree. Guy #1 should not be nervous because Historical Guy #1 (who experienced the exact same scenario historically) turned out to be ok. Guy #1 should only worry if he thinks his 50 years are going to be worse than Historical Guy #1--in which case the 3% rule would not have worked anyway.

You might allow for the remote possibility that perhaps it is you that are misunderstanding the rule.

@Mark: Yes it does.

Posted by: Retired Syd | March 27, 2012 at 08:52 AM

Syd- Its not a question of understanding the rule. (I think I could write an article on it at this point).

The issue is that I believe the rule, as written, has a (perhaps significant) deficiency.

It does not properly handle conditional probabilities as it relates to back-testing. The proof is in the above example if you would just take off your blinders and let your mind soak it in....I have a Masters degree in this topic, but until now haven't ever applied it to this area. And in an engineering context, this "science" would get laughed out of the room for its indefensible deficiencies.

And if people don't understand this, they are going to get hurt, I'm just trying to point that out, as a matter of helping my fellow wanna-be-retirees.

So what I was hoping to to evolve the discussion to was finding a "reason" why withdrawing 3% from a retirment portfolio might make inherent sense, beyond relying on this shallow research.

Posted by: New at this | March 27, 2012 at 10:17 AM

@New: It's not science, it's financial modeling.

Let me try one more thing:

Over the 80 years studied 1 guy had the worst 50 years, another guy had the 2nd worst 50 years, another guy had the 3rd worst 50 years another guy had the 4th worst 50 years and someone had the absolute best 50 years.

You with me so far?

The guy that had the worst 50 years would have been ok at a 3% withdrawal rate, the guy with the 2nd worst period would have ultimately been ok with a little higher withdrawal rate, and so on and so on. The guy with the best 50 years would have done ok with a really high rate, maybe 5 or 6%, maybe higher (I didn't go back in the article to see what that would have been for the best years because no one really cares.)

Stay with me:

In your example above you are trying to give Guy #1 AND Guy #2 the absolute worst 50 years. But, in that new historical 80 year period which we are assuming the same as the 80 year historical period, they can't both have the worst 50 years if we already gave that award to one of them. So the other one could have taken out more. The 3% just applied to the guy that was the worst.

I know this didn't work, but I really am trying.

Posted by: Retired Syd | March 27, 2012 at 11:05 AM

@New: P.S. It ONLY makes sense to withdraw 3% if you happened to be the guy with the worst 50 years. For everyone else, they need to get in their time machines, go back to the beginning of their retirements, and take out a higher %.

Posted by: Retired Syd | March 27, 2012 at 11:20 AM

Syd - pls quit reciting from the Bengen article and do some independent thinking.

Ask yourself, how is it that Bengen's own research gives guy1 and guy2 a diff. Prob of success despite having identical assets at the same point in time?

Posted by: New at this | March 27, 2012 at 11:26 AM

For some crazy reason....I completely understand what Syd is saying.....but I ALSO understand what New is getting at too. You guys both understand the concept.....however it is certainly revealing how engineer minds think differently than accountants:)......in any case retirement, as in life is not an exact science. Retirement and LIFE is about engaging flexibility in your decision making when needed.

Posted by: nicole | March 27, 2012 at 12:24 PM

"@New: It's not science, it's financial modeling." BINGO! These are tools in an ever changing world. Will history repeat? Are we in such a period of staggering debt that using the past as a reference is meaningless? Will personal circumstances be completely different than imagined? Who knows? Be flexible using all available tools, keep options open and most importantly - enjoy life!

Posted by: Steve | March 27, 2012 at 05:07 PM

So I thought about this a little more as I stared at a 100 year chart of the Dow...

I think all Bengen was trying to do, was provide a short-cut method for taking a first-cut at whether someone is in the ballpark of being able to retire. And to that end, think he did a great job....

But as pointed out with the above guy1/guy2 example, I don't think this gives the all-clear to declare your retirement date at "outlier" moments in time like March 2000, (when the Tech Boom was peaking)...(I know Syd disagrees, but I respectfully don't care)

A little judgement is warranted, and some type of over-ride should be implemented to make sure your not an anomoly....The problem is however, that you don't always know you're at an outlier moment in time until several years after-the-fact...

To that end, as I decide how to implement my own version of the 3% rule, I'm going to start by calculating 3% of the value of my assets, "nomalized" to what their value would be if the dow was between 10,000 and 11,000.

Then, going forward if my Current Withdrawal amount gets out of bounds with the 3% rate along the way, I'm going to think seriously about scaling back for the moment....

With this adjustment, I think I have found the answer I set out looking for...which is a retirement plan that allows me to sleep....

Thanks for engaging in my journey, and best of luck with yours, Syd, and all the rest on this forum....I cannot wait to make it official!!!

But alas, have decided to stick it out long enough to collect one last bonus in December....Then, I am joining the rest of you...!!!

Posted by: New at this | March 28, 2012 at 03:00 PM

Consuming 3% of a 50/50 equity/fixed portfolio in year one, then consuming that amount adjusted for inflation every subsequent year for 50 years will deplete that portfolio if equity returns, interest rates, and inflation are no worse than what they were in the worst 50 year period of the past 80 years. If conditions turn out better, one dies with an estate, if conditions turn out worse, well, that’s what some folks worry about.

Limiting present consumption to follow the rule treats future evils of the like of the Great Depression, World War II, and the Oil Shock as certainties. What does one expect, an alien invasion?

Posted by: Cyclesafe | January 17, 2013 at 06:10 PM

How did you compute the $34,500 in the original scenario? I understand the first year $30,00 but don't understand what/how to come up with the inflated base amount.

Posted by: Judy | February 02, 2013 at 07:14 PM

Judy:

No problem, here's the math. So you got the original 30k--3% of Dick's nest egg. That's year 1. But each year, inflation increases that 30k basket of goods by the inflation rate:

Year 2's expenses (and withdrawal) will then be 30,000*1.035 (increasing by the inflation rate)= $31,050.

Then Year 3's withdrawal will be 3.5% higher than that: 31,050*1.035 = 32,137 and so on:

Year 4: 32,137*1.035 = 33,261

Year 5: 33,621*1.035 = 34, 426 (I rounded up to 34.5 for the post)

That's how the rule works--a common misconception is that you simply take 3% of your ending nest egg out each year. You actually only take that % out the first year and then the percentage increases each year. For an illustration of how that would work, assuming a 4% withdrawal rate see this post: http://retiredsyd.typepad.com/retirement_a_fulltime_job/2012/03/see-average-joe-retire.html

Hope that helps!

Posted by: Retired Syd | February 02, 2013 at 07:47 PM

It helped very much so! This is the first time I've ever read a good explanation of withdrawing retirement funds. I have a lot of research to do in the next five years and both enjoy, and are learning from, your blog.

Posted by: Judy | February 03, 2013 at 12:26 PM