SWR Question - Yearly Rebalancing Is Good?

[bbuzzard]

I am certain this has been asked before, but I cannot find it in the archives.

I feel I have a clear understanding of SWRs, and understand why the 4% rule works. However, it seem that the real world will most likely allow significantly higher WRs. THis is what I mean:

The 4% rate comes from assuming the worst possible returns based on historic data. In reality, actual returns are mopst likely going to be much higher than this. If you withdrawal 4% a year of you initial portfolio, adjusted yearly for inflation, you heirs are most likely going to be very, very wealthy. OTOH, I do not want to live on Alpo when I am 70, so I have to start conservative at 4%. However, why can't this be adjusted to 4% of my current balance at the start of each year. In other words, I restart my retirement each year. If the market went up 20% last year, then my SWR also goes up 20% (still 4% of the portfolio, but +20% in toital dollars). No need to ever adjust downward because the 4% rule states I am safe as long as I limit myself to 4% of the portfolio from any given starting date.

In summary, there is no difference between someone who has been retired for 10 years and has a $3M portfolio and someone whjo is retiring this year with a $3M portfolio. Both should be able to safely withdrawal $120k. However, under current logic (as I understandin it), the fellow who has been retired for 10 years may only be withdrawling $85k because than is 4% of his original portfolio (plus inflation adjustment).

What am I missing? Am I stating to obvious? Am I stupid and missing the obvious?

 

[CybrMike]

Your missing the fact that you only have 20 years left to live, whereas the new person has 10 more years to recover from a bad market correction/crash.

 

[bbuzzard]

Your missing the fact that you only have 20 years left to live, whereas the new person has 10 more years to recover from a bad market correction/crash.

I respectively disagree. The younger person has to ensure that their money will last 10 years longer, and therefor has more, not less, risk for a given WR.

 

[CybrMike]

not when you take into account the fact that the "older" person has went through a "good" market period, which supposedly increases the chances of a "bad" market period coming very quickly.

 

[CybrMike]

So I put it into firecalc.

Seems if you do a $3M portfolio with a 20 year lifespan the SWR = 5.21 whereas the 30 year lifespan SWR = 4.52

Although this obviously does not take the odds of starting after a bull market has had a run.

 

[charlie]

Hi bbuzzard,

4% of your current balance is a good SWR method and is likely
to give you a higher withdrawal amount on average.

There has been a lot of discussion on this subject but I am too
lazy to look up the threads for you.

Check out his link for more insight:

http://www.gummy-stuff.org/sensible_withdrawals.htm

Cheers,

Charlie

 

[cute fuzzy bunny]

Boy is this starting to sound like the "SWR 6.21%" thread...

If the question is a real one, the answer is simple. You may model any return or any expectation of a major downside event as you wish, the reality is going to be different.

If your portfolio cant survive an event like the depression or 'stagflation', and one of those happens, you'll run out of money. You'll be in good company though.

If you have 25x your annual spending in your portfolio and you have a well diversified portfolio, you'll probably make it even if something bad like that happens.

 

[Nords]

The 4% rate comes from assuming the worst possible returns based on historic data. In reality, actual returns are mopst likely going to be much higher than this.

Good luck with that.

Before you crunch the numbers I'd recommend you read "Triumph of the Optimists. The authors point out that the historical record is filled with data contaminated by "easy numbers" (like the Ibbotson data starting in 1926), survivor bias, unrealistic assumptions on expense ratios, and bad CPI indexes.

The 21st century's returns for an all-equity portfolio are predicted to be about 5% over inflation. Inflation for the last century has run about 3%, although for the last 30 years it's been more like 5%. And the Gordon Equation ratchets predicted returns down to about 6-7%. So maybe your 8% number is good for pre-tax, pre-inflation returns. After taxes & inflation you'll probably be looking for low-cost investments.

When you start reading that retirement spending declines with age, remember that the study didn't assess the effects of rising healthcare costs...

 

[bbuzzard]

You guys always make things to complicated. I was looking for a "why yes, you can do this with do risk what so-ever." Instead, you dump cold water on my head .

 

[cute fuzzy bunny]

You can do this with do risk what so-ever.

 

[bbuzzard]

While there are always risk regardless of what we do (the future is always unknown and unknowable), I still think I am essentially correct. For example, suppose you retire at age 50 with a $1M portfolio. You dutifully begin withdrawing $35k/year. Five years later, due to a great market, you have $2M in the bank. I think it would be crazy to limit youself to 3.5% of your starting assets (All numbers are inflation adjusted to constant value dollars in this example). I would feel more than comfortable to increase my stipend to 3.5% of my current portfolio, especially knowing that I have been living on 1.75% of my current portfolio for the past five years, and could do it again if needed.

 

[TromboneAl]

Here's a simpler way of looking at things.  You've discussed two SWR strategies. 

1. Taking 4% out the first year and adjusting for inflation for future years.

2. Taking 4% out every year.

These will give significantly different results.  FireCALC lets you estimate what the result of strategy #1 is.  If you want to use stragegy 2, you'll have to find a different app to estimate the results.

 

[bbuzzard]

Here's a simpler way of looking at things. You've discussed two SWR strategies.

My point is, they are the same. Look at is this way. Image two twins (obviously the same age). Twin #1 retires Jan 1, 2006 with assets of $1M. Based on FireCalc, she chooses a SWR of 4% and withdraws $40,000 on the first day of the year for living expenses. In 2006, inflation is 3%, and investment returns are 18%. Therefore, Twin #1 has $1,132,800 on January 1, 2007. She adjust her withdraw for 3% inflation and takes out $41,200 for living expenses in 2007.

Twin #2, retires on January 1, 2007, exactly one year after her sister. She happens to have exactly $1,132,800 on this date, the same as Twin #1. Based on FireCalc, she also chooses a SWR of 4%, taking $45,312 our of her account for 2007.

Riddle me this, Batman: What is the difference between the two twins? Why is $45,312 safe for one, but only $41,200 is safe for the other? What is the difference? My point is, readjusting you withdraw rate upward each year to 4% of you current assets does not effect the probability of portfolio success predicted by FireCalc relative to you original success rate.

 

[Zipper]

Twin #2 will have periods of volatility when her draw will be less than the year before.

The key is to have balance (60/40) so you don't blow up in a bad year or a series of bad years.

The same applies in a stretch of good years. Don't go nuts and get greedy. Stick to the plan and take only the 5% (less if you ER'd) in good and bad years.

A lot of pension plans got into trouble because they raided the cookie jar when times were good.

 

[samclem]

bbuzzard,

Charlie gave you a link to a good site--Gummy explains the two approaches well, and he even has an intermediate aproach that he likes.

The two approaches (4% plus inflation vs 4% of the yearly end value) are very different. Look at Gummy's spreadsheets and it will be clear. If you can handle the volatility (e.g if you have another source of inflation-adjusted income for your bare necessities), then I believe basing your withdrawals on your end balance each year is much safer. Think of it this way--there's no reason to believe a particular portfolio (unless it is all TIPS, etc) is going to track with inflation--equities and various bonds have fallen well behind inflation for many years at a time, especialy when inflation is high. I'm much more comfortable pegging my withdrawals every year to the performance of my portfolio rather than to inflation. Taking a percentage of the yearend balance also is like value averaging-- you sell fewer securities when the share proces are lower. By taken out less when the portfolio has poor performance you leave enough money in to permit a recovery when the market clibs. Finally, you can never totally run out of money with this approach--but if your withdrawal percentage is too high you'll experience constantly declining real withdrawals oer the years.

 

[MasterBlaster]

Charlie:

Thanks for the fabulous link.

He puts some science into portfoio withdrawal rates.

 

[Rich_by_the_Bay]

bbuzzard,

Not sure any of the other posts mentioned this: your "4% of assets each year" method means that your annual income will vary significantly from year to year. You may have years of 15% or 20% less than the initial income, and years way above that. You may have years of down income in the face of ongoing inflation. In other words, in the end you will probably be fine regarding preserving your balance but income will be an annual rollercoaster.

With the 4% annually adjusted for inflation, you will have gentle fluctuations in income of, say, 3-5% which are guaranteed to preserve your purchasing power. Less aggressive a stance, but much less volatility of income.

If you have enough money that you could tolerate a major income reduction in a down year or two or three or four, go for it. If you need a bit more predictability, use plan A.

BTW, Stein and Demuth recommend a compromise: reset the thermostat every 5 years by restarting at 4% of actual assets, but live with inflation-adjustments-only in between. That's what I hope to do.

 

[rrspike]

bbuzzard,

Not sure any of the other posts mentioned this: your "4% of assets each year" method means that your annual income will vary significantly from year to year. You may have years of 15% or 20% less than the initial income, and years way above that. You may have years of down income in the face of ongoing inflation. In other words, in the end you will probably be fine regarding preserving your balance but income will be an annual rollercoaster.

With the 4% annually adjusted for inflation, you will have gentle fluctuations in income  of, say, 3-5% which are guaranteed to preserve your purchasing power. Less aggressive a stance, but much less volatility of income.

If you have enough money that you could tolerate a major income reduction in a down year or two or three or four, go for it. If you need a bit more predictability, use plan A.

BTW, Stein and Demuth recommend a compromise: reset the thermostat every 5 years by restarting at 4% of actual assets, but live with inflation-adjustments-only in between. That's what I hope to do.

This is a great discussion/thread. Near and dear to me, too.

The approach of taking 4% of assets as-of beginning of each year would seem to me to require steadfast discipline with an acknowledgement that you are letting your contemporary asset base (and therefore the vagaries of the current market) rule your actions, rather that letting historical perspective over the last 80 years rule your actions. This can lead to your emotions making decisions, instead of hard facts driving the decisions.

Without a strutured plan, in this scenario how many will ask twice before withdrawing 1.5 or 2x last year's distribution if the portfolio increased in value 1.5 or 2x? On the other hand, how many will ask twice before withdrawing only 50% or 60% of their previous year's distribution if the portfolio assets took a dump? Will they have the discipline to do it again the next, and the next year, if there is a protracted down market. Or does the method fly out the window when the pain gets too much to bear, and distribution amounts are somehow arrived at arbitrarily/emotionally ? Regardless of the discipline, how many (who have based their withdrawal model on any SWR) can handle such potentially huge swings in withdrawals needed to cover their living expenses?

If there are those who think they can live off of $40,000 one year and tighten the belt and live on $25,000 the next year, then their biggest challenge may not be as esoteric as deciding on how to implement the 4% SWR.?

I think i would rather believe in the numbers, but that's just me - i'm a scientist. 4% SWR is based on historical inflation rates and market returns of a fairly rigid asset allocation mix (statistically proven to be the most efficient) over a series of timeframes over the last 80 years or so. It's better than leaving it up to chance. But i like the 'reset the thermostat' idea as a compromise..maybe every 5 or 10 years (again, depending on moving averages).

 

[dory36]

My point is, they are the same. Look at is this way. Image two twins (obviously the same age). Twin #1 retires Jan 1, 2006 with assets of $1M. Based on FireCalc, she chooses a SWR of 4% and withdraws $40,000 on the first day of the year for living expenses. In 2006, inflation is 3%, and investment returns are 18%. Therefore, Twin #1 has $1,132,800 on January 1, 2007. She adjust her withdraw for 3% inflation and takes out $41,200 for living expenses in 2007.

Twin #2, retires on January 1, 2007, exactly one year after her sister. She happens to have exactly $1,132,800 on this date, the same as Twin #1. Based on FireCalc, she also chooses a SWR of 4%, taking $45,312 our of her account for 2007.

Riddle me this, Batman: What is the difference between the two twins? Why is $45,312 safe for one, but only $41,200 is safe for the other? What is the difference? My point is, readjusting you withdraw rate upward each year to 4% of you current assets does not effect the probability of portfolio success predicted by FireCalc relative to you original success rate.

You are right about the differences, and you can throw out the old plan based on the lower balance and start a new one based on the higher one.

This was discussed a lot at http://early-retirement.org/forums/index.php?topic=5572.27 starting about halfway down the page, so I won't repeat it all here.

 

[cute fuzzy bunny]

Besides that explanation, there may be another.

A brief side story...the reason why many unmanipulated monte carlo simulations produce such different numbers from real data sequences like the ones firecalc uses is that they lose any year to year correlations. From psychological (the markets cheap/expensive, i'm gonna buy/sell) or the various "laws" and "rules" that show that over 20+ year periods market returns represent company earnings and growth and the psychological aspects are muted.

Starting a run at the point of a downturn, at its bottom, or somewhere in the middle...backtested against the historic, fully correlated data, produces different results and different withdrawal rates. The thing that confuses people is that the real 'safe' rate going forward will be the same regardless of the scenario 'cherry picked' from the past.

In short, someone starting at the beginning of a downturn (say in 2000) vs at the beginning of an upturn (say in 2003) will get different numbers even with the same starting portfolio size. One is pre-disastered, the other hasnt gotten there yet. History says the predisastered one will have a shorter time to recovery because downturns have always had an end. Reality going forward will produce a completely different number.

In real short, its an estimate and an imperfect tool thats good for rough guidance, not a crystal ball.

In very short, if firecalc says your withdrawal rate survived all of its historic models, then you could have made it through the depression and stagflation. Unless we have something worse, if you dont take more than the stated withdrawal rate, you'll be fine. You can probably take more. At your peril.

 

[sgeeeee]

Here's a simpler way of looking at things.  You've discussed two SWR strategies. 

1. Taking 4% out the first year and adjusting for inflation for future years.

2. Taking 4% out every year.

These will give significantly different results.  FireCALC lets you estimate what the result of strategy #1 is.  If you want to use stragegy 2, you'll have to find a different app to estimate the results.

Just a note. You can use FIRECalc to evaluate strategy 2 also. Simply enter your initial withdrawal rate as $0 per year and increase your expense ratio by 4% (to 4.18% if you normally use the default of .18%).

Of course the answer you get will be 100% survival. Becuase no matter how small your portfolio gets, there will always be 4% available. In order to get any value from the simulation you have to look at the detailed results and see what 4% of the portfolio value really is.

A better way to look at this is to assume that some level of spending is required for survival and use this as an iniitial withdrawal rate. Then increase the expense ratio till the simulation fails. The total of the inflation adjusted survival level of withdrawal plus the percentage of portfolio value over your actual expense ratio is what you can spend from year to year.

I posted some results from this kind of simulation on the board somewhere, if anyone is interested in doing the search.

 

[sgeeeee]

My point is, they are the same. Look at is this way. Image two twins (obviously the same age). Twin #1 retires Jan 1, 2006 with assets of $1M. Based on FireCalc, she chooses a SWR of 4% and withdraws $40,000 on the first day of the year for living expenses. In 2006, inflation is 3%, and investment returns are 18%. Therefore, Twin #1 has $1,132,800 on January 1, 2007. She adjust her withdraw for 3% inflation and takes out $41,200 for living expenses in 2007.

Twin #2, retires on January 1, 2007, exactly one year after her sister. She happens to have exactly $1,132,800 on this date, the same as Twin #1. Based on FireCalc, she also chooses a SWR of 4%, taking $45,312 our of her account for 2007.

Riddle me this, Batman: What is the difference between the two twins? Why is $45,312 safe for one, but only $41,200 is safe for the other? What is the difference? My point is, readjusting you withdraw rate upward each year to 4% of you current assets does not effect the probability of portfolio success predicted by FireCalc relative to you original success rate.

You are right bbuzzard. To the extent that we believe one answer, we should believe it for both. There are really two issues here that we tend to get wrapped around becuase we confuse them:

1) do we believe FIRECalc so completely and believe we know our lifespan exactly so that we will make our plans to run out of money the year we die? Or do we figure that there is some probability that we live a very long time or that we face financial times worse than we have seen in the past 120 years?

2) if we understand and believe FIRECalc has some validity and value, can't we reapply it in future years after we have retired and assume the results have the same validity?

The answer to the first question is, "Maybe planning a cushion makes sense."

...and the answer to the second question is, "Yes." Look at the detailed FIRECalc results. Dory actually plots them on the opening page for a $1M portfolio with high probability of success. Some of those sequences ended up with over $6M (remember those are inflation adjusted dollars, too). Clearly, retirees who are fortunate enough to end up on one of those trajectories can afford to spend more. If after a few years of retirement you find yourself with a SWR much higher than your actual spending, you can decide what to do with it. You can use it to buy safety and a further cushion (see point 1). Or you can spend some of it. FIRECalc results are as accurate at that point in time as they are today. That statement does not imply that they will actually predict the future for either case.

 

[free4now]

The numbers do indeed show that 4% of initial withdrawls produces the same 100% historical survivability as 4% of yearly balance.

But the subtle thing to realize is that the guy who took out 4% of initial will have a bigger nest egg at the end (and most points between) than the guy who took 4% of yearly balance. It makes sense... if you withdraw less you keep more.

If at any point in the future things get worse than history, the one with the bigger nest egg is in the better position. The one with a just-big-enough-for-historical nest egg is in trouble.

 

[cute fuzzy bunny]

If we hit something worse than the great depression, nest eggs will be the least of our problems.

Further, as I like to point out, we all have better financial 'sneakers' than 90+% of the rest of the people in the world, and therefore are fare more likely to be able to outrun that 90% while we're all running from the bear.

A mass financial catastrophy like that wouldnt faze me a bit. My home is paid for. I can shovel dirt to make enough money to feed my family. And excepting complete destruction of our country and economy, I'll still be relatively wealthy compared to my neighbors. Heck, I'd probably be a major buyer of real estate and equities at pennies on the dollar.

If it all goes so to hell that seven figures doesnt "survive", eight figures probably wouldnt be any different.

 

[MasterBlaster]

Fireme:

The 4% SWR method only produces a bigger nest egg if markets are rising compared to the 4% annual take.

In either case an ER is unlikely to let his stash rise so large that he doesn't increase his withdrawals. Similarly in hard time an ER is unlikely to keep taking out large withdrawals to the detriment of his stash just cause the formula says he can.

That's why I liked the gummy-suff analysis. It gave some mathematical rigor to the likely behavior of real-world retirees. In great markets you'll probably spend some extra. In poor market you'll cut back.