Greater than 4% Withdrawal Rate

[brewer12345]

If you're ER'ing with at a 95% "success rate", you've got a pretty slim chance (~1 in 20) of having picked the worst year or two that leads to a busted ER. A few years later, with a little growth, you've grown away from the 95% rate, perhaps up to 100% and it's smooth sailing.

But if you reset after every up year, you're right back up to that 1 in 20 chance, again and again...do that 10 or 20 times and there's a mighty good chance you'll find that troublesome patch.

Cb

But you are also a year older every time you reset. Do it 10 or 20 times and you will be 10 or 20 years older and the money won't need to last as long (AKA you r withdrawal period will be shorter). So there is a counter-balance.

 

[Rich_by_the_Bay]

But you are also a year older every time you reset. ...and the money won't need to last as long (AKA you r withdrawal period will be shorter). So there is a counter-balance.

... and less time to recover after a prolonged down market, no? Resetting too often seems to increase your volatility risk either way.

 

[brewer12345]

... and less time to recover after a prolonged down market, no? Resetting too often seems to increase your volatility risk either way.

I think we need a statistics wonk to tell us, because so much of stats is counter-intuitive.

I don't imagine it would be smart to push the envelope, but resetting every few years after you atrted out with a very safe WR should be OK. If you expect a 30 year draw and make it 15 years with your capital still intact, you can afford to do a lot of damage in the remaining years.

 

[FIRE'd@51]

I'm sure others have probably done this. I ran the standard FIREcalc for a whole bunch of equity/bonds mixes and found that the highest SWR which gives a 100% success rate for a 30-yr time frame is 3.75%, which occurs for a 36/64 equity/bond mix. (For a 75/25 mix the highest SWR is 3.59%). So, if one believes FIREcalc provides the worst-case scenario, why not simply choose the highest SWR that gives a 100% success rate. Then one should be able to "restart the clock" with 100% confidence whenever one wants.

 

[brewer12345]

I'm sure others have probably done this. I ran the standard FIREcalc for a whole bunch of equity/bonds mixes and found that the highest SWR which gives a 100% success rate for a 30-yr time frame is 3.75%, which occurs for a 36/64 equity/bond mix. (For a 75/25 mix the highest SWR is 3.59%). So, if one believes FIREcalc provides the worst-case scenario, why not simply choose the highest SWR that gives a 100% success rate. Then one should be able to "restart the clock" with 100% confidence whenever one wants.

I think the problem is that there is an implicit assumption that the future will be no worse than the past. But we a splitting harirs here.

 

[MasterBlaster]

I'm sure others have probably done this. I ran the standard FIREcalc for a whole bunch of equity/bonds mixes and found that the highest SWR which gives a 100% success rate for a 30-yr time frame is 3.75%, which occurs for a 36/64 equity/bond mix. (For a 75/25 mix the highest SWR is 3.59%). So, if one believes FIREcalc provides the worst-case scenario, why not simply choose the highest SWR that gives a 100% success rate. Then one should be able to "restart the clock" with 100% confidence whenever one wants.

From the Retire early homepage study on SWR's...

http://www.retireearlyhomepage.com/restud1.html

Note: Check your asset allocation as it disagrees with the chart

 

[Cb]

But you are also a year older every time you reset. Do it 10 or 20 times and you will be 10 or 20 years older and the money won't need to last as long (AKA you r withdrawal period will be shorter). So there is a counter-balance.

No doubt, aging offsets a bit of the roulette game you play with frequent resets.

Here's how I look at it:

Bailing once your portfolio is in the 95% zone is probably safe enough...if the sh7t hits the fan in the early going you're still fairly young and marketable should it come to that. If you keep running right back up to the 95% threshold with regular resets you might just find that rough patch when you're 68 with only Windows 95 skilz...

Cb

 

[Cb]

A few years ago I was thinking hard about the different variable wihdrawal schemes I was finding on the internets...Gummy's "Sensible Withdrawals", Guytron's 67 Rules, SG's FIREcalc expense ratio trickery, Bob90210's website, etc.

The "safety bonus" or somewhat higher SWR they offered was appealing, but I was still reluctant to go into retirement on that basis because whatever the variation in the WR would obviously have to come out of the discretionary half of our spending.

But a while back I had something of a brainstorm (by my standards)...I realized that of my discretionary budget items, a goodly percentage were for "long-cycle" things like car replacements, every-third-year Hawaii trips, home redecorating, etc, as opposed to "short-cycle" discretionary expenses like dinners out, rounds of golf, and entertainment.

It then occurred to me that a variable withdrawal strategy really wouldn't be all that painful in reality...we could simply time our long-cycle expenditures - in other words, putting off a Hawaii trip or car replacement during a market downturn. Delaying those expenditures until markets improve should allow us to leave our routine discretionary expenditures pretty much alone.

Cb

 

[Rich_by_the_Bay]

But a while back I had something of a brainstorm (by my standards)...I realized that of my discretionary budget items, a goodly percentage were for "long-cycle" things like car replacements, every-third-year Hawaii trips, home redecorating, etc, as opposed to "short-cycle" discretionary expenses like dinners out, rounds of golf, entertainment.

It then occurred to me that a variable withdrawal strategy really wouldn't be all that painful in reality...we could simply time our long-cycle expenditures - in other words, putting off a Hawaii trip or car replacement during a market downturn. Delaying those expenditures until markets improve should allow us to leave our routine discretionary expenditures pretty much alone.

Exactly. A "percent of assets" strategy and some flexibility is appealing.

You basically never worry about running out of money. You could run out of "lifestyle," though, so you need to be sure that the fat you might need to trim during down years is just that, fat. That is, you need a good cushion over and above your standard, comfortable expenses. But when stocks are up, it's playtime.

 

[sgeeeee]

I still don't understand how resetting on good years can in any way impact your initial chance of success that's based on a certain withdrawal rate.

Jane retires in 2005 with 1,000,000 and decides on a 4% withdrawal rate (40,000) that gives 95% chance of success. What a great year, Jane's portfolio in 2006 is now 1,200,000. Jane resets and decides to take 4% from her new stash so now she's living large with 48,000 withdrawal. 4% rate, 95% chance of success.

Joe retireds in 2006 with 1,200,000 and decides on a 4% withdrawal rate. 95% chance of success.

Is Joe's chance of succcess any different from Jane's? They are both have the same sized stash and the same withdrawal rate. If their chances of success are the same (95%) starting in year 2006 then Jane's cherry picking her good year didn't matter.

You have it right. A lot of people do not understand that FIRECalc is a worst case calculator or understand what that means. Averages are irrelevant. Additional wealth puts you on a different worst case trajectory and it only takes one good year, one lottery winning, or one windfall of any kind to put you on that new trajectory. You don't need to stack up good years before you have a different situation than the original simulation.

Being more conservative than the FIRECalc maximum SWR won't hurt anyone though and there are lots of other valid reasons not ot take a historical simulation result as a precise value. People don't have to understand the mathematical nuances of SWR simulations to have a safe retirement plan.

 

[JohnEyles]

I still don't understand how resetting on good years can in any way impact your initial chance of success that's based on a certain withdrawal rate.

Jane retires in 2005 with 1,000,000 and decides on a 4% withdrawal rate (40,000) that gives 95% chance of success. What a great year, Jane's portfolio in 2006 is now 1,200,000. Jane resets and decides to take 4% from her new stash so now she's living large with 48,000 withdrawal. 4% rate, 95% chance of success.

Joe retireds in 2006 with 1,200,000 and decides on a 4% withdrawal rate. 95% chance of success.

Is Joe's chance of succcess any different from Jane's? They are both have the same sized stash and the same withdrawal rate. If their chances of success are the same (95%) starting in year 2006 then Jane's cherry picking her good year didn't matter.

Sounds good to me. Not to mention (assuming they retired at the same age, and
ignoring gender discrepancies in longevity), Jane's chance of success is actually a little
better because she'll probably die a year earlier.

 

[rmark]

"A lot of people do not understand that FIRECalc is a worst case calculator or understand what that means."

Actually Firecalc uses the historical gross annual returns for the U.S., ignoring the outcomes of other industrialized nations. Given that virtually all the other countries used by Dimson, Marsh, and Staunton in Irrational Optimists have lower real returns, in some cases with more volatility, its not truly a worst case calculator.

Note also the default investing cost in Firecalc is based on Vanguards index funds, yet these were not available before about 1975, and certainly not popular before 1985, yet it implies it is appropiate to use this with data from prior years, when investing costs were higher, and net returns to the investor are likely smaller.

These returns should be further reduced by the fact that investors tend to invest more late in a bull market, leading to lower returns on dollar weighted basis. I believe Bogle has mentioned this either in his books or in one of his online speeches.

Borrowing from Willam Bernstein, who suggests that you end up about 2% below average from a mostly stock portfolio, and 1% if mostly bonds, due to volatility , we end up with something like this -

Past U.S. stock real return = 7% - 2% = 5%
Past U.S. bond real return = 3% - 1% = 2% so a 60/40 mix = the famous 4%

Past U.S. stock real return = 7% - 2% - 2% costs = 3%
Past U.S. bond real return = 3% - 1% - 1% costs = 1% so our 60/40 = 2%

Reduced further for international returns, dollar weighting, more conservative retiree 40/60 portfolios = near 0%

The developed nations have had about a 2% real growth rate over the last 200 or so years, with net savers making a bit more, say 3%, and net sellers a bit less , say 1%, which is taxed on nominal return, netting about 0%.

 

[Cb]

Just be careful you don't kill yourself trying to make a living, rmark.

Cb 8)

 

[rmark]

I work in a university library storage facility, and some days I'm kind of like the Maytag repairman. Except we get online subscriptions to current financial periodicals. (and a lot of stuff on cows, engineering, etc)

 

[Animorph]

I agree you can reset your conventional n% inflation adjusted WR IF you are using the 100% safe SWR.

If you are using 95% (or lower), you do risk having the market go down just after you adjusted and being in that bad 5% unless you readjust downward.

Dan

 

[Hydroman]

Reduced further for international returns, dollar weighting, more conservative retiree 40/60 portfolios = near 0%

The developed nations have had about a 2% real growth rate over the last 200 or so years, with net savers making a bit more, say 3%, and net sellers a bit less , say 1%, which is taxed on nominal return, netting about 0%.

So I guess what you are saying is that ER/SWR is a fools game. This forum should be taken down and we all need to save up enough to live on principle for the rest of our lives because there is no such thing as a positive ROI. I always thought that if retirement was impossible I would like to work in a Library. Are you hiring RMark?

 

[rmark]

By late life, with conservative portfolios, yes I suspect 0% returns. Thats what my reading suggests. I think you may be just dividing your money over your remaining lifetime.

Arnott “Disentangling value and size” article
Arnott, Casscells “Demographics and capital market returns” paper
Cooley, Hubbard, Walz “Sustainable withdrawals from your retirement portfolio”
Dichev “What were stock investors actual historical returns?” paper
Dimson, Marsh, Staunton “Irrational optimism” paper
Ervin “Shortfall risk, asset allocation, and over funding a retirement account” paper
Frazzini “Dumb money: mutual fund flows and the cross-section of stock returns”
Grossman, Shore “Cross section of stock returns before WW1” paper
Jorion “Long term risks of global stock markets” paper
Jorion, Goetzmann “Global stock markets of the 20th century” paper
Lewellen “Predicting returns with financial ratios” paper
Milevsky “Retirement ruin and the sequencing of returns” paper
Taleb “Focus on the exceptions that prove the rules” article

 

[sgeeeee]

"A lot of people do not understand that FIRECalc is a worst case calculator or understand what that means."

Actually Firecalc uses the historical gross annual returns for the U.S., ignoring the outcomes of other industrialized nations. Given that virtually all the other countries used by Dimson, Marsh, and Staunton in Irrational Optimists have lower real returns, in some cases with more volatility, its not truly a worst case calculator.

Actually, it is. It takes the data for S&P500, bond and CPI and looks for the worst case historical retirement survival sequence. Thus, it is by definition a worst case calculator. You can argue about whether fixed expense ratio, or any other data within it are appropriate, but the simulation portion of the calculator is a worst case engine.

These returns should be further reduced by the fact that investors tend to invest more late in a bull market, leading to lower returns on dollar weighted basis. I believe Bogle has mentioned this either in his books or in one of his online speeches.

How you use the results from FIRECalc is, of course, open for debate. If you choose to trade, invest in additional asset classes, or pay high fees and trading costs, your mileage will vary. But, again, that does not change the fact that FIRECalc is a worst case simulation engine applied to a specific set of data.

Borrowing from Willam Bernstein, who suggests that you end up about 2% below average from a mostly stock portfolio, and 1% if mostly bonds, due to volatility , we end up with something like this -

Past U.S. stock real return = 7% - 2% = 5%
Past U.S. bond real return = 3% - 1% = 2% so a 60/40 mix = the famous 4%

That's not how everyone reads William Bernstein and that is not where the 4% number comes from. The Trinity Study and Guyton are the ones that first published historical worst case simulator results that provided the 4% results. If your averages were applied to the problem, then the 4% withdrawal rate would preserve net egg value forever. That is not the case for the Trinity Study results which looked at specific retirement lengths up to 30 years. Guyton, I believe, extended the studies to 40 years.

To some extent, none of this matters for the typical retiree. The 4% number should be viewed as a guideline, not a rigid rule.

 

[Rich_by_the_Bay]

I agree you can reset your conventional n% inflation adjusted WR IF you are using the 100% safe SWR.

If you are using 95% (or lower), you do risk having the market go down just after you adjusted and being in that bad 5% unless you readjust downward.

Exactly. You can reset any time you want and will be just like someone else just starting out on FIRE. But if you do it too frequently and only do so for the UP years, you can run into trouble as you state. Do it for both up and down years and you are back to pcnt of assets strategy.

 

[ESRBob]

Here's how I'm understanding it now -- whenever you reset, that is like wiping the board clean -- you reset around that portfolio value and whatever shenanigans you've done in years past are no longer relevant. You have x in the portfolio and you start setting your withdrawals accordingly, raising them by inflation etc over the following years.

If you reset again relatively frequently, you'll never have a chance to find out whether your approach was going to be safe or not for the 30 or 40 year timeframe.

It is only when you really follow one of these annual-inflation-raise plans for a long time that you find out if it is safe or not, and by then it's too late to do anything about it if you are in a worst case scenario. The safety was compromised by having given yourself the full inflation raises over the intevening years, even when that meant you were withdrawing substantially more than 4% of that year's portfolio value.

If you never or rarely go beyond the 4% of portfolio value (akin to resetting every year) then my research shows you've got a very high chance of keeping the portfolio intact in real terms over the medium and long runs.
In other words, resetting is not the problem -- the problem is in continuing to give yourself annual inflation raises during a period of high inflation and low market return, when you are in effect withdrawing at much higher-than-4% rates.

 

[Nords]

To some extent, none of this matters for the typical retiree. The 4% number should be viewed as a guideline, not a rigid rule.

Oh, great, that'll start a new thought process:

"But... but... but... if I don't follow anyone's rigid rules and I run out of money, then it's my own fault and I can't sue anyone!!
Maybe I should evade all personal financial responsibility keep working for another year or two to make sure I have enough... yeah, that's it... just three or four more and then I'm really done..."

 

[sgeeeee]

Oh, great, that'll start a new thought process:

"But... but... but... if I don't follow anyone's rigid rules and I run out of money, then it's my own fault and I can't sue anyone!!
Maybe I should evade all personal financial responsibility keep working for another year or two to make sure I have enough... yeah, that's it... just three or four more and then I'm really done..."

I think you should go back to work, Nords. Now, if you don't follow this advice, you can't sue me.

 

[landover]

I still don't understand how resetting on good years can in any way impact your initial chance of success that's based on a certain withdrawal rate.

Jane retires in 2005 with 1,000,000 and decides on a 4% withdrawal rate (40,000) that gives 95% chance of success. What a great year, Jane's portfolio in 2006 is now 1,200,000. Jane resets and decides to take 4% from her new stash so now she's living large with 48,000 withdrawal. 4% rate, 95% chance of success.

Joe retireds in 2006 with 1,200,000 and decides on a 4% withdrawal rate. 95% chance of success.

Is Joe's chance of success any different from Jane's? They are both have the same sized stash and the same withdrawal rate. If their chances of success are the same (95%) starting in year 2006 then Jane's cherry picking her good year didn't matter.

I think that if Jane decides to reset her withdrawal rate to 4% of rising portfolio at some point she will not be able to reset to higher withdrawal. After n year(s) of rising portfolio there will be down year(s) and during those down years and after that she will never be able to raise her withdrawal rate as portfolio will never reach 25x of inflation adjusted number of nth year's withdrawal. Basically Jane is starting at the worst start year of firecalc simulation on year n+1. This can be tested with modified firecalc/excel.

This means one can keep on resetting upwards as long as possibly. Of course assuming future is like past.

 

[jdw_fire]

Basically Jane is starting at the worst start year of firecalc simulation on year n+1. This can be tested with modified firecalc/excel.

This means one can keep on resetting upwards as long as possibly. Of course assuming future is like past.

If she starts at "the worst start year of firecalc simulation " and the sim has a 95% success probability then she probably picked one of the 5% failures. Of course this depends on how the market does in reality, because maybe there is no starting year that will fail in her 30 yr period. However picking "the worst start year" to start her retirement makes it certian that if there is a start year that produces a failure she picked it.

 

[WRBT]

and poor Joe, who had managed to accumulate a much higher net worth before retiring and takes the same conservative 4% withdrawal rate... goes down in flames along with Jane, victim of that remaining 5%