failure years

[mathjak107]

so looking at firecalcs results a 95% success rate would allow up to 5 failures to happen .

but the 5 worst failures were 1907,1929 ,1937 ,1965,1966 . except for 1907 those dates are what the 4% safe withdrawal rate is based on .

so when firecalc says you have a 95% success rate is it eliminating the very dates the 4% safe withdrawal rates is based on ?

[Senator]

Quote:

Originally Posted by mathjak107 so looking at firecalcs results a 95% success rate would allow up to 5 failures to happen . but the 5 worst failures were 1907,1929 ,1937 ,1965,1966 . except for 1907 those dates are what the 4% safe withdrawal rate is based on . so when firecalc says you have a 95% success rate is it eliminating the very dates the 4% safe withdrawal rates is based on ?

I think it just takes your input, applies that to each year since the beginning of its data, and applies the returns that the market got that year.

No magic, no skipping, just going through each years returns and applying them. Then subtracting your expenses.

[mathjak107]

correct that is exactly how it does it .

but if 95% success has 5 failure periods it has to be the 5 worst periods above .

but those are the dates the 4% rule is based on . if we say a 95% success rate is acceptable then it seems to me we are accepting failures in the time frames the 4% safe withdrawal rate concept was supposed to protect us from .

[bingybear]

Quote:

Originally Posted by mathjak107 so looking at firecalcs results a 95% success rate would allow up to 5 failures to happen . but the 5 worst failures were 1907,1929 ,1937 ,1965,1966 . except for 1907 those dates are what the 4% safe withdrawal rate is based on . so when firecalc says you have a 95% success rate is it eliminating the very dates the 4% safe withdrawal rates is based on ?

If you ran a standard 30 year case and lumped retirement funds at the beginning, is that not what one would expect?
I'm not sure about 95% success allow up to 5 failures unless the number of cases is 100 (or very close to that) as 5% would be expected.
Now if you run a more involved case where pensions are larger or inherited $ come in a different times, you may see different results.
Consider the 30 something RE who might use 60 or 70 year retirements. I would not expect 1966 to fail with a 60 to 70 year retirement run on firecalc as there is not enough data to back test using a 1966 RE date and 60 or more years.

[bingybear]

Quote:

Originally Posted by mathjak107 correct that is exactly how it does it . but if 95% success has 5 failure periods it has to be the 5 worst periods above . but those are the dates the 4% rule is based on . if we say a 95% success rate is acceptable then it seems to me we are accepting failures in the time frames the 4% safe withdrawal rate concept was supposed to protect us from .

What were the parameters for your firecalc run? Were they in line with the assumptions of the 4% rule?

[mathjak107]

it is 116 cycles . a 50/50 mix up to 70/30 at 4% withdrawal rate shows 6 failed cycles at 94.8% success .

that can only be the 6 worst cycles of which 5 of them would be the worst case scenarios to date. i have to come down to 3.50% withdrawals to get 100% and pass all worst case scenarios .

this is strictly based on withdrawals from portfolio that are needed , no ss or pensions included since that part is covered , but looking at just what the portfolio needs to generate it seems to me that a 95% acceptable success rate is allowing you to fail the worst case scenario's the 4% rule was benchmarked by .

[RunningBum]

Quote:

Originally Posted by bingybear Consider the 30 something RE who might use 60 or 70 year retirements. I would not expect 1966 to fail with a 60 to 70 year retirement run on firecalc as there is not enough data to back test using a 1966 RE date and 60 or more years.

1966 won't even be part of the calculation for another 10 or 20 years. That's why we've been seeing a higher success rate for 50 years over 30 or 40, because 1965 and 66 weren't considered. Now that they pop up, the 50 year success rate will go down. So will the 60 and 70 year retirements eventually. It doesn't matter what happens in the next 10-20 years, once you've hit zero, you can't recover. I suppose you could, but Firecalc doesn't account for borrowing money to get back and stay above water.

[ERD50]

I don't understand the question.

95% success does mean 5% failures (the 6 failed of 116 cycles at 94.8% success you mentioned). What else can it mean? To not see failures, you need to go to 100% success.

Maybe I'm missing your intent, as that is just what the words imply. How do you interpret it any other way?

And yes, I recall ~ 3.5% WR (or less) was required to hit 100% historical success. Some people feel that 95% success (or even less) is a good enough approach, so that's what they go with. I feel better with 100%, at least I can say that my plan survived the worst in history. But assuming we don't hit that, the people spending more will have optimized it better, but we can't know that ahead of time. Just depends on yyour comfort level, and if you could make deep cuts to your spending (some budgets have a lot of very discretionary spending).

-ERD50

[mathjak107]

Quote:

Originally Posted by RunningBum 1966 won't even be part of the calculation for another 10 or 20 years. That's why we've been seeing a higher success rate for 50 years over 30 or 40, because 1965 and 66 weren't considered. Now that they pop up, the 50 year success rate will go down. So will the 60 and 70 year retirements eventually. It doesn't matter what happens in the next 10-20 years, once you've hit zero, you can't recover. I suppose you could, but Firecalc doesn't account for borrowing money to get back and stay above water.

1966 ended with the group in 1996 and should be included already in a 30 year time frame

[mathjak107]

Quote:

Originally Posted by ERD50 I don't understand the question. 95% success does mean 5% failures (the 6 failed of 116 cycles at 94.8% success you mentioned). What else can it mean? To not see failures, you need to go to 100% success. Maybe I'm missing your intent, as that is just what the words imply. How do you interpret it any other way? And yes, I recall ~ 3.5% WR (or less) was required to hit 100% historical success. Some people feel that 95% success (or even less) is a good enough approach, so that's what they go with. I feel better with 100%, at least I can say that my plan survived the worst in history. But assuming we don't hit that, the people spending more will have optimized it better, but we can't know that ahead of time. Just depends on yyour comfort level, and if you could make deep cuts to your spending (some budgets have a lot of very discretionary spending). -ERD50

the question is if the supposed 4% rule was supposed to last throught the worst case scenario's then why is 95% acceptable if those worst case scenario's are the ones that failed .

[RunningBum]

The claim is that the 4% rule came about to protect against even the very worst years such as 1929 and 1966. But, a 4% WR fails when starting your retirement in those years.


I'm not sure the 4% rule ever was claimed to be as bullet-proof as the OP says it was. The Wiki for Bengen, who came up with the 4% rule, says

Quote:

Based on his early research of actual stock returns and retirement scenarios over the past 75 years, Bengen found that retirees who draw down no more than 4.2 percent of their portfolio in the initial year, and adjust that amount subsequent every year for inflation, stand a great chance their money will outlive them.

"A great chance" isn't the same as "would never fail".


The other probability is that Firecalc uses different assumptions and factors than what Bengen and others used in coming up with a 4% SWR. For example, there is no such thing as a single inflation rate that applies to all. And they probably have different assumptions on how a retiree would invest.


What it comes down to is, "person A said this was guaranteed, and person B's tool says it's not." Person A may or may not have made that guarantee and person B is not bound by it.

[RunningBum]

Quote:

Originally Posted by mathjak107 1966 ended with the group in 1996 and should be included already in a 30 year time frame

Yes, that's my point. It is counterintuitive that a 30 year time frame would have a lower success rate than a 50 year time frame, but the reason is that a 30 year time frame includes 1966, but the 50 year time frame does not yet include that bad start.

[ERD50]

Quote:

Originally Posted by mathjak107 the question is if the supposed 4% rule was supposed to last throught the worst case scenario's then why is 95% acceptable if those worst case scenario's are the ones that failed .

I believe the 4% 'rule' included a 5% failure rate. I don't know for sure, you'd have to look at the original Trinity study, which I think is what you are referring to. I'm pretty sure that's why FIRECalc defaults to 95% as well.

edit (see Running Bum's answer as well, I cross posted with that)

-ERD50

[MBSC]

The FireCalc portfolio tab defaults to long-term corporate bonds.

Quote:

William Bengen used intermediate-term government bonds, which allows for a 100% success rate for the inflation-adjusted 4% withdrawal rate rule. Meanwhile, the Trinity study uses more volatile long-term corporate bonds, which caused the maximum inflation-adjusted withdrawal rate for new retirees in 1965 and 1966 to dip below 4%.

Source: https://www.bogleheads.org/wiki/Trinity_study_update

Quote:

While Mr. Bengen’s original research combined the S&P 500 index with 5-year intermediate term government bond returns, the Trinity Study used long-term high-grade corporate bond returns instead. With more volatile corporate bonds, the sustainable withdrawal rate dipped slightly below 4% in 1965 and 1966. This led people to hear that the 4% rule has a 95% chance for success, though that is only the historical data, and it does not imply that today’s retirees will enjoy the same odds for success. Nonetheless, a 95% chance for success has a special meaning in statistics about providing a good level of confidence, as so it is easy to misunderstand this point.

Source: Safe Withdrawal Rates for Retirement and the Trinity Study

[bingybear]

Quote:

Originally Posted by RunningBum 1966 won't even be part of the calculation for another 10 or 20 years. That's why we've been seeing a higher success rate for 50 years over 30 or 40, because 1965 and 66 weren't considered. Now that they pop up, the 50 year success rate will go down. So will the 60 and 70 year retirements eventually. It doesn't matter what happens in the next 10-20 years, once you've hit zero, you can't recover. I suppose you could, but Firecalc doesn't account for borrowing money to get back and stay above water.

the bold text was my point to the OP. Without knowing the case being run... lots of observations can be assumed.

I do like your insight as to why we see higher success on long retirement runs. I've used 50 years for quite a while. Firecalc is not the only calculator I use, so not too worried about the usefulness of the predictions.

[mathjak107]

so the bottom line is over a 30 year time frame the 4% rule is more an acceptable withdrawal rate then a safe one . considering it failed at least 5 of our worst time frames i wouldn't call that exactly safe . especially since already the y2k retiree is on par with the 1929 retiree 15 years in and that was a failure point for that group starting in 1929.

if anything the really safe withdrawal rate should be called 3.50% and 4% an acceptable withdrawal rate .

[mathjak107]

found a blog with a very good explanation

What is Your Retirement Number - The 4% Rule - Go Curry Cracker!Go Curry Cracker!

[DrRoy]

But isn't it the case that unless the future is a mirror of the past, then the Firecalc number is still just a guess? What if markets cannot match the growth of the last 116 years? Also, growth is easier from a smaller base. For example, one reason China cannot grow as fast as before is that it's economy is so much larger now that the same growth rate as in the past cannot be sustained.

[RunningBum]

I think a more likely scenario is that one will miscalculate expenses. Forget to account for house maintenance or car replacement, have a bounce-back kid, underestimate out of pocket medical expenses or the increase in insurance rate hikes, and your problem becomes spending way more than 4% rather than whether 4% will last you.

[ERD50]

Quote:

Originally Posted by DrRoy But isn't it the case that unless the future is a mirror of the past, then the Firecalc number is still just a guess? What if markets cannot match the growth of the last 116 years? Also, growth is easier from a smaller base. For example, one reason China cannot grow as fast as before is that it's economy is so much larger now that the same growth rate as in the past cannot be sustained.

Unless you have a proven Crystal Ball, what other options do you have?

I would not go so far as to say it's a 'guess' - anyone can guess. It is a number derived from past cycles. Look at the description on the FIRECalc page, with their analogy to temperatures - we don't guess that it will be hot in the summer and cold in the winter here in the Midwest, it's based on historical cycles. Yes, we can get lower lows and higher highs, records are made to be broken.

And the 4% 'rule' isn't based on the 'growth of the last 116 years', it is based on the worst cycles in those 116 years. In most other cycles, the 4% is way conservative. IIRC, even 6% survives 50% of the time.

-ERD50