4% rule gone for good?

[pb4uski]

Quote:

Originally Posted by David1961 +1. The article lists 1966 as one of the worst times (if not the worst time) to retire. Since we are at about the 50 year mark, it'd be interesting to see what the 50 year SWR would have been for someone retiring in 1966. I imagine that's built into Firecalc, though.

Quote:

Originally Posted by REWahoo Yes it is.

Assuming that is the worst period it looks like a 3.15% WR would have still succeeded. ($31,500 initial withdrawal, $1m portfolio, 49 year time horizon, all other assumptions default gives 100% success).

[mathjak107]

i find knowing that we need about a 2% real return average the first 15 years of a retirement gives us a pretty good guide as to where we stand.

you know that y2k retiree who is under 2% better cut back withdrawals well in advance of failure.

that correlation between failures and numbers means historical data is not historical data that has to play out to be true.

if 8 plus 2 is 10 and 9 plus 1 is 10 and 6 plus 4 is ten all we need to know is if we see the number 10 some event will happen regardless of how the sum is arrived at .

so anytime we see less than 2% real return a red flag should go up.

[papadad111]

Quote:

Originally Posted by audreyh1 The problem is there are far fewer 50 year periods so the statistical sample is much smaller.

Yes that is the problem in trying to determine a safe 50 year withdraw rate. Also the data that we DO have is USA only and the economy is so much more global that it ever was even in the 30 year sample periods. So .... I wonder mathematically if one can string together a "series of worst returns" to simulate a worst case 50 year period and then determine a safe withdraw rate. I think 3 percent or more may be too high.

Something on order of 2.3-2.7 percent maybe? . Of course the likelihood of some real black swan happening also is higher as is the likelihood of winding up with a huge pot of dough at the end .... Guess best is Variable withdraw after the first 15 or 20 years ( a " higher risk" period).

[2017ish]

Quote:

Originally Posted by papadad111 Yes that is the problem in trying to determine a safe 50 year withdraw rate. Also the data that we DO have is USA only and the economy is so much more global that it ever was even in the 30 year sample periods. ....

We do have data from other markets; the 20th century in the USA markets was a golden era. The Shocking International Experience of the 4% Rule (The paper discussed in the link is the first I read from Wade Pfau.)

[David1961]

Quote:

Originally Posted by papadad111 Yes that is the problem in trying to determine a safe 50 year withdraw rate. Also the data that we DO have is USA only and the economy is so much more global that it ever was even in the 30 year sample periods. So .... I wonder mathematically if one can string together a "series of worst returns" to simulate a worst case 50 year period and then determine a safe withdraw rate. I think 3 percent or more may be too high. Something on order of 2.3-2.7 percent maybe? . Of course the likelihood of some real black swan happening also is higher as is the likelihood of winding up with a huge pot of dough at the end .... Guess best is Variable withdraw after the first 15 or 20 years ( a " higher risk" period).

A rule of thumb that I have is that with a 0% real rate of return (your investments just keep up with inflation), a 2% WR should last you 50 years. I'm sure there are scenarios where this may not work, but for me 2% would be a very safe WR for 50 years. Especially if you can adjust your spending if you suffer some bad years at the beginning.
But with this approach, you'd be much more likely to die before you run out of money (and thus could leave a lot of money to your heirs).

[ERD50]

Quote:

Originally Posted by David1961 A rule of thumb that I have is that with a 0% real rate of return (your investments just keep up with inflation), a 2% WR should last you 50 years. ...

To be clear - that is not a 'rule of thumb' it is simple arithmetic. 100/50=2.

However, that simple arithmetic does not take into account sequence of returns.

You can model this in FIRECalc. Set portfolio entry "Start" tab at $1M, period at 50 years, $0 spending. Set 'other spending' tab to $20,000 (2%), no inflation adjustment. Set "Your portfolio" tab to:

A portfolio with random performance, with a mean total portfolio return of 0 % and variability (standard deviation) of 0 %. Assume an inflation rate of 0 %.

And you'll see a straight line decline to $0 on year 50, as expected.

Now change the standard deviation to a mere 10% (to simulate some sequence of returns variability)... and success rate drops to 33%.

Quote:

The lowest and highest portfolio balance at the end of your retirement was $-533,105 to $1,541,546, with an average at the end of $-60,275.

-ERD50

[papadad111]

Quote:

Originally Posted by 2017ish We do have data from other markets; the 20th century in the USA markets was a golden era. The Shocking International Experience of the 4% Rule (The paper discussed in the link is the first I read from Wade Pfau.)

This is just partial data -- for example where is the data for world #2 economy China? Perhaps china is a 21st century story so the data is not included. But can not be ignored...

Such an economic mega shift ( the rise of China ) certainly will alter the forward looking long view of all of these economies and related asset returns (USA included).

[2017ish]

Quote:

Originally Posted by papadad111 This is just partial data -- for example where is the data for world #2 economy China? Perhaps china is a 21st century story so the data is not included. But can not be ignored... Such an economic mega shift ( the rise of China ) certainly will alter the forward looking long view of all of these economies and related asset returns (USA included).

Agree on the difference going forward, if China continues on its upward trajectory. Pfau specifically addressed the omission of China and the survivorship bias of his data:

Quote:

Keep in mind also that I am only looking at the 19 developed market countries in the dataset, with data going back to 1900. Travel back in time to 1900, though, and ask people to put together a list of 19 developed market countries for the 20thcentury, and you would probably find frequent mention of countries like Argentina, Russia, and China, among others. As those countries never made the dataset, even the results I describe here include survivorship bias.

If you could remove that bias, however, the 20th century results would be more dismal....

[audreyh1]

Quote:

Originally Posted by papadad111 Yes that is the problem in trying to determine a safe 50 year withdraw rate. Also the data that we DO have is USA only and the economy is so much more global that it ever was even in the 30 year sample periods. So .... I wonder mathematically if one can string together a "series of worst returns" to simulate a worst case 50 year period and then determine a safe withdraw rate. I think 3 percent or more may be too high. Something on order of 2.3-2.7 percent maybe? . Of course the likelihood of some real black swan happening also is higher as is the likelihood of winding up with a huge pot of dough at the end .... Guess best is Variable withdraw after the first 15 or 20 years ( a " higher risk" period).

My instincts (based on prior reading/discussion) are that around 3.3% is probably a "perpetual" number that would reasonably survive 50 years or even indefinitely.

But it is best to have a response plan and be able to cut back if one finds oneself on one of the a "worst case" trajectories.

Otherwise, being too conservative may mean depriving oneself during retirement and leaving a lot behind. If that is a goal - OK.

My instinct is that dropping below 3% especially if you are doing % of remaining portfolio will leave a huge amount on the table after you die. But of course I can't prove that.

[daylatedollarshort]

Quote:

Originally Posted by ERD50 To be clear - that is not a 'rule of thumb' it is simple arithmetic. 100/50=2. However, that simple arithmetic does not take into account sequence of returns. You can model this in FIRECalc. Set portfolio entry "Start" tab at $1M, period at 50 years, $0 spending. Set 'other spending' tab to $20,000 (2%), no inflation adjustment. Set "Your portfolio" tab to: A portfolio with random performance, with a mean total portfolio return of 0 % and variability (standard deviation) of 0 %. Assume an inflation rate of 0 %. And you'll see a straight line decline to $0 on year 50, as expected. Now change the standard deviation to a mere 10% (to simulate some sequence of returns variability)... and success rate drops to 33%. -ERD50

I think with this methodology, the Boglehead wiki refers to it as liability matching (my husband calls it the TIPS and a beach condo (or in general low baseline expenses) plan), the idea is to generally invest in inflation adjusted, low volatility asset classes, particularly I bonds and TIPS ladders, with a zero or greater real yield, always holding the bonds to maturity.

With even a zero real return over a 40 year retirement horizon, one could have a SWR of 2.5%, and real yields, though still low by historical standards for TIPS, currently are higher than zero for TIPS with a maturity date greater than 5 years:

United States Government Bonds - Bloomberg

This means, depending on your rolling average real yields, your SWR might be 3%+ with no worries except in the case of a government collapse or Treasury bond default.

[David1961]

Quote:

Originally Posted by ERD50 To be clear - that is not a 'rule of thumb' it is simple arithmetic. 100/50=2. However, that simple arithmetic does not take into account sequence of returns. You can model this in FIRECalc. Set portfolio entry "Start" tab at $1M, period at 50 years, $0 spending. Set 'other spending' tab to $20,000 (2%), no inflation adjustment. Set "Your portfolio" tab to: A portfolio with random performance, with a mean total portfolio return of 0 % and variability (standard deviation) of 0 %. Assume an inflation rate of 0 %. And you'll see a straight line decline to $0 on year 50, as expected. Now change the standard deviation to a mere 10% (to simulate some sequence of returns variability)... and success rate drops to 33%. -ERD50

Good point. It really is math more than a rule of thumb.

[Running_Man]

Quote:

Originally Posted by David1961 +1. The article lists 1966 as one of the worst times (if not the worst time) to retire. Since we are at about the 50 year mark, it'd be interesting to see what the 50 year SWR would have been for someone retiring in 1966. I imagine that's built into Firecalc, though.

Historically 1966 is viewed as the worst year to retire in yet in 1966 10 year bonds yielded 4.6% and the S&P500 yielded 3.5%.

The proponents of total return investing with 10 year yields at 2 percent and S&P500 yield at 1.8% are being given the chance to prove their point over the next 30 years.

[youbet]

The problem with retiring in 1966 was inflation, not yields.

I retired into the Great Recession. Since then, there's been a nice recovery and all seems to be working out (so far). But the folks who retired into that inflationary period have never seen a recovery from it. The impact of the 1966 and subsequent years of inflation are still with us today. The growth of CPI slowed but never deflated and there has never been a recovery in prices since that period.

[haha]

Quote:

Originally Posted by youbet The problem with retiring in 1966 was inflation, not yields. I retired into the Great Recession. Since then, there's been a nice recovery and all seems to be working out (so far). But the folks who retired into that inflationary period have never seen a recovery from it. The impact of the 1966 and subsequent years of inflation are still with us today. The growth of CPI slowed but never deflated and there has never been a recovery in prices since that period.

If by recovery you mean a reversal of the inflation, I don't think that ever happens in modern central bank economies.

No matter what they say, governments like inflation. Look at our federal reserve chairs. They have been are falling all over themselves to jack up inflation, but after all it is actually higher than the world average since 18th century. Tax r3eceipts go up, people feel more flush, there is some redistribution from holders of money to borrowers of money -all things that democratic governments find copacetic.

Ha

[dallas27]

Quote:

Originally Posted by haha If by recovery you mean a reversal of the inflation, I don't think that ever happens in modern central bank economies. No matter what they say, governments like inflation. Look at our federal reserve chairs. They have been are falling all over themselves to jack up inflation, but after all it is actually higher than the world average since 18th century. Tax r3eceipts go up, people feel more flush, there is some redistribution from holders of money to borrowers of money -all things that democratic governments find copacetic. Ha

They like controlled inflation, because deflation freezes all commerce. High inflation is not much better. Moderate inflation is the sweet spot.


Sent from my iPhone using Early Retirement Forum

[haha]

Quote:

Originally Posted by dallas27 They like controlled inflation, because deflation freezes all commerce. Sent from my iPhone using Early Retirement Forum

This is commonly said. But look at England's history from Waterloo to WW1. Perhaps the longest period of excellent economic progress ever experienced, and accompanied by steady gentle deflation. When productivity increases, consumer prices should deflate. THe story of semiconductors over the past ~50 years.

Ha

[Totoro]

Quote:

Originally Posted by dallas27 They like controlled inflation, because deflation freezes all commerce.

I haven't seen as of yet convincing evidence (in reality!) that deflation has a negative effect.

Even the theoretical reasoning employed is shaky at best. For example people supposedly would start postponing consumption with sustained deflation. News flash counterexample: electronics get cheaper every year and people still buy them! Likewise, investment returns mean money today is worth less than tomorrow, and consumption still happens.

What certainly does have negative effects are volatile and/or high inflation/deflation rates. It just works both ways.

I do have high (irrational?) hopes that central banks have become better at playing the inflation/deflation game in the past 100 years. One risk factor less.

[dallas27]

Japan.


Sent from my iPhone using Early Retirement Forum

[ERD50]

Quote:

Originally Posted by daylatedollarshort I think with this methodology, the Boglehead wiki refers to it as liability matching (my husband calls it the TIPS and a beach condo (or in general low baseline expenses) plan), the idea is to generally invest in inflation adjusted, low volatility asset classes, particularly I bonds and TIPS ladders, with a zero or greater real yield, always holding the bonds to maturity. With even a zero real return over a 40 year retirement horizon, one could have a SWR of 2.5%, and real yields, though still low by historical standards for TIPS, currently are higher than zero for TIPS with a maturity date greater than 5 years: United States Government Bonds - Bloomberg This means, depending on your rolling average real yields, your SWR might be 3%+ with no worries except in the case of a government collapse or Treasury bond default.

Thanks for reminding me of that. I'll consider that when I'm further along, for a portion of my portfolio.

-ERD50

[Hamlet]

Quote:

Originally Posted by haha This is commonly said. But look at England's history from Waterloo to WW1. Perhaps the longest period of excellent economic progress ever experienced, and accompanied by steady gentle deflation. When productivity increases, consumer prices should deflate. THe story of semiconductors over the past ~50 years. Ha

I think that it is important to distinguish between deflation caused by increased productivity (like anything electronic), and deflation caused by a collapse in demand/incomes (like a recession). The second kind of deflation destroys any economy with sizable borrowing being done.

Likewise, inflation can be caused by a rise in demand/incomes, or it can be caused by a supply shock that reduces productivity.

Ultimately, it all boils down to real changes in incomes. No one cares about a little inflation if their pay is going up faster than prices. And people don't mind a little deflation as long as their incomes aren't going down too.