Author Topic: 150 years of retirement plan outcomes  (Read 6503 times)

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
150 years of retirement plan outcomes
« on: May 27, 2017, 08:21:23 AM »
In a handful of posts over last few weeks, people have discussed or mentioned the financial risk of equity investments and discussed how challenging it is to understand the risk baked into a savings plan and an asset allocation.

I'm hoping the graphic below (comes from some writing I've been doing) will contribute to topic. The red line in the line chart shows how much someone would accumulate at the end of a 35 year, $5500 a year savings program using a 75% stocks and 25% bonds allocation any time over the last roughly 150 years. The first scenario modeled, for example, starts in 1872...


The thing that's important to take away from line charts like this is that outcomes vary widely around that median outcome (shown as that dark black line) that many of us focus on. The median outcome for this scenario is around $600K for example. But the range of possibilities runs from $200K to $1M roughly. The standard deviation is close to $200K.

That's a lot of variability.

Tip: One can gain great insights into the variability of one's own savings plan using http://www.cfiresim.com/ Note that I used cfiresim to create the data that "sits" behind the line chart.

A final note: Your personal line chart will look more or less jagged depending on your asset allocation formula.


Clean Shaven

  • Pencil Stache
  • ****
  • Posts: 821
  • Location: Wild Wild West
Re: 150 years of retirement plan outcomes
« Reply #1 on: May 27, 2017, 09:23:31 AM »
What comprises the stock and bond parts of the graph data? SP500, total US, mix with international?

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #2 on: May 27, 2017, 10:55:10 AM »
What comprises the stock and bond parts of the graph data? SP500, total US, mix with international?

The cFIREsim FAQ says this: "...historical stock/bond/gold/inflation data from 1871 to present..." which is little vague but in an interview Bo the creator indicates that the stock market data come from Shiller's US stock market data:

http://www.econ.yale.edu/~shiller/data.htm

BTW, you get nearly the same results if you use FIRECalc... The numbers don't seem to match perfectly, but they get pretty close.

And speaking of FIRECalc, it sounds like the FIRECalc author needed to finesse things a little bit in order to deal with bond interest returns in the early years. At the FIRECalc website, they go into the machinations necessary to come up with bond returns near the bottom of this page:

http://firecalc.com/intro.php

respond2u

  • Stubble
  • **
  • Posts: 119
Re: 150 years of retirement plan outcomes
« Reply #3 on: May 29, 2017, 07:35:49 PM »
Thanks Seattle--that's a pretty shocking difference in results! The reverse of saving is spending, so I bet if you were to chart out portfolio depletion for retirees, longevity would be opposite...

Did you plot that against CAPE by any chance? I would expect there to be a fair amount of correlation (saving in high CAPE times isn't worth much).

I guess the only other response I can think of is that I never experienced a steady "5500/year" time for savings. It varied quite a bit depending on whether checks came in December or January, though generally went up as I earned more.

sokoloff

  • Handlebar Stache
  • *****
  • Posts: 1191
Re: 150 years of retirement plan outcomes
« Reply #4 on: May 29, 2017, 07:56:14 PM »
Using some simplistic assumptions, for an end portfolio value of $300K, $600K, and $900K under those base conditions (and all contributions at the "end" of the year), I get CAGRs of 2.46%, 5.90%, and lucky 7.77%

Those are lower than I'd have thought, but those were also 75/25 portfolios.
Stephen, do you a similar chart for 100/0 (all equity) portfolios?

https://docs.google.com/spreadsheets/d/1nLlpdjhebYoNnfLpd1labFk8bdBf4RaGC220e0KH6mM/edit?usp=sharing (cell E2 and F40)

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #5 on: May 30, 2017, 07:10:12 AM »
Sokoloff, here's link to post what describes steps for doing math with cfiresim: http://evergreensmallbusiness.com/cfiresim-retirement-plan-b-calculations/

BTW I didn't save 100% stocks, but I can pretty confidently predict you end up with a higher median result and a bigger standard deviation. And just FYI, the 2.46% result roughly equals what someone who retired in 1981 experienced over their accumulation years and that 7.77% result is about what someone who retired in 1964 experienced over their accumulation years.

Respond2u, I agree the results shock. Also, I think you're probably right about those above median and below median results reversing themselves over retirement. E.g., if you retired in 1999, you were looking at an excellent outcome (roughly $1M) for a $5500 a year for 35 years retirement plan. I calculate the average return for that person to be 8.28% (that's a real return... yikes). But you would think at the end of 30 years, that retiree likely won't have seen her or his good fortune continue. For one thing this retiree's first financial experience was the tech crash.

Sharing a perspective like this can make me seem like Debbie Downer, but I suggest we consider this variability in our planning. It represents a powder keg for people's retirement plans. Especially since in many ways, we've had a great two decade run and so people are desensitized to the issue.

P.S. I've got a table of percentile ranks at the backgrounder blog post I did about this last week which might provide further perspective:

http://evergreensmallbusiness.com/retirement-plan-b-need-one/

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #6 on: May 30, 2017, 07:36:14 AM »
This isn't a chart, but a quick calculation of a 100% stocks allocation over 35 years with a $5500 a year "IRA" contribution  (and then the other tweaks to get cfiresim to do the math) delivers following results:

Average:   $766,731
Median:   $696,362
St. Dev.:   $293,929
Highest:   $1,535,966
Lowest:   $250,852

Here's same data for the 75% stocks and 25% bonds (though I think I used Excel rather than cfiresim calculated these values so they may not match what someone calculates with cfiresim)

Average   $614,262
Median   $620,358
StDev   $179,489
Max         $1,007,843
Min      $231,314 

Heroes821

  • Pencil Stache
  • ****
  • Posts: 604
Re: 150 years of retirement plan outcomes
« Reply #7 on: May 30, 2017, 09:26:30 AM »
This isn't a chart, but a quick calculation of a 100% stocks allocation over 35 years with a $5500 a year "IRA" contribution  (and then the other tweaks to get cfiresim to do the math) delivers following results:

Average:   $766,731
Median:   $696,362
St. Dev.:   $293,929
Highest:   $1,535,966
Lowest:   $250,852

Here's same data for the 75% stocks and 25% bonds (though I think I used Excel rather than cfiresim calculated these values so they may not match what someone calculates with cfiresim)

Average   $614,262
Median   $620,358
StDev   $179,489
Max         $1,007,843
Min      $231,314

I find that amazing that the low with stock and bonds (most people think it's safer) was lower than pure stocks.
« Last Edit: May 30, 2017, 09:47:38 AM by Heroes821 »

sokoloff

  • Handlebar Stache
  • *****
  • Posts: 1191
Re: 150 years of retirement plan outcomes
« Reply #8 on: May 30, 2017, 09:42:51 AM »
Sharing a perspective like this can make me seem like Debbie Downer, but I suggest we consider this variability in our planning. It represents a powder keg for people's retirement plans. Especially since in many ways, we've had a great two decade run and so people are desensitized to the issue.
On the contrary, I find this work fascinating and hugely valuable.

Pretending that drawdowns and bear markets don't happen is utter folly.

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #9 on: May 30, 2017, 10:11:14 AM »
This isn't a chart, but a quick calculation of a 100% stocks allocation over 35 years with a $5500 a year "IRA" contribution  (and then the other tweaks to get cfiresim to do the math) delivers following results:

Average:   $766,731
Median:   $696,362
St. Dev.:   $293,929
Highest:   $1,535,966
Lowest:   $250,852

Here's same data for the 75% stocks and 25% bonds (though I think I used Excel rather than cfiresim calculated these values so they may not match what someone calculates with cfiresim)

Average   $614,262
Median   $620,358
StDev   $179,489
Max         $1,007,843
Min      $231,314

I find that amazing that the low with stock and bonds (most people think it's safer) was lower than pure stocks.

I think people work with different definitions of the concepts of "safe" and "safer"...

E.g., if by unsafe or risky you or I mean worst case, a 75% stocks and 25% bonds allocation is riskier because that $231K minimum is less than the $251K minimum.

But if by unsafe or risky you or I mean variability in returns, then 100% stocks is riskier because you're looking at a standard deviation of $179K vs. $294K,

I suspect people need to think about all this carefully because if people aren't looking at the standard deviation or the minimum but rather that median return they actually are not considering the risk...

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #10 on: May 30, 2017, 10:15:38 AM »
Sharing a perspective like this can make me seem like Debbie Downer, but I suggest we consider this variability in our planning. It represents a powder keg for people's retirement plans. Especially since in many ways, we've had a great two decade run and so people are desensitized to the issue.
On the contrary, I find this work fascinating and hugely valuable.

Pretending that drawdowns and bear markets don't happen is utter folly.

I think so too.

DavidAnnArbor

  • Handlebar Stache
  • *****
  • Posts: 2266
  • Age: 58
  • Location: Ann Arbor, Michigan
Re: 150 years of retirement plan outcomes
« Reply #11 on: May 30, 2017, 02:34:04 PM »
what's the probability of 1 standard deviation ?

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: 150 years of retirement plan outcomes
« Reply #12 on: May 30, 2017, 03:01:01 PM »
That's a remarkably low savings rate, for a very long period.  The more relevant chart for most of this audience is probably more like $50k per year for 10 years, since people here are mostly higher than average income paired with ridiculous savings rates.  And for that chart, the variability is likely to be even greater.

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #13 on: May 30, 2017, 03:43:18 PM »
what's the probability of 1 standard deviation ?

Ending up a standard deviation low puts you at just about the 20th percentile and at about $440K in savings at start of retirement according to this scenario.

Ending up a standard deviation high puts you at about the 80th percentile and at about $800K in savings at the start of retirement according to this scenario.

Not super likely... but pretty conceivable... and maybe something people want to plan for (at least on the downside?)

BTW I've got a table of percentiles here about half way down the page: http://evergreensmallbusiness.com/retirement-plan-b-need-one/

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #14 on: May 30, 2017, 03:48:49 PM »
That's a remarkably low savings rate, for a very long period.  The more relevant chart for most of this audience is probably more like $50k per year for 10 years, since people here are mostly higher than average income paired with ridiculous savings rates.  And for that chart, the variability is likely to be even greater.

Sol, here's a quick cFIREsim calculation with $50K a year for 10 years using a 75% stocks and 25% bonds allocation.

Average:   $729,360
Median:   $716,759
St. Dev.:   $180,733
Highest:   $1,403,655
Lowest:   $361,992

The variability here is pretty striking. You could actually end up with less money that you put in (10 years of $50K could end up as $361K.)

P.S. Probably be good to run your own numbers... and then double-check using FIRECalc.

daverobev

  • Magnum Stache
  • ******
  • Posts: 3961
  • Location: France
Re: 150 years of retirement plan outcomes
« Reply #15 on: May 31, 2017, 11:24:16 AM »
I guess really what's needed is not two separate things (did the market do well while you were saving, and, did the market do well while you were retired), but a single, start-of-saving-til-death thing?

If you had a shitty set of returns before retirement, chances are you'll see better than average returns while retired. And the converse; if you had great returns before retiring, you've got a good chance of lower returns after - it makes sense, as things smooth out over time, right?

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #16 on: May 31, 2017, 11:57:42 AM »
I guess really what's needed is not two separate things (did the market do well while you were saving, and, did the market do well while you were retired), but a single, start-of-saving-til-death thing?

If you had a shitty set of returns before retirement, chances are you'll see better than average returns while retired. And the converse; if you had great returns before retiring, you've got a good chance of lower returns after - it makes sense, as things smooth out over time, right?

Daverobev, that seems/sounds right to me... But I personally think this is surprisingly tricky to get a good handle on.

My own sense is one wants to have both the psychological and financial flexibility to handle a situation that differs from the median outcome which I tend to focus on.

And this point about financial risk: It seems to me as if someone wants to always look beyond the expected return or future value and at least quantify the variability associated with one's asset allocation.

Obvious points from my blog post about this: The workarounds include saving more, working longer, compounding interest longer, earning a higher return, and spending less during retirement... but then don't forget the just as real possibility which is that you could end up an outcome that's way above average.


Esteban

  • 5 O'Clock Shadow
  • *
  • Posts: 5
Re: 150 years of retirement plan outcomes
« Reply #17 on: May 31, 2017, 01:15:08 PM »

And this point about financial risk: It seems to me as if someone wants to always look beyond the expected return or future value and at least quantify the variability associated with one's asset allocation.

  ... but then don't forget the just as real possibility which is that you could end up an outcome that's way above average.

Interesting graph. The takeaway I get is that volatility in a bad drawdown could cause you to "crap out", AKA the Gambler Ruin Problem, i.e., once you are down sufficiently you no longer have the capital to "keep playing", which necessarily means you cannot participate in future updrafts.

Howard Marks of Oaktree defines risk as the possibility of permanent loss. I redifine it in terms of a drawdown so massive that it takes an excruciating long time to build back up. That's what we wish to avoid, no?

http://www.zerohedge.com/news/2014-09-05/oaktrees-howard-marks-explains-difference-between-volatility-and-risk

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: 150 years of retirement plan outcomes
« Reply #18 on: May 31, 2017, 02:19:55 PM »

And this point about financial risk: It seems to me as if someone wants to always look beyond the expected return or future value and at least quantify the variability associated with one's asset allocation.

  ... but then don't forget the just as real possibility which is that you could end up an outcome that's way above average.

Interesting graph. The takeaway I get is that volatility in a bad drawdown could cause you to "crap out", AKA the Gambler Ruin Problem, i.e., once you are down sufficiently you no longer have the capital to "keep playing", which necessarily means you cannot participate in future updrafts.

Howard Marks of Oaktree defines risk as the possibility of permanent loss. I redifine it in terms of a drawdown so massive that it takes an excruciating long time to build back up. That's what we wish to avoid, no?

http://www.zerohedge.com/news/2014-09-05/oaktrees-howard-marks-explains-difference-between-volatility-and-risk

I maybe see this differently... or possibly I'm saying the same thing differently... but to me the "answer" is to look beyond the median return or median future value. We want to recognize--whatever your or my number is--that the actual number might will surely vary quite a bit from forecasted number.

Another way to say this same thing: " Yes, yes, I'm saving money for FIRE... and given the inputs i'm using, sure I get a median forecast of $600K (say) but I expect to end up within this range of outcomes that runs from $500K to $700K (or whatever)... and I'm okay with that..."

In the blog posts I did, I assumed (using my example numbers) that people's plan "a" equaled $600K... but then needed a plan "b" for the $500K outcome.

Caveat: Your numbers will vary. The above numbers are for example only.