Probability of Success at varying % Equity AA

[Midpack]

As a corollary to the earlier Portfolio vs AA thread (link below), though this topic hasn't come up in a while. Again, the exact $ are not the central point, the graphs are meant to compare the relative portfolio required to spend $30K/yr (inflation adjusted) over a 30 year retirement at 90%, 95% and 100% success at various asset allocations. Once again, I used FIRECALC to generate all the numbers. The vertical axis is portfolio value, the horiz axis is % equity and the colored bars are success rates.

If the second chart seems confusing, it's was meant to more clearly show how much more or less you would need to retire at various success rates and asset allocations as compared to a 60:40 AA and a 95% success rate (often the baseline for SWR discussions). For example, if you want to retire with a 100% success rate and % equity - history suggests you'd need 167%, or 67% more than another retiree with a 60:40 AA and 95% success rate. That's a LOT more years working to retire, or dramatically less spending in retirement. We all have to assume some risk, it's just deciding what trade off between risk and when to retire (or how much to spend) we can live with.

Again, I leave it to readers to draw their own conclusions (instead of blathering on further myself).

[For you eagle eyes who note the $ results are slightly different than the earlier thread, it's because the first thread I used the default long interest rate for fixed income, and in this one I used 5 year treasuries at the suggestion of a member on the earlier thread.]

What's different in the charts on both threads is success rates are given, whereas more often charts use portfolio $ as a given. That's why they may seem a little counter intuitive.

http://www.early-retirement.org/forums/f28/portfolio-required-v-equity-asset-allocation-101660.html

 

[audreyh1]

It’s interesting that in these charts for 95% success, green goes to the 60% allocation as requiring the lowest starting portfolio. But for 100% success, 40% equities has the slight edge in terms of lowest starting portfolio.

 

[Midpack]

It’s interesting that in these charts for 95% success, green goes to the 60% allocation as requiring the lowest starting portfolio. But for 100% success, 40% equities has the slight edge in terms of lowest starting portfolio.

Yes, but from a broader perspective, once again the difference at any % success between 40%, 60% and 80% equity is relatively minor - as was the point of the first thread. What varies more dramatically is the range of residual values, not the $ required or the AA.

And as you know, the sensitivity at 100% is extreme - a statistical outlier can dramatically skew 100%. So much so that I wouldn't recommend anyone use it. YMMV

 

[levindb]

To me, it appears that there is not a significant difference in AA choices between 40% equities and 80%...certainly within the margins of error. So, I can just quit reading about everyone's AA and use that found time to do more important things in life, like take a nap.

 

[REWahoo]

To me, it appears that there is not a significant difference in AA choices between 40% equities and 80%...certainly within the margins of error. So, I can just quit reading about everyone's AA and use that found time to do more important things in life, like take a nap.

+1

Figuring out a "good" AA isn't what derails a retirement portfolio, it's maintaining that AA when the fit hits the shan and it gets tough to take that nap.

 

[Midpack]

To me, it appears that there is not a significant difference in AA choices between 40% equities and 80%...certainly within the margins of error. So, I can just quit reading about everyone's AA and use that found time to do more important things in life, like take a nap.

That was the central point in the earlier thread. I repeatedly see people agonizing over minor changes in AA when it takes a pretty big change to make much difference (until you get to the 100% and especially 0% extremes)...

 

[audreyh1]

Yes, but from a broader perspective, once again the difference at any % success between 40%, 60% and 80% equity is relatively minor - as was the point of the first thread. What varies more dramatically is the range of residual values, not the $ required or the AA.

That’s right, in choosing between one of the middle allocations, one should look at annual volatility and worse case drawdown over multiple years, versus
worst case and average size of the residual portfolio. Worst case drawdown with the traditional method is portfolio going to 0 unless 100% success is used. And even then it’s designed to drawdown to close to zero worst case over the period specified anyway.

For some people, residual portfolio matters, or they hope to leave at least X amount behind. For others, residual portfolio is of little importance.

I evaluated different withdrawal rates for the %remaining portfolio case, and starting income and max drawdown were major criteria, but also noted the residual portfolio. I only looked at the 50/50 and 60/40 cases though. Evaluating worst case draw down took identifying the worst year and running multiple individual scenarios, so it was a lot of work using FIREcalc. You never went to 0, but worst case drawdown (inflation adjusted) usually took many many years and resulted in a much smaller portfolio and thus much smaller annual income compared to the starting income.

 

[Midpack]

+1

Figuring out a "good" AA isn't what derails a retirement portfolio, it's maintaining that AA when the fit hits the shan and it gets tough to take that nap.

Absolutely. Hopefully most here have some idea how they'd really react based on 1987, 2000 and/or 2008 - were you able to nap without panic selling? A lot more real than historical blips on charts.

 

[audreyh1]

+1

Figuring out a "good" AA isn't what derails a retirement portfolio, it's maintaining that AA when the fit hits the shan and it gets tough to take that nap.

Absolutely! The sleep at night factor cannot be ignored. How many people here rebalanced at the end of 2008 or during the the first quarter of 2009? How many panicked and took some drastic action that they later regretted? It made Bill Bernstein throw up his hands and completely change his investment advice for retirees since so many of his clients bailed and then didn’t get back in and then were stuck with a much lower portfolio. How many here were caught like deer in the headlights and did nothing? - definitely way better than bailing.

 

[OldShooter]

Absolutely. Hopefully most here have some idea how they'd really react based on 1987, 2000 and/or 2008 - were you able to nap without panic selling? A lot more real than historical blips on charts.

Yes. I think the ubiquitous "risk tolerance assessment" questionnaires found on the internet are completely useless. I have always liked Fred Schwed's discussion of the issue:

There are certain things that cannot be adequately explained to a virgin either by words or pictures. Nor can any description that I might offer here even approximate what it feels like to lose a real chunk of money that you used to own.
​

The first hit is the worst. 1987 for us. With each subsequent wild ride it gets easier to relax and enjoy the views. We have never sold even a dime's worth of stock in any downturn.

 

[samclem]

And as you know, the sensitivity at 100% is extreme - a statistical outlier can dramatically skew 100%. So much so that I wouldn't recommend anyone use it. YMMV

It seems to me that the differences between the way portfolios perform at 95% success and 100% success may provide some important hints at what might happen if the next 30 years feature performance that is different than the historical data sets. IOW, disregarding the "edge cases" in our present data requires an implicit assumption that the future will be a lot like what we've already experienced. I'm not comfortable with that assumption, but I may be in the minority. True, we've only got history to guide us, but I'm not OK placing a huge bet on a high-fidelity repeat of it, so the edge cases are important to me. Again, I understand others may differ in their assessment.

That’s right, in choosing between one of the middle allocations, one should look at annual volatility and worse case drawdown over multiple years . . . .

I evaluated different withdrawal rates for the %remaining portfolio case, and starting income and max drawdown were major criteria, but also noted the residual portfolio. I only looked at the 50/50 and 60/40 cases though. Evaluating worst case draw down took identifying the worst year and running multiple individual scenarios, so it was a lot of work using FIREcalc. You never went to 0, but worst case drawdown (inflation adjusted) usually took many many years and resulted in a much smaller portfolio and thus much smaller annual income compared to the starting income.

I agree that worst-case drawdown is an important metric. When we are riding that chart line in real life with no a priori knowledge that the line eventually turns up and everything works out, it could be scary (whether our portfolio is diving due to a rapid market slump, or we are just have a chronic failure to keep up with inflation). This is the kind of thing that could result in investors abandoning their previous plans. Do you happen to recall any differences you found in 50/50 vs 60/40 as far as max inflation-adjusted drawdown is concerned?

 

[REWahoo]

...How many people here rebalanced at the end of 2008 or during the the first quarter of 2009? ...How many here were caught like deer in the headlights and did nothing?

What about those of us who were invested primarily in balanced funds (Wellesley & Wellington) and did both?

 

[NW-Bound]

“Everyone has a plan until they get punched in the mouth” -- Mike Tyson

 

[audreyh1]

What about those of us who were invested primarily in balanced funds (Wellesley & Wellington) and did both?

Well then obviously you didn't have to worry about doing the rebalancing yourself and it didn't matter whether or not you were a deer in the headlights.

 

[Midpack]

It seems to me that the differences between the way portfolios perform at 95% success and 100% success may provide some important hints at what might happen if the next 30 years feature performance that is different than the historical data sets. IOW, disregarding the "edge cases" in our present data requires an implicit assumption that the future will be a lot like what we've already experienced. I'm not comfortable with that assumption, but I may be in the minority. True, we've only got history to guide us, but I'm not OK placing a huge bet on a high-fidelity repeat of it, so the edge cases are important to me. Again, I understand others may differ in their assessment.

I agree wholeheartedly, we’re closer to a 200% success rate (IOW we’re drawing about half the 100% spend rate).

I meant both threads only to illustrate relative cases, not to recommend withdrawal rates or success rates. Each of us make those decisions according to our needs and expectations. Again, that’s why I’ve tried to discourage focus on the exact $ numbers, they’re only good relative to each other.

 

[audreyh1]

It seems to me that the differences between the way portfolios perform at 95% success and 100% success may provide some important hints at what might happen if the next 30 years feature performance that is different than the historical data sets. IOW, disregarding the "edge cases" in our present data requires an implicit assumption that the future will be a lot like what we've already experienced. I'm not comfortable with that assumption, but I may be in the minority. True, we've only got history to guide us, but I'm not OK placing a huge bet on a high-fidelity repeat of it, so the edge cases are important to me. Again, I understand others may differ in their assessment.


I agree that worst-case drawdown is an important metric. When we are riding that chart line in real life with no a priori knowledge that the line eventually turns up and everything works out, it could be scary (whether our portfolio is diving due to a rapid market slump, or we are just have a chronic failure to keep up with inflation). This is the kind of thing that could result in investors abandoning their previous plans. Do you happen to recall any differences you found in 50/50 vs 60/40 as far as max inflation-adjusted drawdown is concerned?

I would have to go back and look. I only ran a few 60/40 scenarios. It was mostly 50/50 total stock market and 5 year treasuries versus withdrawal rate.

Max drawdown was very heavily influenced by withdrawal rate (%) as you can imagine, yet higher % withdrawal also meant higher initial income. So even if your income dropped more drastically, it might still be higher than a more modest withdrawal rate under the max drawdown worst case starting year.

I still have the overview results, but it will take a bit of work to present them again. Otherwise - I have posted the results a couple of times in this forum, but it's been a while.

 

[REWahoo]

...it didn't matter whether or not you were a dead in the headlights.

True, but I definitely was!

 

[audreyh1]

Ooops! LOL!

 

[samclem]

I still have the overview results, but it will take a bit of work to present them again. Otherwise - I have posted the results a couple of times in this forum, but it's been a while.

No, that's fine. I recall your previous posts, I'll paw around for them when I get some time. Thanks!

 

[marko]

Great thread!

I'm often out in the weeds, but in the unlikely event that I'm reading all this correctly, the main difference between a 20/80 and an 80/20 is the volatility during the course of the study and perhaps what is left as a balance at life's end.

If so, if you're not overly concerned about leaving a large estate why would you not drift away from 60/40 and more toward something with less volatility?

With that in mind, for those of us dividend investors, wouldn't a heavier bond profile (say, 40/60) allow perhaps a bit more income while ending up at the same place balance-wise?

 

[Dd852]

Yes. I think the ubiquitous "risk tolerance assessment" questionnaires found on the internet are completely useless. I have always liked Fred Schwed's discussion of the issue:

There are certain things that cannot be adequately explained to a virgin either by words or pictures. Nor can any description that I might offer here even approximate what it feels like to lose a real chunk of money that you used to own.
​

The first hit is the worst. 1987 for us. With each subsequent wild ride it gets easier to relax and enjoy the views. We have never sold even a dime's worth of stock in any downturn.

In 1987 I had no money so it was easy! I tracked everything even back then and our net worth at end year was $7,000 (first year of marriage too!). During other downturns I was employed and making good money so I happily continued to buy all the way through and am coasting on that today. (Net worth 143,000% higher !). The real question is how much of a downturn would shake my confidence now that I don’t have a paycheck coming in and never will again.... I hope I don’t get really tested.

 

[RunningBum]

Great thread!

I'm often out in the weeds, but in the unlikely event that I'm reading all this correctly, the main difference between a 20/80 and an 80/20 is the volatility during the course of the study and perhaps what is left as a balance at life's end.

If so, if you're not overly concerned about leaving a large estate why would you not drift away from 60/40 and more toward something with less volatility?

With that in mind, for those of us dividend investors, wouldn't a heavier bond profile (say, 40/60) allow perhaps a bit more income while ending up at the same place balance-wise?

As you say, 80/20 probably leaves you with more money at the end. If you don't care about your ending balance, doesn't that mean you likely can spend a little more, either for unexpected expenses, or to splurge?

 

[big-papa]

One of the reasons I like PMT based withdrawal methods instead. For a given number of years you want to plan for, you can set the terminal value of the portfolio to be exactly $0 or any other number you want it to be. One trades off the possibility of running out of money prematurely for a more volatile income stream, something that I find more palatable for my situation, especially once SS starts.

One of the things you find when looking at non-SWR-ish types of withdrawal methods is that many of the rules of thumb and even terminology for withdrawals were made assuming SWR. And those might not apply to a variable withdrawal method, or at least not in the same way. For example, you no longer talk about probability of running out of money for an 80/20 vs 60/40 portfolio, but rather what the volatility of withdrawals might look like. And the details of that depend on the actual variable withdrawal method chosen.

 

[marko]

As you say, 80/20 probably leaves you with more money at the end. If you don't care about your ending balance, doesn't that mean you likely can spend a little more, either for unexpected expenses, or to splurge?

I suppose. But I'd guess a 40/60 might also get you there to the same place via slightly larger bond payments.

Academic for me as I'm 60/40 and will likely stay there for the foreseeable but assuming I'm following all of this correctly, an interesting discussion.

 

[HI Bill]

I believe that the success rates shown include both historic bond and stock yield averages. However, IMHO, it appears that we'll never see really high bond yields again in our lifetime, given the Fed's manipulation of the $/interest rates, and the need for cheap $ to allow us to continue deficit spending. My theory is that the model over-accounts for the safety given by bonds, since the returns will never return to what they were. Just my theory.... Vanguard's 'balanced asset mix' chart shows bond returns from 1926-2012.