Trinity Study update

[DblDoc]

Updated tables with data through 2009 and interesting discussion on "adaptive" withdrawal rates for when times are good/bad.

Portfolio Success Rates: Where to Draw the Line

Discussion started over at bogleheads here: Bogleheads :: View topic - Trinity Study Authors update their results

DD

 

[W2R]

Thank you, DblDoc! Interesting update that deserves some careful reading, so I have no comments yet. Definitely worth reading IMO, and food for thought.

 

[W2R]

Thank you, DblDoc! Interesting update that deserves some careful reading, so I have no comments yet. Definitely worth reading IMO, and food for thought.

I love this quote from the article:

In a sustained bear market such as 2008 to early 2009, clients preserve the opportunity for portfolio recovery by reducing withdrawal amounts to no more than the dividend and interest income from the portfolio, thus avoiding the liquidation of shares or bonds at low values. Planning budgets in retirement and allocation to liquid assets should include the possibility of such market conditions.

That is how I plan to approach another crash, if we should have one.

 

[chinaco]

A Comment from the OP at BH. His last statement about 75% chance of success at 7%.

I haven't had a chance to read the whole article yet, but this line from their executive summary is a bit surprising:

"We conclude that if 75 percent success is where to draw the line on portfolio success rates, a client can plan to withdraw a fixed amount of 7 percent of the initial value of portfolios composed of at least 50 percent large-company common stocks."

Well, the actual result is not suprising, but it is surprising that they would emphasize it. Taking a 25% chance for failure does sound a bit risky for my tastes.

The key words "a bit risky for my tastes"

The way I understood the conclusion was that the author's suggest that 75% might be a lower boundary for what "might be" reasonable for people willing to take "a little more risk", but also want some realistic chance of achieving success. Not that he is recommending it.

IMO - One of the authors' of this study [in interviews] stated that the study (the original one) is not to be used as a rule, but a measure of risk for portfolio success with certain WR% that one might use to plan or understand a potential outcome. If one has an overly optimistic plan, that means that they are more likely to have to make course corrections along the way. [For example] if one chooses to take a fixed 7% (no cola) [and that translated to 75% chance of success] they have a 25% chance that they may need to course correct if they expereince the "bad scenario". They have a 75% chance of experiencing the "good scenario" and taking the 7% nominal for 30 years. But even 75% chance of success could turn out to be a 100% failure because the past may not exactly looks like the future.


It is not an immutable law of financial physics! But is is still very valuable to me. I get it.

Warning Label: Read and understand the instructions and use with caution!

 

[donheff]

Reassuring for us ER types using low SWRs. Those median ending portfolio amounts promise the possibility of first class plane seats in our mid 70s when our joints are aching.

 

[Midpack]

OK, 75% success at 7% WR got my attention.

But in reading I see that's for someone with a 20 year horizon. Nice read, but nothing really new.

I'm still comfy at 60/40 AA and a 2.5% WR. Odds are we'll be able to increase spending in retirement, and if not we'll still be fine. We'd rather that than to start out overly optimistic and have to substantially decrease spending later of course.

 

[donheff]

OK, 75% success at 7% WR got my attention.

But in reading I see that's for someone with a 20 year horizon. Nice read, but nothing really new.

I'm still comfy at 60/40 AA and a 2.5% WR. Odds are we'll be able to increase spending in retirement, and if not we'll still be fine. We'd rather that than to start out overly optimistic and have to substantially decrease spending later of course.

Take another look. 7% with a 50/50 portfolio has an 85% success rate for 30 years. But if you want inflation adjustments you have to go lower.

 

[Midpack]

Take another look. 7% with a 50/50 portfolio has an 85% success rate for 30 years. But if you want inflation adjustments you have to go lower.

I can't imagine anyone would actually use Table I, without inflation adjustment, as a basis for a 20-30 or more year plan - unless they're planning to live on cat food at the end of life. So I went immediately to Table II, inflation adjusted. The authors seem to acknowledge same in their conclusions where they don't refer to the data in Table I at all...

If a client expected a 20-year payout period after retirement and was willing to accept a 75 percent success rate, he or she could select a 7 percent or even an 8 percent annualized withdrawal rate for fixed withdrawals from a portfolio of at least 50 percent stocks.

The sample data suggest that clients who plan to make annual inflation adjustments to withdrawals should also plan lower initial withdrawal rates in the 4 percent to 5 percent range, again, from portfolios of 50 percent or more large-company common stocks, in order to accommodate future increases in withdrawals.

 

[nun]

In a sustained bear market such as 2008 to early 2009, clients preserve the opportunity for portfolio recovery by reducing withdrawal amounts to no more than the dividend and interest income from the portfolio, thus avoiding the liquidation of shares or bonds at low values. Planning budgets in retirement and allocation to liquid assets should include the possibility of such market conditions.

Isn't this as obvious as "the nose on your face". Wow, don't eat into capital. A bear market is the reason to have a few years set aside in cash or a CD ladder and to have income that is not dependent on the stock market.

 

[chinaco]

Most people's situations are not as simple as pile of money in a portfolio. In the US, most have some sort of other resources (pension, SS, or both) and hopefully some sort of savings that are invested.

I believe those studies just offer insight... and a rough planning assumption that will need to be (or want to be) adjusted (up or down).

The high nominal WR% (or some variation) might be appropriate for some people. Whether or not it goes on for 30 years or not is the question.

If one hopes to take a pile of money in a portfolio and create a cola or non-cola pension like income stream for (pick your years)... then this helps one to understand it a little better.


My income streams look roughly like an isosceles triangle with the ends clipped. Those income sources are staggered as they come on line (Pension + SPIA ladder + SS1 + SS2) in a stepped fashion. At the peak they might come close to covering most of our planned income but inflation and eventually the death of one of us will reduce it to the level that the portfolio is needed to fund the gap. We will spend more heavily on the front side of the triangle (ER years) and back side (late years).

 

[youbet]

Reassuring for us ER types using low SWRs. Those median ending portfolio amounts promise the possibility of first class plane seats in our mid 70s when our joints are aching.

Don..... Please keep in mind that those landing near the "median" will be rare. Most will wind up with ending portfolios significantly higher or lower than "median."

That'a a concept I find many folks struggle with. Most I chat with about it seem to assume they'll wind up near the median. Actually, most will wind up far removed from the median.

I think this is caused by the common use of the term "average" as meaning "typical." In our use here, hitting near median is not typical but rather rare in occurance and simply a measure of central tendancy of data which, for most asset allocations and historical time periods, is widely dispursed.

So, yes, if you wind up with "median" or better, those first class plane seats will be at your disposal. But it will be common to not hit median. And, of course, being below "median" doesn't necessarily mean you run out of money. It may simply mean you have a smaller residual at the end of the withdrawal period.

 

[FIRE'd@51]

Actually, most will wind up far removed from the median.

How do we know this? As far as I can tell, the authors didn't present a standard deviation of the ending values, which would have been helpful (you would think an academic study would have included this ). Then we could estimate the probability of winding up a specified amount from the median.

 

[Spanky]

Fortunately, even if the future returns on the S&P500 as a whole are fairly tame, it should be possible to do substantially better than a 4% draw by using strategic asset allocation and by finding the right risk/return balance. I have shown examples here of a portfolio which supports a 6.15% draw.

Hurray! I should stop my plan for 2% SWR - not.

 

[haha]

[For example] if one chooses to take a fixed 7% (no cola) [and that translated to 75% chance of success] they have a 25% chance that they may need to course correct if they expereince the "bad scenario".

This kind of thing is pure trash, resume padding by someone with no experience of what really gets retirees into trouble. It is absolutely irresponsible and meaningless to talk about nominal withdrawals in a real world.

True that some multimillionaires on this board have way more money than they need to fund an expected simple life at home, so a nominal draw might seem reasonable, especially if they could really stand life at home with only the missus for distraction, or even more likely if she could stand it. But these people are not the ones who will go for 7%. 7% will appeal to people who detest work or who are getting itchy to travel NOW! and who may not have a whole lot of excess.

Say realistically a couple could squeeze 20% out of the budget (and unless one is a government worker, how could he know what their insurance and medical care might cost, or what might befall one or the other?) Or what congress might do to SS and Medicare?

Then go back and look at what would have happened to that 20% surplus in the 1970s, with any allocation that could have been guessed in advance.

They would be very severely stressed just to pay property tax and utilities and buy food and gasoline, etc. etc. And if P.V. had not shown up, these stressed retirees might have been on breadlines.

Ha

 

[Danmar]

Isn't this as obvious as "the nose on your face". Wow, don't eat into capital. A bear market is the reason to have a few years set aside in cash or a CD ladder and to have income that is not dependent on the stock market.

I guess I'm going to have to spend more than simply dividends? My expected estate value after 30 years would be astronomical if we only spend dividends.

 

[nun]

I know that many will be relying on CGs as well as dividends in retirement which is why it's important to have at least a couple of years cash on hand so you don't need to sell low to cover expenses.

My ER income stream looks like this: rental income, this is already in place and will cover half my ER expenses the rest will come from either a small 72t or spending down after tax savings equivalent to a 1% annual withdrawal from my portfolio. I'm leaning towards the 72t to give me extra liquidity and reduce RMDs. I'll also do annual IRA to ROTH rollovers. At 59.5 I get a small COLA pension, and at 66 I'll get SS and UK retirement pensions that will total $28k in today's dollars. So I won't actually need to make any withdrawals from my IRAs etc after 66. I go into another accumulation phase.

 

[rescueme]

This kind of thing is pure trash, resume padding by someone with no experience of what really gets retirees into trouble. It is absolutely irresponsible and meaningless to talk about nominal withdrawals in a real world.

True that some multimillionaires on this board have way more money than they need to fund an expected simple life at home, so a nominal draw might seem reasonable, especially if they could really stand life at home with only the missus for distraction, or even more likely if she could stand it. But these people are not the ones who will go for 7%. 7% will appeal to people who detest work or who are getting itchy to travel NOW! and who may not have a whole lot of excess.

I agree with your statement/analysis.

While we're currently in excess of that "magic 4%", we also don't have all our retirement sources "on-line" at this time.

We're still waiting on two small defined benefits (e.g. pensions) for DW, along with three sources of SS (DW at FRA age of 66, my claim of 50% against her at that time, and finally my SS at age 70).

While we certainly will exceed that 4% currently, and for the next 6+ years (till I turn 70, and claim SS), at that time our combined WD rate will be well under 4% (actually, currently computed at just under 2.5%).

Do we really want to impact our current standard of living just to adhere to a "suggested number"? I think not.

While such "guidance" may be of value to those that never considered the subject, for those that have really computed the possibilities, the media has not given coverage for the many variances of retirement, but continue to pander to the "great unknown" (or clueless)....

 

[FIRE'd@51]

We're still waiting on two small defined benefits (e.g. pensions) for DW, along with three sources of SS (DW at FRA age of 66, my claim of 50% against her at that time, and finally my SS at age 70).

As I understand it, you can do either of these, but not both.

 

[youbet]

How do we know this? As far as I can tell, the authors didn't present a standard deviation of the ending values, which would have been helpful (you would think an academic study would have included this ). Then we could estimate the probability of winding up a specified amount from the median.

Oh. Hmmmm..... You're right. My comments were based on FireCalc output graphs, not on anything this author presented. So if future outcomes are different than the historical-data driven outcomes from FireCalc, perhaps the dispersion will be small and most people will be right at the median.

That doesn't pass my common sense test. But, who knows, I suppose it's possible that folks retiring in any year of the next several decades will all have similar outcomes unlike folks who retired in the past and typically had wide ranging outcomes depending on the year they retired.

Do you agree that at least with FireCalc output data, few outcomes are at or very near the median while many outcomes are widely dispursed from the median? That's the way it looks to me. So to assume (which many do) that your outcome is likely to be near the median is not justified.

 

[rescueme]

As I understand it, you can do either of these, but not both.

Sure I can. I turn 66 (my FRA) in January, she in May of the same year and will apply for SS at that time. Once that happens, I can claim against her till I turn age 70.

Read the following example (under the heading "Team Play"):

Two Ways To Optimize Social Security Benefits

Our situation is a bit different since she is not going to take SS till her FRA, but it dosen't change the manner in which we can apply for benefits. However, it will give me a bit over 3.5 years of 50% of her benefit till I turn 70, and at that time (assuming I die first) will give her a much greater benefit beyond hers, after my passing.

 

[FIRE'd@51]

Oh. Hmmmm..... You're right. My comments were based on FireCalc output graphs, not on anything this author presented. So if future outcomes are different than the historical-data driven outcomes from FireCalc, perhaps the dispersion will be small and most people will be right at the median.

That doesn't pass my common sense test. But, who knows, I suppose it's possible that folks retiring in any year of the next several decades will all have similar outcomes unlike folks who retired in the past and typically had wide ranging outcomes depending on the year they retired.

Do you agree that at least with FireCalc output data, few outcomes are at or very near the median while many outcomes are widely dispursed from the median? That's the way it looks to me. So to assume (which many do) that your outcome is likely to be near the median is not justified.

I was only questioning your statement that most (as opposed to many) ending values were far from the median. To know that we would need to know the actual distribution of ending values (and its moments). This is true for Firecalc as well.

 

[youbet]

I was only questioning your statement that most (as opposed to many) ending values were far from the median. To know that we would need to know the actual distribution of ending values (and its moments). This is true for Firecalc as well.

The fact that end outcomes are widely dispursed around the median is readily apparent with FireCalc output graphs with just a quick glance. Sometimes when just playing around with the numbers, I'll change AA values just to watch the distribution widen or tighten up.....

But, if you need to have descriptive statistics of the FireCalc ending values, they're readily available from the downloadable spreadsheets. I went through that exercise a number of times before coming to the conclusion that, since eyeballing the graph pretty much tells the story and the distribution is skewed anyway, I was measuring with a micrometer and cutting with an axe.

Did you see anything different when you applied the usual statistical formulas/calculations to the data in the spreadsheets?

My suggestion to Don was simply that his observation about the median value was correct but that any assumption that the median value is "typical," "common" or "probable" is likely a stretch.

 

[chinaco]

This kind of thing is pure trash, resume padding by someone with no experience of what really gets retirees into trouble. It is absolutely irresponsible and meaningless to talk about nominal withdrawals in a real world.
...

Ha

Have you read the report?

Do you think my interpretation of the authors conclusion (or data) is wrong (as stated in the example) or are you saying you think taking a nominal withdrawal is a bad idea or increasing the risk of failure (i.e., must be 100% success)?

It kinda looks like high nominal WD caused a reaction.

There are scenarios and circumstances where nominal can make sense just as there are scenarios and circumstances where it would be a bad idea.

My main point was... it is not a rule, rather insight that can be applied to planning.

Personally, I am fairly conservative.

 

[donheff]

What is with the big stats argument? I started it by fantasizing how I would spend my median ending portfolio. For my fantasy it doesn't much mater if there is a fat tail. I have a better than even chance of flying first class. If I am on the bad end I will stick with coach.

 

[chinaco]

I assume he thought I was making a general recommendation for a high nominal WR%.


To use the Otar vernacular... Red and gray zoners would be taking on the risk of a bad outcome... with the possibility of having to cut their base lifestyle.


Green zoners might also if they pushed it too far.


Still the key part of managing any portfolio for income is: to begin with a realistic plan and make adjustments based based on reality as the future unfolds... don't plan on the market saving you!