Juggling Trinity Study Data AOF Style

[Dr. Mom]

In case you missed it, the link below is AOF messing around actuary style with the Trinity Study data.  Super interesting to me.  Reinforces sequence of returns issues as well.  Happy reading...

https://rockstarfinance.com/trinity-study-another-dimension/

[VagabondMD]

I read it (and the companion article) already this morning. Great stuff!

It validates, to some degree, my concern of being overly complacent about the “4% Rule” and its variants. On the other hand, in most possible scenarios and regimes, the 4% Rule is too conservative. The future is unknowable, but will likely not be like the past which makes basing future decisions on the past somewhat risky.

I also like that he explodes the myth that stock returns are virtually guaranteed if you hold on long enough. There is certainly no absolute rule of law or of the universe that this is the case though we often assume that it is.

[Dont_know_mind]

I think a retirement strategy should be robust to mean reversion in valuation at a minimum.

My strategy is to assume historical returns in setting my asset allocation, but to assume mean reversion in stock valuation and mean reversion in bond yields over the next 10 years for return estimates.

Assuming mean reversion in both these, I don't expect a greater than 2% nominal return over the next 10 years in a 70/30 portfoilio. This maybe excessively pessimistic, but assuming mean reversion is not expecting a Great Depression or anything, merely a bear market at some stage in the next 10 years. They have been wrong for a while, but I suspect the Oaktree/GMO 10 year forward return estimates to pan out in the end. Perhaps they won't and that would be a bonus.

[Molar Mechanic]

It's a neat trick, but doesn't really change anything.  The Trinity study, with arguably more real world results came up with 4% as safe 95% of the time.  He used a set of real numbers but in an order that had never occured in real life, thereby creating more outliers, and came up with 3.8% as 95% safe.  That doesn't seem to be much of a real world game changer to me.  So now we save to 26.3x expenses instead of 25x?  I hope there aren't that many folks in our industry that are so miserable that they are just waiting for a number in the bank statement to walk away.  If you make enough to save that aggressively, shouldn't you be able to change your lifestyle to balance things first?

I think most can agree that some amount of flexible drawdown strategy is appropriate for most, especially those who aren't going for bare bones, subsistence living "Lean FIRE" type retirement.  In the down years, a few more burgers and a few less steaks, a little more National Parks and a little less Europe, etc in the lean years to make the numbers work.  Most of us who retire well will want to increase spending as savings allow.

I want to see the study done that can create an actionable drawdown strategy.  Maybe it's a plan of minimum safe spending (Say 3%) plus a percentage of gains greater than "X".  Perhaps it just works out to be a periodic resetting the number to take 4% of, or perhaps is decade of life minus 2% (forties you take 2%, sixties you take 4%, eighties you take 6%.  I don't have a clue what the "right" formula will turn out to be, but I'm pretty sure there is one out there that will show safety greater than 95% with average spending greater than the 4% number.

[The White Coat Investor]

I think the big mistake in his methods is assuming the returns this year have nothing to do with the returns from last year. I don't think that's true. When bond prices go down, the yields go up. When stock prices go up (assuming the fortunes of the company haven't changed), the dividend yield goes down.

In short, I think it's a worthless exercise that doesn't tell you anything of value.

The conclusion, that you should be careful especially if you run into the buzzsaw of poor returns and high inflation early in retirement, is valid, but the study doesn't show that.

[Complete_newbie]

The White Coat Investor wrote:

I think the big mistake in his methods is assuming the returns this year have nothing to do with the returns from last year. I don’t think that’s true. When bond prices go down, the yields go up. When stock prices go up (assuming the fortunes of the company haven’t changed), the dividend yield goes down.

In short, I think it’s a worthless exercise that doesn’t tell you anything of value.

The conclusion, that you should be careful especially if you run into the buzzsaw of poor returns and high inflation early in retirement, is valid, but the study doesn’t show that.

Click to expand...

So you have the right to be definitive (calling it worthless...strong words) and he doesn't?

Here is my conclusion: It doesn't make it worthless. Also if you are so sure about correlation between year to year stock returns, then prove it?

I would say that yea there may be correlation between stock returns YoY (bogleheads are momentum investors with long term horizon) and if so would like to see it as a study. Furthermore its just an exercise in resampling the data with replacement written fancily for your average reader.

[Dr.V. Investor]

This was an interesting article yesterday. I really am not a believer of a glide path, it is interesting how many people dissect these things.  The more savings, the safer, hoping to leave some for children.

Since this requires subscription, I am adding some of the text:  With both stocks and bonds expensive by historic measures, and people having longer retirements, researchers are rethinking these rules to better manage the risk of a market decline.  The expectation for future returns are lower.

Consider a dynamic approach, Mr. Blanchett says. Many dynamic strategies set their initial withdrawal rate for a 30-year retirement at about 5%. But they require users to cut spending in a year in which their portfolio loses value.

Inflation-adjusted income is likely to decline over time.A second method, dubbed “the guardrail approach,” provides more latitude to raise spending.

Say you retire with $1 million in a portfolio with 60% in U.S. and foreign stocks and 40% in bonds and withdraw 5%, or $50,000, in year one. At year-end, you must recalculate your withdrawal amount as a percentage of your new balance. Assuming your portfolio declines 20% to $800,000, your $50,000 withdrawal—plus an annual adjustment for inflation—now represents more than 6% of your new $800,000 balance.

Any time your withdrawal rate rises above 6%, the rule imposes a 10% pay cut for the next year, says Jonathan Guyton, a financial adviser and co-creator of this strategy. As a result, after adjusting the $50,000 initial withdrawal—to $51,000, assuming 2% inflation—the method imposes a 10% pay cut, of $5,100, to produce a $45,900 withdrawal in year two.

The good news: You can take a 10% raise following years in which your withdrawal rate falls below 4%.

In years in which your withdrawal rate is between 4% and 6%, simply adjust your most recent withdrawal—$50,000 in this example—to keep up with inflation. (But don’t take the inflation adjustment following any year in which your investments.

During a bull market, your balance may rise enough to tempt you to save less or retire early—setting you up for trouble if the markets reverse course. They recommended focusing on saving a fixed amount rather than focusing on reaching a number.

But a recent study by Mr. Pfau and Michael Kitces, director of wealth management at Pinnacle Advisory Group Inc. in Columbia, Md., finds that those who take the opposite approach—by reducing stock exposure in the initial years of retirement and then gradually raising it over time—are likely to make their money last longer.

According to the research, those who start retirement by reducing their stock holdings to 20% to 30% of their portfolio and end up with 50% to 70% in stocks can withdraw 4% of their balance per year and give themselves annual raises to compensate for inflation over 30 years, even in the worst market scenarios. (The authors based these simulations on historical market returns rather than accounting for the lower interest rates facing today’s retirees.)

In contrast, those who keep 60% in stocks throughout retirement or who taper to a 30% stock allocation from 60% are likely to run out of money after 28 years in the 5% of worst-case scenarios, says Mr. Pfau.

Of course, if stocks fare well in the early part of retirement, those who use the conventional approach will come out ahead. But the new approach provides better downside protection in the years right after retirement, when retirees are most vulnerable to financial losses. If a bear market occurs then, a portfolio can quickly be depleted by market losses and withdrawals.

 

https://www.wsj.com/articles/forget-the-4-rule-rethinking-common-retirement-beliefs-1518172201

[Zaphod]

Totally agree with MM and WCI, those are big flaws. Yes the minimum was quite lower, but the results were basically the same, which is it really that surprising? Im no math wiz but if you're simply taking a similar data set and reusing the same numbers but in different combinations...you're going to end up at a similar average. The more you do it the more you just come to the true average. It would be different if the ranges in the expanded set were crazy different, but since thats unlikely and they were rolling periods its just more similar data.

[ENT Doc]

If you're betting on a 4% rule to enable success I'll just say May the 4th be with you.

[Complete_newbie]

Zaphod wrote:

Totally agree with MM and WCI, those are big flaws. Yes the minimum was quite lower, but the results were basically the same, which is it really that surprising? Im no math wiz but if you’re simply taking a similar data set and reusing the same numbers but in different combinations…you’re going to end up at a similar average right?

Click to expand...

Kinda true. You wont end up with SAME results, but what this dude is doing is essentially creating confidence intervals (he didn't mention it explicitly but you could see it easily) something like 3.8%-4.2% withdrawal has 95% certainty of lasting etc. You can get higher significance for 99% of the time based on your simluations.

Anyways, been thinking really hard about an analytical blog and am close to starting it (mostly to call BS on alot of numbers thrown without context). Stay tuned.

[Zaphod]

Complete_newbie wrote:

Zaphod wrote:

Totally agree with MM and WCI, those are big flaws. Yes the minimum was quite lower, but the results were basically the same, which is it really that surprising? Im no math wiz but if you’re simply taking a similar data set and reusing the same numbers but in different combinations…you’re going to end up at a similar average right?

Click to expand…

Kinda true. You wont end up with SAME results, but what this dude is doing is essentially creating confidence intervals (he didn’t mention it explicitly but you could see it easily) something like 3.8%-4.2% withdrawal has 95% certainty of lasting etc. You can get higher significance for 99% of the time based on your simluations.

Anyways, been thinking really hard about an analytical blog and am close to starting it (mostly to call BS on alot of numbers thrown without context). Stay tuned.

Click to expand...

Right. I expanded a tiny bit. Otherwise I think MM discussed it well.

Making the study more math appropriate and robust does not mean it is automatically more robust in real life. Trinity study also supposes you spend more every single year of your life, which we know isnt true either.

[Zaphod]

Dr.V. Investor wrote:

 

In contrast, those who keep 60% in stocks throughout retirement or who taper to a 30% stock allocation from 60% are likely to run out of money after 28 years in the 5% of worst-case scenarios, says Mr. Pfau.

Of course, if stocks fare well in the early part of retirement, those who use the conventional approach will come out ahead. But the new approach provides better downside protection in the years right after retirement, when retirees are most vulnerable to financial losses. If a bear market occurs then, a portfolio can quickly be depleted by market losses and withdrawals.

Click to expand...

That is simply another way of stating that SORR is a real issue that needs to be respected. Lots of ways to do it, x number of years, gliding up equities, etc...Bonds decrease volatility, thereby making withdrawal rate more steady, but at the cost of increasing longevity risk aka you run out of money. So basically you simply need to make it out of that SORR time frame.

For most here your balance is going to be large enough it wont matter at all and we'll likely have some fixed portion of the portfolio to address early years. For others, a much larger concern if they have less and are not able to essentially choose their retirement date.

[Complete_newbie]

Zaphod wrote:

Complete_newbie wrote:

Zaphod wrote:

Totally agree with MM and WCI, those are big flaws. Yes the minimum was quite lower, but the results were basically the same, which is it really that surprising? Im no math wiz but if you’re simply taking a similar data set and reusing the same numbers but in different combinations…you’re going to end up at a similar average right?

Click to expand…

Kinda true. You wont end up with SAME results, but what this dude is doing is essentially creating confidence intervals (he didn’t mention it explicitly but you could see it easily) something like 3.8%-4.2% withdrawal has 95% certainty of lasting etc. You can get higher significance for 99% of the time based on your simluations.

Anyways, been thinking really hard about an analytical blog and am close to starting it (mostly to call BS on alot of numbers thrown without context). Stay tuned.

Click to expand…

Right. I expanded a tiny bit. Otherwise I think MM discussed it well.

Making the study more math appropriate and robust does not mean it is automatically more robust in real life. Trinity study also supposes you spend more every single year of your life, which we know isnt true either.

Click to expand...

Well yea. Can't account for individual needs. Its not assuming much though. And your statement is general. Like saying this drug shows x% improvement doesn't mean its gonna work on you. Well, true but then you're not going to draw conclusions from it? Then why are we doctors and doing research?

Its not math appropriate, its just simulation. To be honest I have never read trinity study (may be I should) but what this guy is doing is not "worthless" and that is my point. It is actually well done and tbh, can you prove to me stock returns YoY are correlated with prior years? if not then this method is valid to calculate different scenarios.

The biggest take away from SOR etc I take is: well if everyone was SOOO comfortable with stock market and index investing then why the heck are we discussing it 20349830482409 of times? Set it and forget it. But that is not so. What are you going to do about it?

[AlexxT]

All of these studies are interesting, but the solution to all of the problems they raise is the same.

When the market is down, spend less.

Bonds are yielding less, and stocks will probably yield less in the future, so expect to withdraw less than 4%.

If this study scares you, then plan on withdrawing less than 4%.

3% is safer than 4%.

2% is safer than 3%.

If you can't make the numbers work on 4%, or 3%, then work for another couple of years and run the numbers again.

If you're retired and the market tanks and you don't have enough money, go back to work.

You're welcome.

 

[Complete_newbie]

Molar Mechanic wrote:

It’s a neat trick, but doesn’t really change anything.  The Trinity study, with arguably more real world results came up with 4% as safe 95% of the time.  He used a set of real numbers but in an order that had never occured in real life, thereby creating more outliers, and came up with 3.8% as 95% safe.  That doesn’t seem to be much of a real world game changer to me.  So now we save to 26.3x expenses instead of 25x?  I hope there aren’t that many folks in our industry that are so miserable that they are just waiting for a number in the bank statement to walk away.  If you make enough to save that aggressively, shouldn’t you be able to change your lifestyle to balance things first?

I think most can agree that some amount of flexible drawdown strategy is appropriate for most, especially those who aren’t going for bare bones, subsistence living “Lean FIRE” type retirement.  In the down years, a few more burgers and a few less steaks, a little more National Parks and a little less Europe, etc in the lean years to make the numbers work.  Most of us who retire well will want to increase spending as savings allow.

I want to see the study done that can create an actionable drawdown strategy.  Maybe it’s a plan of minimum safe spending (Say 3%) plus a percentage of gains greater than “X”.  Perhaps it just works out to be a periodic resetting the number to take 4% of, or perhaps is decade of life minus 2% (forties you take 2%, sixties you take 4%, eighties you take 6%.  I don’t have a clue what the “right” formula will turn out to be, but I’m pretty sure there is one out there that will show safety greater than 95% with average spending greater than the 4% number.

Click to expand...

Once I am not working full time I can do that (by then it'd already be done lol). I have it in my head but the programming required is a bit involved. Sigh. Time, the most precious entity!