Putnam: 10% Is Optimal Retirement Equity Allocation

[GaryInCO]

Quote:

Originally Posted by MasterBlaster You need to buy at the bottom of the cycle and sell at the peak. That's the trick - Buy low sell high ! Whatever you do don't reverse this tried and true method of success.

Sound advice. I can state from personal experience that the "buy high sell low" approach doesn't work nearly as well.

[ERD50]

Quote:

Originally Posted by chinaco ERD - you should to reread that research paper. His entire point is that tools like FIRECALC do not analyze certain aspects of the problem. Here are a few quotes... just to make that point. Hopefully no one is putting too much faith in any calculators to divine their financial future in the security markets. They merely provide some insight to formulate an approach and plan. They do not provide your specific answer! IMO - His paper points out some often overlooked issues and a different form of analysis that includes some of those issues. He is making a similar point as that of Zvi Bodie and some others. Take no more risk than is necessary with assets that will be needed to fund one's living expenses in retirement. He pegs the (low risk portfolio) equity level to be fairly low (5% to 25%). Now, he also suggests that those who have a bequest motive (and excess assets for a residual estate) may choose to invest those excess assets (and take some risk with the estate) with the hope of acquiring more for the estate... he suggests a 35% to 45% equity allocation might be appropriate. I thought the paper was informative. Too bad it did not show illustrations for 55 yo early retirees.

I have not read every word of the paper, but I did read the excerpts you quoted. It still leaves me wondering.

FIRECALC shows the higher EQ AA to provide a higher success %. Going down below 35% equities drops the success % for the scenarios I've run (including the default one). So why does this paper define 5~25% equities as a higher success rate?

Forget about how much you leave on the table ('positive RPV') - I'm talking about the risk of running out of money ('negative RPV)'. It seems to me that FC is all about negative RPV (failures). And FC says that something much higher than 25% equities is best (avoiding failure) for a pretty wide range of scenarios. I just don't get why this report differs in AA so greatly from what FC reports.

-ERD50

[chinaco]

Quote:

Originally Posted by ERD50 I have not read every word of the paper, but I did read the excerpts you quoted. It still leaves me wondering. ... I just don't get why this report differs in AA so greatly from what FC reports. -ERD50

And you can guarantee you will not "get it" if you do not read it. Although, You probably won't fully understand (the issues you are concerned about) even if you do read it.

His model seems to be completely different and he is using different data (some variables appear to be derived then applied as base assumptions).

His portfolio is "specially optimized".... on and on.

Quote:

One final methodological issue needs to be discussed. The asset allocations used throughout this report are optimized so as to minimize retirement downside (depletion) risk for any given scenario. Given the complex nature of the problem we are examining, we are forced to use a stochastic optimization process to seek out the best asset allocation mix for any set of assumptions. This approach is different than that used for conventional optimization in that thousands of simulations are made with each step of the algorithm in its search for the best solution.

Your observation is probably related to the fact that each of those tools and research papers look at the issues differently (trying to uncover or shed light on specific issues), include different data, and include certain assumptions.

His goal was to make a point.... He obviously was not randomly running simulations and accidentally happened across an "ah ha" type discovery.

The market behavior of the last decade has put some popular notions under more scrutiny.

Quote:

The study suggests, in short, that the higher equity allocations used in many popular retirement investment products today significantly underestimate the risks that these higher-volatility portfolios pose to the sustainability of retirees’ savings and to the incomes they depend on.

I do not take any of those reports or any tool to be anything other than something that provides some insight about the problem... perhaps they suggest an outcome that could be more likely (assuming the model is followed and the past looks like the future).... and assuming one does not scr3w up somehow.

[Ken11]

I only scanned the paper- but this caught my eye:

"Of course, the duration of any specific plan or portfolio
will vary because of the uncertainty of how long one will
live. But RPV analysis captures and integrates all of these
dynamic components — flows, returns, and longevity —
and then discounts them into a positive or negative value

expressed in today’s dollars"
-----------------

THis means the analysis involves some averages.
Thats why the 7% SWR is mentioned.
It seems that, for any individual, longevity risk remains uncovered.
The analysis seems to isolate "investment risk"
I dont know any smart person that recomments a 7% initial withdrawal rate.

Even just isolating the investment risk I'm still left wondering:
Was mindless annual rebalancing assumed?
How were the correlations that exist between the various returns/factors modelled?
Mote carlo suggests these correlations were not considered
(Note that Firecalc inherently does get this right.
The natural correlations are built into the Firecalc data.)
What maturity were the bonds assumed to have?

[MasterBlaster]

As I understand it, they do a multi-dimensional risk variance normalization where they normalize for the distribution of portfolio return risk, longevity risk, and sequence of withdrawal risk.

It's a model. FIRECALC is also a model. Maybe the model(s) is/are spot on. maybe it's not too good at predicting what to do.

What I can tell you though is that except for those retiring right before a bear market or before massive inflation hits, that the ubiquitous 4% SWR is way too conservative and severely limits your real-lifestyle in retirement.

The discussed model makes an attempt to risk-normalize that.

[ERD50]

Quote:

Originally Posted by Ken11 [ THis means the analysis involves some averages. Thats why the 7% SWR is mentioned. It seems that, for any individual, longevity risk remains uncovered. The analysis seems to isolate "investment risk" I dont know any smart person that recomments a 7% initial withdrawal rate. Even just isolating the investment risk I'm still left wondering: Was mindless annual rebalancing assumed?

Those were the same Q's that I wondered about - I will try to find time later to read it all, but from my scanning, I'm not expecting to find the answers.

Quote:

Originally Posted by MasterBlaster What I can tell you though is that except for those retiring right before a bear market or before massive inflation hits, that the ubiquitous 4% SWR is way too conservative and severely limits your real-lifestyle in retirement.

And unless one has a crystal ball to know if they are retiring before one of those events or not (or some unforeseen event), 4% doesn't seem so conservative.

It is true that most of the historical runs turn out fine @ 4%, or even higher. But I'm not so concerned about the 'most', I'm concerned about the 'what ifs'. Most of the time, I don't get in an accident when I drive my car. I don't think that leads to the decision that we shouldn't bother with seat belts.

-ERD50

[MasterBlaster]

Quote:

Originally Posted by ERD50 And unless one has a crystal ball to know if they are retiring before one of those events or not (or some unforeseen event), 4% doesn't seem so conservative. It is true that most of the historical runs turn out fine @ 4%, or even higher. But I'm not so concerned about the 'most', I'm concerned about the 'what ifs'. Most of the time, I don't get in an accident when I drive my car. I don't think that leads to the decision that we shouldn't bother with seat belts. -ERD50

In my opinion you are missing the point of the linked article.

We can all sit in our bunker saving our nuts for the FIREacolypse. Or we can do some sort of variable scheme that just may allow us more income during the period that matters. If success is defined as never going broke during retirement and always having a baseline standard of living, then either scheme will work but the former (fixed SWR) scheme leaves much to be desired in terms of lifestyle.

I believe that what the linked study attempts to do is optimize withdrawal. You may live to be 100 but it isn't probable. There may be another depression but it isn't very likely. and so on.

The linked article then uses a risk normalization scheme based on historical data to suggest what you can withdraw given a current nestegg and risk and reward profiles.

[ERD50]

Quote:

Originally Posted by MasterBlaster In my opinion you are missing the point of the linked article. We can all sit in our bunker saving our nuts for the FIREacolypse. Or we can do some sort of variable scheme that just may allow us more income during the period that matters. If success is defined as never going broke during retirement and always having a baseline standard of living, then either scheme will work but the former (fixed SWR) scheme leaves much to be desired in terms of lifestyle. I believe that what the linked study attempts to do is optimize withdrawal. You may live to be 100 but it isn't probable. There may be another depression but it isn't very likely. and so on. The linked article then uses a risk normalization scheme based on historical data to suggest what you can withdraw given a current nestegg and risk and reward profiles.

Well, I read it (most of it, still skimmed a bit from p 10-14). I guess I 'get' his point - but it's just not the way that I want to look at my retirement. It reminds me of the old saw regarding the 6 foot tall statistician drowning in the pool of an average 4 foot depth.

Also, can you point out where it talks about a variable spending model? I missed that (or are you inferring it, from the higher WR?).

A few quotes (bold mine):

Quote:

In fact, 9.96% of the outcomes have a negative present value. This represents a roughly one-in-ten chance of exhausting the portfolio’s assets well before the retiree’s death. Negative RPV outcomes can be thought of as situations in which the retiree has to borrow money from his heirs in order to support the desired spending level (or perhaps move in with them).

Personally, I want more buffer than that to reduce the chance of needing to borrow money from my kids to get by. In my view, if I knowingly (as much as we can 'know' these things), took that high a risk, I should have kept working. At a minimum, I should have pulled my kids in for a consultation - ' Hey kids, how do you feel about 1 in 10 odds of needing to chip in for Mom and Dad, even though I'm quiting work early?' Personally, I wouldn't ask that of them.

Who knows, it may still come to that, but at a WR closer to 3%, I think the odds are pretty low indeed.

And overall and with all this mortality factor considered, I'm still confused as to why such a low EQ% is suggested. He mentions sequential down year risk, but FIRECALC hits these runs too, and they are indeed tied to the failures, but it still doesn't suggest that such low EQ% give better success rates.

-ERD50

[clifp]

I also read the paper and I am very unimpressed. The useful thing in the study is to a factor in longevity risk when calculating SWR. Frankly this is much more important for people retiring at 65 than since 50. The probability that 50 year old mN will die in the first 15 years are only 13%. The probability that 65 year old will die in the first 15 years is 40%. I contend for a 50 year planning on a early death doesn't make sense, for a 65 year taking out 7% isn't a bad idea, as long as the primary back up plans is spending down the kids inheritance.

However his methodology for estimating future returns isn't just different than FIRECalc it is flat out wrong. He assumes a 6% real return and 16% variance for equities. That figure seems ok I am less sure about fixed income real returns.

However the mistake is treating the stock market performance like a flipping a coin. We all know that coin have no memory and the odds that heads will appear in the next flip are 50% even if you've had 6 heads or tails in row. There is no evidence that market doesn't have a memory, and lot of evidence that there is a strong bias toward a reversion to the mean. Other than the great depression, the Dow hasn't had two years in row with statistically significant drop since 1900. If Dow/market returns were uniformly distributed this is something we'd see far more often.

This approach makes equity look far more volatile than they are. If you want to use a different model than FIRECalc, fine but you shouldn't use an event like retirement with 15-40 year time frame and than look at returns on annual basis. I am confident that if he looked at 5 year averages and volatility of equity returns, he end up with a much different asset allocations. Or alternativey he could just use FIRECalc .

[Chuckanut]

I remember John Templeton being interviewed by Rukeyser. When asked what was the best way to generate income, he replied "invest for growth and then sell some of the growth". Of course, Templeton always seemed to be a courageous optimist. I understand he made a fortune during the Depression by buying 100 shares of every stock he could find that was less than $1. Most went belly-up, but the ones that survived !!!!!!

[Huston55]

https://content.putnam.com/literature/pdf/PI001.pdf

I just read this paper (a little late to the party), and the resulting posts. I thought the discussion was useful, still very relevant two years later, which is why I 'bumped' the thread - hoping to stimulate some discussion about AA when entering FIRE in the current 2013/14 environment.

Although Harlow seems very academically qualified ("not a light weight" as someone said), I am like Clifp...very unimpressed by the paper. Here's why:

1. The 6%-8% withdrawal rates for a 65 yo are absurd; especially when contained in a paper purporting to be 'conservative' compared to the norm.
2. Using the SS longevity table, which seems to drive the high WDR, is also not 'conservative', and is likely to leave the retiree wishing they'd not lived so long.
3. The 6%/3%/1% stock/bond/cash 'real' returns seem essentially eliminate inflation from the analysis; again, absurd, since inflation risk from longevity is something that must be addressed.
4. And, last but not least: Appendix B, at the end of the paper, which doesn't seem to be explained, shows exactly the opposite AAs than the rest of the paper. IOW, it contradicts the paper's basic conclusions about optimum AAs. Can't tell if numbers were just mistakenly transposed or what but seems a little sloppy.

I'd like to hear other thoughts on this, especially contrary views - those that support a very low equity allocation entering FIRE, to test the AA and methodology I currently plan to use.

In the interest of full disclosure, my situation is below:
• Planned FIRE AA: 65/25/15
• Will use Guyton-Klinger VWR approach
• Keeping 15% cash to ride out bears
• Have a small/moderate COLAd pension beginning at 60 and retiree health benefits beginning upon retirement.

[Gatordoc50]

10 % equity was an expensive decision over the past few years. Ouch! I'd rather work a few more years if presented with a bad sequence of returns early on than leave so much " good" SD on the table. Actually, that's what I did after the crash. I worked a few more years until some stability returned.

[Ready]

Quote:

Originally Posted by Huston55 In the interest of full disclosure, my situation is below: • Planned FIRE AA: 65/25/15

I think you've discovered the secret to achieving a high WR. Make sure you AA adds up to 105%!

[REWahoo]

Quote:

Originally Posted by Ready I think you've discovered the secret to achieving a high WR. Make sure you AA adds up to 105%!

Hey, don't give it away! It'll lose effectiveness if everyone does it...

[braumeister]

As an old teacher of mine used to say,

"Two plus two equals five, for sufficiently large values of two."

[MRG]

Quote:

Originally Posted by braumeister As an old teacher of mine used to say, "Two plus two equals five, for sufficiently large values of two."

Your teacher was O'Brien, goodness were you in room 101?

MRG

[Chris M]

Quote:

Originally Posted by Huston55 https://content.putnam.com/literature/pdf/PI001.pdf I just read this paper (a little late to the party), and the resulting posts. I thought the discussion was useful, still very relevant two years later, which is why I 'bumped' the thread - hoping to stimulate some discussion about AA when entering FIRE in the current 2013/14 environment. Although Harlow seems very academically qualified ("not a light weight" as someone said), I am like Clifp...very unimpressed by the paper. Here's why: 1. The 6%-8% withdrawal rates for a 65 yo are absurd; especially when contained in a paper purporting to be 'conservative' compared to the norm. 2. Using the SS longevity table, which seems to drive the high WDR, is also not 'conservative', and is likely to leave the retiree wishing they'd not lived so long. 3. The 6%/3%/1% stock/bond/cash 'real' returns seem essentially eliminate inflation from the analysis; again, absurd, since inflation risk from longevity is something that must be addressed. 4. And, last but not least: Appendix B, at the end of the paper, which doesn't seem to be explained, shows exactly the opposite AAs than the rest of the paper. IOW, it contradicts the paper's basic conclusions about optimum AAs. Can't tell if numbers were just mistakenly transposed or what but seems a little sloppy. I'd like to hear other thoughts on this, especially contrary views - those that support a very low equity allocation entering FIRE, to test the AA and methodology I currently plan to use. In the interest of full disclosure, my situation is below: • Planned FIRE AA: 65/25/15 • Will use Guyton-Klinger VWR approach • Keeping 15% cash to ride out bears • Have a small/moderate COLAd pension beginning at 60 and retiree health benefits beginning upon retirement.

Huston,
I think your Point No. 3 on inflation is crucial to shedding light on why these results seem to fly in the face of FIRECalc, as well as the published historical analyses of safe withdrawal rates and asset allocation. At least for the period since 1926, the primary risk to retirement portfolio survival came from high inflation, not bear markets. This can best be seen in the original historical analysis of safe withdrawal rates, done by William Bengen and published in the Oct. 1994 Journal of Financial Planning (http://www.retailinvestor.org/pdf/Bengen1.pdf). While the Trinity study is more up-to-date, and Bengen did some hokey things to extend his analysis beyond 1994, what I like about Bengen’s article is that he actually shows which starting retirement years led to success and which ones failed for various asset allocations and withdrawal rates. And most portfolio failures occurred for retirements that included the high-inflation years of the 1970s, not the Depression. For example, for a 50 stocks/50 Treasuries portfolio and a 5 percent withdrawal rate, 6 Depression-era portfolios failed in less than 40 years (the ones for retirements beginning in 1929-31, and 1934-39). In contrast, 17 retirements that started between 1955 and 1974 were bankrupted in less than 40 years (the only two that lasted beyond 40 years began in 1958 and 1959).

But for a 5 percent withdrawal rate and a 75 stocks/25 Treasuries asset allocation, the number of retirements that were bankrupt in less than 40 years was reduced to only 3 in the Depression era (1929, 1930, and 1937), and to 14 in the high-inflation era (which Bengen referred to as the “Big Bang,” in contrast with the “Big Dipper” and “Little Dipper” bear markets during the Depression).

Although the Putnam paper uses real returns, I think you’re right that it does not really capture the risk posed by volatility in inflation—especially the risk of an extended period of high inflation such as we had from the late sixties through early eighties. And the risk of outliving your savings in retirement comes from two sources—bear markets and inflation. At least in the past, the inflation risk seems to have been the bigger of the two. Any analysis of asset allocations that does not address the inflation risk is bound to yield distorted results—and may be more likely to favor portfolios weighted heavily towards fixed income.

I don’t think the inflation issue is the only problem with the paper—but I think it’s the biggest problem. I would not consider changing your plans based on this paper.

Chris

[pb4uski]

Reminds me of the old joke of an accountant interviewing for a job and the prospective employer says "I only have one question for you - how much is 2 plus 2?".

The accountant replies "How much do you want it to be?"

He got the job.

[Huston55]

Quote:

Originally Posted by Ready I think you've discovered the secret to achieving a high WR. Make sure you AA adds up to 105%!

Oops. Meant 60/25/15.

Tough to give back that 5% though.

[NW-Bound]

Quote:

Originally Posted by Ready I think you've discovered the secret to achieving a high WR. Make sure you AA adds up to 105%!

OK, here's an easy way to do it, and you do not need to stop at 105% either.

There are many double and triple-leveraged ETFs on the market now. Your $1 in these ETFs will gain or lose as much as $2 or $3 invested in the equivalent indices.

I have used some of these in the past to do short-term market timing, when I did not want to sell some of my positions to fund the purchase, nor wanted to redeem from I-bonds which I could not easily buy back at the same rate.

I don't think I ever had more than 2 to 3% of portfolio in these 2X and 3X ETFs though. They are also not good for long-term holding, as the 2X or 3X factors are only guaranteed for short-term.

Quote:

Originally Posted by REWahoo Hey, don't give it away! It'll lose effectiveness if everyone does it...

Nah! Most people do not have the guts.