Retiring In Secular Cycles

[Cut-Throat]

A question for the diehard rebalancers in the audience:

If you've bought into the idea of rebalancing based on some arbitrary period or %-out-of-whack, why doesn't the idea of rebalancing based on valuation metrics appeal to you? Or does it?

Hoc*us does! - He's rebalanced into 0% equities and is waiting for the big drop to get back in!

 

[audreyh1]

A question for the diehard rebalancers in the audience:

If you've bought into the idea of rebalancing based on some arbitrary period or %-out-of-whack, why doesn't the idea of rebalancing based on valuation metrics appeal to you? Or does it?

Well, I guess if there were a paper out there that demonstrated the benefit and provided a simple method, I might consider it. But otherwise it doesn't really appeal to me.

I mean - rebalancing already is doing a comparison of relative valuations of different asset classes. But, of course, it's all relative and depends on when you started. And if all asset classes seem - "overvalued" - well, I always hold a certain % in cash anyway. Sometimes overvaluations are resolved by a sharp correction, sometimes they are resolved by moving sideways for a long time. Sometimes overvaluations last a long, long time. Either way rebalancing seems to work reasonably well.

The valuation trends can take decades, generations. Each generation has a different attitude towards different asset classes based on their prior experience, and this can drive valuations up or down for long periods.

Audrey

 

[justin]

This is still data mining. Someone decided on using quartiles instead of quintiles or sextiles. I'd guess if you had quintiles, today's PE would fall within the second quintile, and we would have ~100% success in the second quintile, and 75% success in the first quintile (where there would be significantly higher PE's).

I might consider PE ratio as a qualitative measure in deciding "Am I there yet?". For example, if PE's doubled to 35 or so in a year even though earnings remained fairly level, I'd be thinking "bubble" and wouldn't count on my portfolio (which would have just doubled in the past year also) to keep growing at the same rate. In fact, I'd be expecting a high probability of a correction. In other words, if my portfolio went from $750,000 today to $1.5 million (my magic number) in 2008 mainly because of the expansion of the PE ratio, I wouldn't feel comfortable ERing.

 

[haha]

This is still data mining. Someone decided on using quartiles instead of quintiles or sextiles. I'd guess if you had quintiles, today's PE would fall within the second quintile, and we would have ~100% success in the second quintile, and 75% success in the first quintile (where there would be significantly higher PE's).

 

[DRiP Guy]

quintiles...
sextiles...

septiles?
quincunx?

 

[crestmont]

Hello, this is Ed Easterling. Hope you don't mind if I post a few comments for your consideration. First, I've enjoyed reading your thread today...there are sure a lot of interesting comments and thoughts. As several have noted, it was not my intention to (1) present a bullish or bearish view of the market, (2) suggest that anyone not invest in equities, or (3) provide an assessment or recommendation for the optimum portfolio or investment approach. The article was intended to use the research and observations from Crestmont Research and Unexpected Returns and apply it to SWR.

Over the past few years (and I think originally in this forum or a related one), I have received a number of questions about SWR. Particularly recently, for whatever reason, a number of investors and advisors have asked about it. As a result, the analysis was performed and "Destitute" was published. Mauldin like it and asked to use it in Outside The Box. It's also available on the Crestmont Research website.

The article does shed insight into the fallacy of averages and, in particular, that valuation matters--it affects the ultimate return from financial assets. Further, the analysis quantifies it. I hope people will not read into the article more than was intended. As with Crestmont's other research and the book Unexpected Returns, the objective is to provide thought-provoking insights and perspectives...not to impose investment outlooks, recommendations, or advice. As visitors to the website know, it's an open-access site without cost, ads, or registration. The site, and visitor feedback, has been instrumental in furthering my research--thanks to any of you that may have sent comments or suggestions.

Lastly, to address a couple of comments about the inclusion of 100% equities in the article. That was (1) to illustrate the impact of stock valuation on success and not complicate the analysis with more assumptions regarding asset mix or rebalancing frequency and (2) to respond to a number of comments that I heard recently about advisors or investors increasing significantly the equity portion because the retirees "needed" higher withdrawals and since they would have a long time (30 years), equities would do just fine. For the younger member of this forum that have a long time, please see the article "Waiting For Average" to understand that time will not make up for above-average valuations. The stock market is not a machine that produces 10% returns (the long-term average) if you stay long enough. The starting level of valuation does matter.

After reading your posts, I took a quick look at including bonds. The results are that the current level of bond yields (5% or so) reduces the success rate (assuming annual rebalancing) at all stock/bond mix ratios. Interestingly (at least to me), yields need to be closer to 7% to provide help to the success rate. Also, for a 100% bond portfolio mix, it appears that it would take over 7.5% yields to provide 100% success for a 4% SWR because of the impact of inflation on withdrawals over longer periods.

Sorry for the long post, but I hope this helps to address some of the comments preceding it in the thread. Thanks for your interest (and criticisms...it drove the look at and insights into adding bonds) in the article. All the best, Ed


P.S. Relating to the comments about the current level of reported P/Es, please see "The Truth About P/Es" in the stock market section of the www.CrestmontResearch.com website. The late stage of the earnings cycle is currently distorting P/Es lower. Using a variety of recognized techniques to adjust earnings (multi-year averages, regressions to the economy, etc.), we are further into the top quartile that it appears.

 

[DRiP Guy]

I for one thank you for your post. Very though provoking... (sound of mouse clicking as I go to dig into the links...)

I do appreciate your very upfront self-characterization [bold added due to applicability to this topic]:

...provocative insights on the financial markets and on the hedge fund industry.

... concentrated on the drivers and characteristics of secular stock market cycles, the impact of inflation and interest rates on the stock and bond markets, and various aspects of hedge funds and the hedge fund industry.

I also think this approach you end your opening page with is laudable, although difficult for most humans (me, too!) to hew to:

We solicit your insights at Info@CrestmontResearch.com, whether supporting or contradicting our presentations. They will assist in furthering our research.

8)

 

[wabmester]

I still think that looking at historical US returns for SWR's is wildly optimistic. And I'm waiting for the Japanese market SWR study.

 

[Cut-Throat]

This all assumes that P/E ratios are the correct valuation of Stocks. - If you can apply this to stocks why not real estate or any other investment?

 

[haha]

Hello, this is Ed Easterling.

Hey, welcome to the board. I hope you stop by as often as you can.

For the younger member of this forum that have a long time, please see the article "Waiting For Average" to understand that time will not make up for above-average valuations. The stock market is not a machine that produces 10% returns (the long-term average) if you stay long enough. The starting level of valuation does matter.

As I see it, the only way this could be falsified would be if the corporate economy were to move to a permanently more profitable position- hard to do from where we are starting IMO.

By the way, I bought, read, and kept you book. I think it is excellent. Now if my bearish positions would only work out better...

I still think that looking at historical US returns for SWR's is wildly optimistic. And I'm waiting for the Japanese market SWR study.

Possibly so, but not on this board. Plug 'er into FireCalc and fire away! Who cares if you don't understand or don't pay attention to nuance?

On this board to even mention that the highest quartile of PEs brings a lower survival rate was not exactly embraced. Stocks are the opiate of the affluent.

This will not change until and if market performance falters dramatically.

Ha

 

[donheff]

After reading your posts, I took a quick look at including bonds. The results are that the current level of bond yields (5% or so) reduces the success rate (assuming annual rebalancing) at all stock/bond mix ratios. Interestingly (at least to me), yields need to be closer to 7% to provide help to the success rate.

This issue is confusing. When I first started looking at various Monte Carlo simulators I came to the same conclusion - higher equities = higher success, period. Thus 100% equities would be the way to go. But a number of posters have cited a variety of studies that indicate that, in addition to reducing volatility, a portion of bonds improves SWR success rates. I assumed this actually related to the same factors that seem to influence your valuation effect -- a prolonged bad market in the early years of the withdrawal phase leave you with too little time to recover. Thus the bonds (and reduced volatility) reduce the overall impact in an early year bear market. (It may also be that bonds tend to rise during stock bears?)

If this is not the case, would it be your opinion (not asking for a recommendation) that, with respect to setting a SWR, 100% stocks must always be the preferred approach over a mixed stock/bond portfolio? I.e. to get the highest practicable SWR from a given investment amount, go for 100% stocks? (I recognize that if you can afford a smaller WR a mixed, or even all fixed portfolio could assure a guaranteed income)

 

[wabmester]

I think Ed modeled bonds only in terms of yield rather than total return. In the past 25 years or so, bonds have returned much more than the yield due to declining interest rates. I think FIREcalc models bonds as if you sell them each year, so it's a mix of yield and cap gain/loss. Same with stocks.

 

[crestmont]

I'm not sure how to reflect referenced quotes (or for that matter, why I have a problem with dryer sheets!!!), but I think the SWR issue is impacted by the interaction of the withdrawal rate and the bond rate (i.e. as the initial withdrawal rate declines, it takes a lower bond yield to provide benefits to a stock portfolio.

Yes, I assumed 'yield' for the quick analysis since the presumption would be that today's investor would be "stuck" with current yields. Even if there were total return gain/losses, the reinvestment rate would be at market yields and thus would imbed the now current yield.

The issue that we're all discussing revolves around the withdrawal rate that can be sustained based upon the initial valuation (which is really 'expected return' or 'market return') of the core financial assets (stocks and bonds).

Further, we're dealing with the fact that under certain--yet limited--situations when high starting valuations drive poor returns. And although the statistics provide for apparently high success rates even in high valuations, the propensity for failure rises significantly. I am awed that almost all scenarios provide such large ending values, yet the minority scenarios expire a portfolio. And since zero is so incredibly serious, the prospect of failure becomes more than just a low probability. At some point for many of us (certainly for me), running out of $ at 80 in completely unacceptable (yet I respect anyone willing to risk it even when the odds are very low).

 

[Nords]

I am awed that almost all scenarios provide such large ending values, yet the minority scenarios expire a portfolio. And since zero is so incredibly serious, the prospect of failure becomes more than just a low probability.

Pretty impressive what happens when you set the feedback gain and then just turn your back for 30-40 years, huh?

I think that the models don't reflect the very real human behavior of reducing spending, raising cash, and possibly even finding part-time work. But they sure scare a lot of people into being more conservative than perhaps is warranted.

 

[DRiP Guy]

...zero is so incredibly serious, the prospect of failure becomes more than just a low probability.

I liked your taking a moment to define risk management on your site, differentiating-- [paraphrasing] it is NOT about maximizing gain, it is about avoiding the few completely unacceptable outcomes.

That said though, once all the analytical smoke clears, I keep coming back to a principle that I think I intuited myself, but first saw reduced to print and a strategy by Peter "gummy" Ponzo -- it would be a foolhardy retiree indeed who simply drew and drew at some preordained fixed "SWR" rate, even as he saw his nest egg dropping far below plan expectations due to poor market conditions, until it was right down into the dirt.

Instead, I think most prudent people would rein in their withdrawals early, to allow some recovery long before a crisis.

That said, though, two things THEN occur to me (do I seem to be vacillating? ) :

1) That self-preserving 'escape clause' should not prevent us working on strategies that are robust enough to avoid worrying that such ad hoc belt tightening would ever be needed,

2) The majority of people I see considering these topics are highly motivated, and deeply engaged in exploring their own retirement planning. I do fear there is coming a class of Early Retirees who will not be nearly so involved in their own planning, and grappling with the risks and decisions as today's FI/RE 'pioneers' seem to be doing. So perhaps they would fly the thing right into the tarmac unless there are warning bells...

 

[Cut-Throat]

I am awed that almost all scenarios provide such large ending values, yet the minority scenarios expire a portfolio. And since zero is so incredibly serious, the prospect of failure becomes more than just a low probability. At some point for many of us (certainly for me), running out of $ at 80 in completely unacceptable (yet I respect anyone willing to risk it even when the odds are very low).

Yet, anyone that has studied this problem would detect that money was getting a bit tight long before age 80 and reduce spending to a point that running out of money would be almost impossible.

Or have you never heard of the "Trout Bum in a trailer park down by the river lifestyle"? 8)

 

[DRiP Guy]

Pretty impressive what happens when you set the feedback gain and then just turn your back for 30-40 years, huh?

I think that the models don't reflect the very real human behavior of reducing spending, raising cash, and possibly even finding part-time work. But they sure scare a lot of people into being more conservative than perhaps is warranted.

Hey, how come I didn't see this before I just posted?

I could have saved my breath, you already covered the same point, only with a LOT less words!

 

[Cut-Throat]

Hey, how come I didn't see this before I just posted?

I could have saved my breath, you already covered the same point, only with a LOT less words!

Me Too!

 

[DRiP Guy]

Me Too!

Well, your post wasn't wasted though... it got me thinking of the LBYM lifestyle, obviously it propels us to be more likely to FI/RE, but also post-accumulation it is great to already be adaptable in the short term (delayed gratification, like we did when we saved our paper route money) in order to preserve the long term health of the plan.

I have not formalized this in any way in my own head, but I would be willing to bet that many ERs think about this the same way I do:

The nest egg driving the available SWR is the golden goose to me. It must be preserved. So during the salad days, if the SWR is good, and I am taking out SEPP or otherwise moving from the Goose to my local storage (MM fund or whatever), I am certainly not going to be blowing all 4.227% of an available 4.227% -- I will likely be socking away a growing (hopefully) mini-surplus, so that if the SWR does need to dip to 3.11% for a while, I can easily use some of the acorns stored closer to my little nest, figuratively speaking.

 

[Gone4Good]

I'm waiting for the Japanese market SWR study.

This would be very interesting.

It may prove surprising too. Certainly an all equity portfolio would not have faired well. And anyone who treated their house like a bond for allocation purposes would be out on the street. But a 60/40 mix, coupled with the general price deflation experienced by Japan, may actually have been survivable for 30 odd years. . . don't know, but it would be interesting to see.

I'm cautiously optimistic, too, that US equity returns over the past 100 years were not an anomaly. US economic growth has largely resulted from a free enterprise focus that large portions of the world are now starting to embrace (continental Europe notwithstanding). Although the US will not likely lead the main charge in the 21st century, places like India and China have the ability to drive global economic growth the way the US did in the century past. No reason we can't ride along.

 

[crestmont]

I'm cautiously optimistic, too, that US equity returns over the past 100 years were not an anomaly. US economic growth has largely resulted from a free enterprise focus that large portions of the world are now starting to embrace (continental Europe notwithstanding).

Please keep in mind that the most-often quoted data--Ibottson's Annual Yearbook--only reflects long-term returns of 10% because the series starts in 1926 when P/Es were 10.2 and ends when P/Es are almost twice as high. Starting P/E impacts the dividend yield component of total return as well as the capital gain component. Had P/Es at the start been the same as today, long-term returns would have been (necessarily) less than 7%. That alone would be a major change to conventional thinking. The starting point has a major impact.

 

[DRiP Guy]

And I know that CRSP provides databases that clean up biases associated with survivorship and M&A issues, but not sure Ibbotson or others do, so that may also be a confounding factor as well...

 

[DRiP Guy]

Hmm... thought this was interesting. I knew small caps were outperforming, but did not know for how long or to what degree:

Ibbotson also breaks out the performance of each size decile, or 10% slice of the market. What’s interesting is that the returns are almost perfectly rank-ordered—the smallest 10% has done the best, and the largest 10% has done the worst.

Since 1926, the smallest decile has returned an average of 13.96% a year. My only caution about micro-cap investing is that although the “outperformance premium” is very real, it’s not very well-behaved. The relative performance is highly cyclical. It’s either feast or famine.

http://www.crossingwallstreet.com/archives/2006/04/ibbotson_yearbo.html


Not likely to change my own style on such intriguing things though. Slow and steady, broadly diversified, market basket of all stocks...

 

[bpp]

I still think that looking at historical US returns for SWR's is wildly optimistic. And I'm waiting for the Japanese market SWR study.

This came up on raddr's board last May:
http://www.raddr-pages.com/forums/viewtopic.php?t=2546

Maybe I'll do an update through 2006 at some point, but the basic results won't change: 3% and diversification starting in 1990 would have left one intact (so far).

 

[DRiP Guy]

Please keep in mind that the most-often quoted data--Ibottson's Annual Yearbook--only reflects long-term returns of 10% because the series starts in 1926 when P/Es were 10.2 and ends when P/Es are almost twice as high. Starting P/E impacts the dividend yield component of total return as well as the capital gain component. Had P/Es at the start been the same as today, long-term returns would have been (necessarily) less than 7%. That alone would be a major change to conventional thinking. The starting point has a major impact.

While not showing P/Es, this chart from FPA 1 IMHO shows pretty well that the 10+% returns averaged since ~1990 are atypical, and if past is prologue, we should look for ~6.7% to return. Of course, you seem to be telling us that if P/Es don't move down, we may not get that, either...

Note 1 - linked here under non-commerical fair-use provisions