4% of what?

[hix9]

hi sg

I have to say that I think use of "reversion to the mean" as a stock investment indicator is pure horse puckey.... Observed "reversion to the mean" may not happen for many years and that "reversion" could happen at a time when the mean is even higher than it is today.

If you need any consolation that you are not alone in the view that mean reversion might not be a certainty useful to investors, there is always SSRN:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=683182

You have to read the paper to determine if you agree with their approach but here are a couple quotes to spark interest:

"Perhaps most interesting, however, is that the nonstationarity of the E/P ratio suggests that this valuation measure can remain below its mean for an extended period of time, and that its reciprocal, the market P/E, can stay above trend for extended periods - and possibly forever, at least theoretically. The P/E ratio's shift from a stationary to a nonstationary series ca. 1960 implies that it no longer has a mean to which it must revert."

...

"It is clear from examining the graph that the general negative relation between starting P/E ratios and 10-year real returns abruptly truncates at P/E ratios greater than 21. Ten-year returns starting from very high P/E ratios have never been severely negative. Although the overall trend of the P/E and stock return relation is negative, the lowest 10-year real returns are earned starting from market P/E ratios between 12 and 20. Significant declines in the value of U.S. equities starting from high levels of the market P/E ratio are rare."

hix9

 

[haha]

I have pretty much given up on trying to discuss these complex things. I read the paper referenced above. Honestly, I can't really evaluate it, at least not without putting more effort in than I think it merits.

The authors choose not PE10, or peak earnings PE, but your standard 12 month PE. Well, there are known problems with this measure. That is why it was rejected by Campbell and Shiller.

Two other things caught my eye, but not enough to really analyse them, or try to make a critique. I'll just mention them briefly. They mention the PE21 breakover, such that although high PEs are associated with low future returns, very high PEs (>21) are associated with normal returns. I know there may all kinds of hoop-jumping expanations for this oddity, but my guess is that they are bogus. It's like saying up to a threshold, the more HIV in your blood the worse off you are. But if you have a hell of a lot of it in there, cool, you'll be fine.

The other problem I have is related- How often have there been PEs over 21, unless PEs were artificially raised by severely depressed earnings? Very seldom, and most instances during the bubble, which in my opinion is still an unresolved chapter in financial history.

So take your pick. As is often the case, people have placed their bets. All we can do is wait for the race to be run.

Mikey

 

[Mr._johngalt]

"reversion to the mean" amounts to an academic
exercise for most people; nothing more.
(Not that there's anything wrong with that ).

Guessing and predicting can be fun.
I enjoy it a lot, but it is of little help in working
out any meaningful financial plans.

JG

 

[unclemick]

Bad company=good stock, good company=bad stock.

I don't care about academic's, I just use mean revision.
Sometimes it's a long wait.(think Japan, gold)

And, and to make life interesting: per Warren Buffett - sometimes a wet cigar - is a wet cigar.

 

[hix9]

hi mikey

I have pretty much given up on trying to discuss these complex things.

i think i enjoy reading these types of discussions but i have given up (or more accurately never started) trying to apply any of the conclusions so i suppose it is not always time well spent.

hix9

 

[unclemick]

This is the bubblegum/dryer sheet forum.

However - sometimes you get a link to something you can use.

 

[unclemick]

I think it's a hoot that most posters jump on the 4% and totally miss the 'of what' part.

Like the famous analogy of blindfolded men feeling parts of the elephant and making decisions/guesses

Blindfolded or not - you don't want to touch the 'wrong' part.

Heh, heh, heh, heh.

Vanguard Balanced Index - 2.53% current yield
Vanguard Wellesley - 3.60% current yield.

Psst - the Norwegian widow doesn't use no blindfold.

 

[haha]

Mikey,
I think most of what we end up discussing here are investment aprroaches. Your approach is similar to Buffett's as you are always looking for value, and investing is something you enjoy.
That doesn't mean that folks that invest in Vanguard's Retirement strategy funds and let her ride are wrong or are even betting that the market won't take a dive soon. What they are saying is that they don't have the confidence or want the hassle of trying to figure out which way the market will go or when it will drop.

Cut-Throat, that is very well put. You are good at pointing things out without stepping on someone's feet. I am trying to get better at that, but it's a slow slog for me

So the index folks always have their bet placed that over a 30-40 year period stocks will overall increase their value from today. They just have no clue when and hence are always invested. Again, this doesn't mean that they think stocks are cheap now or that stocks won't plunge by 70% in the next ten years. It just means that they'd rather go fishing and not have to worry if they should be calling their broker when they get off the stream.

Neither approach is wrong!

Again, well said. I think stocks will do well also. I just think (hope) that I can continue to do somewhat better or somewhat safer than buy and hold.

I was idly thinking about this as a Bayesian problem. What we tend to argue on this board is "what is the proper reference group or base case?" Buy and hold indexers implicitly or explicitly believe in the broadest reference group, while timers or stock pickers believe in identifiable, reliable sub-categories.

Gives food for thought anyway. Thanks for a wise post.

Mikey

 

[peteyperson]

I have pretty much given up on trying to discuss these complex things. I read the paper referenced above. Honestly, I can't really evaluate it, at least not without putting more effort in than I think it merits.

The authors choose not PE10, or peak earnings PE, but your standard 12 month PE. Well, there are known problems with this measure. That is why it was rejected by Campbell and Shiller.

Two other things caught my  eye, but not enough to really analyse them, or try to make a critique. I'll just mention them briefly. They mention the PE21 breakover, such that although high PEs are associated with low future returns, very high PEs (>21) are associated with normal returns. I know there may all kinds of hoop-jumping expanations for this oddity, but my guess is that they are bogus. It's like saying up to a threshold, the more HIV in your blood the worse off you are. But if you have a hell of a lot of it in there, cool, you'll be fine.

The other problem I have is related- How often have there been PEs over 21, unless PEs were artificially raised by severely depressed earnings?  Very seldom, and most instances during the bubble, which in my opinion is still an unresolved chapter in financial history.

So take your pick. As is often the case, people have placed their bets. All we can do is wait for the race to be run.

Mikey

Mikey,

I have seen similar reasoning from Siegel in papers prior to his recent book. In the case of Siegel he made the point that even buying the nifty-fifty stocks decades ago at their peak would have worked out fine over history even when their collective PE was high and some stocks had PEs over 80! The reason is because whilst the market fell 40% in 73-4 and the most highly valued lost 80% odd, you were then able to reinvest higher dividend payments to buy many more shares of the ones that fell the most and this compensated for the original high price.

There are several problems with this analysis. Firstly as Siegel seems to commonly do, it ignores the reality of taxes on dividend payments. This makes any findings inaccurate and possibly completely wrong. Secondly, one has to be able to cope with a bad decade, then a massive decline in the nifty-fifty, then reinvest for another decade plus to be able to get the average return even ignoring fees and taxes. This may work for someone at the start of a 20-year accumulation phase but it absolutely will not work for someone living off their investments. You'll be living off the dividends not reinvesting them so large capital declines cannot be recovered from using the S&P 500 historical real growth rate of 1.8%. Just not gonna happen. So this is pretty irresponsible to put forward the idea that buying at high prices - even above PE 21 - works. The truth is it only works in specific circumstances, exposes the investor to high risks to capital value over reasonable investor timelines and prevents the ability to later live off investments. Just ask the year 2000 retiree how they feel with the S&P 500 flat 5+ years later ignore fees, taxes and inflation! Theory and practice are two entirely different things.

The paper is BS. Demonstrates a lack of intelligent insight, and just takes numbers on their face value.

Petey

 

[peteyperson]

"reversion to the mean" amounts to an academic
exercise for most people; nothing more.
(Not that there's anything wrong with that  ).

Guessing and predicting can be fun.  
I enjoy it a lot, but it is of little help in working
out any meaningful financial plans.

JG

Not necessarily.

Appreciation of valuations would presently have one placing more int'l and EM and less in the US which has been on a tear and sits rich in valuation as a result. That is to say, reversion-to-the-mean is likely to happen and likely to be more costly for those invested in the broad US market. If one rebalances to set allocations that can also wring out some of the excess but it is far from academic. It can be used both for risk control and to generate excess returns through tactical allocation.

Petey

 

[peteyperson]

Mikey,

When Petey says that the 4% SWR guideline is plain wrong. This is his opinion. And in my Opinion he's wrong. That is what propelled this thread - different approaches.

For clarification, I feel that the US market is sufficiently overvalued that returns when allowing for mean reversion will be poor over the next decade or two. As such a SWR of 4% from the US total market is not viable. Perhaps with value style. In terms of a global asset allocation, with the addition of oil & gas, precious metals and global REITs & TIPS, this provides enough diversity than a SWR of 4%-5% should be possible.

The key is that not all other asset classes are priced the same and so return expectations are different. It will also be possible to obtain a rebalancing bonus when sufficiently diversified and this will help also. Rebalancing either the portfolio or with new capital deployed into asset classes with the highest future expected returns.

An allocation of 60% to S&P 500/Russell 3000/Wilshire 5000 and 40% to US bonds will not provide a 4% w/r over the next 1-2 decades unless markets do not mean revert, dividend payout ratios increase, real earnings growth leaps up or stocks get bid up substantially into a bubble pushing capital returns higher than supported by fundamental earnings. This viewpoint is backed up by numerous observers including Warren Buffett, Jeremy Grantham, etc., all whom see low real returns for a decade or more due to the reasons outlined above.

All the best,
Petey

 

[Cut-Throat]

An allocation of 60% to S&P 500/Russell 3000/Wilshire 5000 and 40% to US bonds will not provide a 4% w/r over the next 1-2 decades

Whoa Petey! Considering that a 60/40 stock portfolio would survive 30 years with a 4.4% withdrawal rate in FireCalc over the last 120 years, you are predicting that it won't survive even 10-20 years. So when is the next Great Depression?

- That's a prediction that Buffet, Bogle and Berstein would all disagree with.

Better check your math dude!

 

[laurence]

Petey, I'm sorry if I'm asking you to repeat yourself, but did you mention how your portfolio is currently allocated? Did someone say you are in Great Britian?

 

[RockMiner]

A simple plan for those in anguish over high valuations:

1) Sell down till you can sleep

2) DCA back in over 5 years or until

3) A Bear crushes the market, then load up.

if P/E's "regress" during that time you might look smart...if they don't, then maybe it is different this time

 

[peteyperson]

Whoa Petey!  Considering that a 60/40 stock portfolio would survive 30 years with a 4.4% withdrawal rate in FireCalc over the last 120 years, you are predicting that it won't survive even 10-20 years. So when is the next Great Depression?

- That's a prediction that Buffet, Bogle and Berstein would all disagree with.

Better check your math dude!

Hi CT,

I think you need to re-read my reply, mate!

I said I did not think it would produce 4% real. Never said the portfolio would not survive!

The 4% comment is based on 1.9% div, 1.8% real growth historical = 3.7% real. PE 19.7x on wider than historical profit margins i.e. earnings higher than would normally be the case. PEs will likely compress as market mean reverts. This will take 1% off over 20 years. Thus 2.7% real. Bond yields are not more than 5% currently, so any stock/bond mix of total market components does not look good for close to 4%.

Stock multiples could get pushed up like 2000 to temporarily boost returns. Real earnings growth could accelerate in a jobless recovery... Dividend payout rates could increase. But will any of that be enough and what state will the market valuation be in then? Any way you look at it - not pretty.

Petey

 

[sgeeeee]

For clarification, I feel that the US market is sufficiently overvalued that returns when allowing for mean revision will be poor over the next decade or two. As such a SWR of 4% from the US total market is not viable. . . .

You are certainly entitled to your opinion. But as far as I can tell, there is no analytical basis for your 4% conclusion. Valuation using P/E does not predict future stock return. I think people are often misled by the term "valuation". It sounds like a number that would be important to stock performance, but it is actually a backwards looking metric that is a very poor predictor of optimum future investment decisions.

Similarly, reversion to the mean is not based in fundamental analysis and even if it were does not predict future stock performance. There are many factors that contribute to P/E and reversion does not require that stock prices drop.

 

[laurence]

Much thanks, Pete. I'm trying to figure out how to best take a position in REITs and natural resources myself. Just don't know enough.

 

[Cut-Throat]

Petey,

I did re-read your reply and you did say 4% w/r. In fact I've got what you wrote quoted. Maybe you should re-read what you wrote.

I would agree with you that the market may not produce 4% real. But this thread is under SWR. And SWR # Real. Two different animals.

I only need  about 1.5% real for a 4% SWR.

So the 4% w/r is a good guideline.

 

[peteyperson]

Petey,

I did re-read your reply and you did say 4% w/r. In fact I've got what you wrote quoted. Maybe you should re-read what you wrote.

I would agree with you that the market may not produce 4% real. But this thread is under SWR. And SWR # Real. Two different animals.

I only need  about 1.5% real for a 4% SWR.

So the 4% w/r is a good guideline.

It depends purely on the asset allocation mix (and I did re-read my reply before, I was clear the first time).

Even if one assumed no reversion-to-the-mean, you're still looking at today's dividend yield and real earnings growth. As we don't know the earnings growth we have to use historical long-run data for that, smooth it out. 1.9% + 1.8% = 3.7%. Add bonds in the traditional US split and you're not seeing 4% real there. All the past research in the world won't make it so. The past research is based on historical dividend yields in the US of 4.5%. Logically a 4% w/r is possible with that and some bonds. The same logic shows the fallacy of relying on what is past to prove what is future. As Buffett said recently, future returns in the US total market he expected to be no more than 6.5% nominal. Fundamental returns are 3.7% real for the US total market, and not 6.3% real as they have been since 1900. The difference is simply lower dividend yields today with no increased real earnings growth to compensate!

If one held a globally diversified portfolio with a mix of assets then the situation changes. A 75/25 mix of S&P 500/US total bonds delivered 4% SWR. A mix of 37.5% S&P 500, 37.5% EAFE index & 25% US bonds would have delivered 4.7% SWR. The addition of value slant would have boosted yields on cheaper purchases, added a value "premium" and boosted SWR still further, as would emphasis on small cap, O&G or REITs. This gets you to 5% SWR from present valuation levels! S&P 500 on present fundamental returns won't do that. They did in the past, but the *makeup of returns* was different.

So we'll have to agree to disagree.

Petey

 

[Mr._johngalt]

Hello everyone. Hey "agree to disagree"........... that's
the gentlemanly way to handle it.

What I am wondering is how many others basically
ignore SWR (in the traditional sense) and manage
ER purely by income/asset (net worth) preservation?
I enjoy all the SWR debate and have even run FIRECALC
but pay absolutely no attention to it in "real life".

JG

 

[peteyperson]

Much thanks, Pete.  I'm trying to figure out how to best take a position in REITs and natural resources myself.  Just don't know enough.

Hi Laurence,

The applicability of REITs (US or Int'l.) depends on what your goals are. In accumulation phase they can diversify you but in distribution phase they come into their own by substantially boosting portfolio yield while maintain capital value buying power. Handy alternative to bonds for many (holding up to 20% US REITs provided higher risk-adjusted return than stocks & bonds alone). Int'l. REITs allow one to diversify more globally that the traditional US-only REIT perspective but you lose the connection to US CPI numbers (as you do with foreign stocks).

More foreign REIT funds are coming out. Range of funds, CEFs and ETFs will be much wider in a year or two than today I suspect. Costs will also come down.

In terms of natural resources, tricky area. Many different opinions on prospects, cheap vs expensive and so on. Having looked at it a bit I think one would really need to dig deep and spend a lot of time to learn & keep up with the sector. For the kind of allocation I would consider not really worth doing that. It is important to note that most try to predict which sectors will perform well but with something like oil & gas it has significant benefits in terms of correlations. Oil rises, oil co. profits but corporate profits fall. Stocks of the two can move in different directions. While oil co. included in indices do influence indice performance, owning separate energy fund allows for rebalancing, profit harvesting and more steady returns during distribution phase. It is also useful to mitigate rising inflation although oil is a small component of CPI overall. I would not own o&g or nat. res. to beat the market but for the diversification it provides in returns during accumulation and distribution phases.

I like Vanguard Energy ETF and Vanguard PM fund. Minimums high on the latter.

Hope that was useful to you.

Petey

 

[Cut-Throat]

Did someone let that flock of geese back in here?

 

[cute fuzzy bunny]

Hey guys, Petey has a different opinion (and it looks like its being expressed as an opinion), and he may be right. At least he's presenting it reasonably.

It bodes well to know and understand all of the viewpoints and prospects for the future. For many of us, that produces a Bogle-ean (is that a word?) "...and then do nothing" response. For others, they might like to have all the info in their brains.

While the "...and its always worked out that way historically for the last 100-and whatever years and something different happening in the future is unlikely to be worse than the depression or the 60's-70's sideways" sounds good. Look at a chart of the S&P500. It was pretty clean and smooth until 10 years ago. Something different HAS happened, and something different MAY happen going forward. Its NOT the same-old same-old.

Having had this feeling for several years now, I've avoided owning TSM/S&P500 type indexes since 2000. So far, thats a good thing as the S&P500 growth over the last 5 years has been zero. Inexpensive actively managed balanced indexes full of value stocks have been pumping out my income and growing my stash in the meanwhile.

 

[peteyperson]

Golllly petey -- THAT"S GREAT NEWS!! -- A couple years ago Buffett was absolutely convinced 7% for the next 5-10yrs was the limit. This prediction was followed by market returns of ~ 28% and 10%. Effectively using up 5 yrs of gains in 2... lol - I was afraid the next 3 yrs were going to be flat...

BTW: Don't you 'mean' reversion as opposed to "revision." Hmmm...
If you want to look at reversion to the mean, it would probably be best to isolate the cyclical stocks <vice entire market> - look for their bottoms to buy and sell at their tops... A strategy like 'dogs of the dow' would probably suit your beliefs. At least with blue chips your downside risk is somewhat mitigated. As you will probably notice with the p/e's and dividends these dogs are carrying - they do have their share of fleas and risk.

Reversion. Indeed. This burning the midnight oil is catching up with me!

Petey

 

[cute fuzzy bunny]

A lot of companies books are getting a mean revision lately