Bogleheads

[cjking]

I calculate fair value for the S&P 500 using "q" as my indicator. I get end-of-quarter data off a chart at Smithers & Co web site and adjust for time and dividends since then to get fair value for the S&P 500 on 1st of July 2008 is 1261. The index looks like it may cross that threshold today. If you look at the chart (I'm not allowed to include a link) you will see this is the first time since about 1990 that US shares have not been overvalued.

(If you google for the website advertising the book "Valuing Wall Street", which explains "q", and which I strongly recommend, you will find a link from there to Smithers & Co's website.)

[cjking]

I use fair value to calculate the real return I expect on shares. For example, when shares were 9.2% overvalued I would have expected a real return of 6.7% / 109.2%, less fund management charges.

According to this method, at the peak of the year 2000 bubble, the expected real return on shares fell below 2%, an all-time low.

[Russell]

Hi CJ --

Boglehead and Economist Ed Tower has done some work with one of his students investigating Tobin's q (motivated by Smither's book, I believe) and came away favorably impressed.

You might interested in their findings -- here's a snippet:

[Ted Valentine]

The bottom is here! Let the speculation begin! This time its different!


_________________
“A bear market is an extended period of time during which people who think this time is different sell all their investments to people who understand that this time is never different.”

[psteinx]

A while back, I played with a bunch of different methods for estimating future stock market returns. The one that was most successful (i.e. best at estimating future returns), was a 'q' variant. It worked better than various measures of PE or EY, even when the latter were adjusted in various ways, using smoothed or multi-year earnings.

That said, I'm reluctant to label any particular value of the S&P as a fair value or a true value. Assuming we can estimate with some accuracy what future cash flows to share holders will be (basically, future earnings), then computing a fair value for the current market depeneds HEAVILY on what discount rate we use to compute the future value of those cash flows.

To turn it around and look at it the other way, we can perhaps say that, absent future changes in valuations (i.e. assuming the market PE 5/10/20 years down the line is about what it is today), and using a stabilized value of the current earnings yield and some plausible assumptions about growth and payouts, we can estimate that future market returns will be about F%. The value for F will change as the market goes up and down. We can then compare F to historical rates of return (let's call that H%). Perhaps one thinks that the market is fairly priced when F ~= H, or perhaps when F >= H. I don't buy that, though. Just because the market returned H historically doesn't mean that it must return that going forward, nor that things are overpriced if future returns (F) are expected to be less than historical returns (H).

Generally, we think of returns as gross values - the return on some calculated index, which makes no provisions for transaction costs, and implicitly assumes that the average investor could historically own that index relatively easily. But we know that transaction costs used to be much higher than they are today, and that it was much more difficult for an investor to hold a a broad, diversified index. Nowadays, it is possible for investors to achieve NET returns that are within a whisker or two of the GROSS returns of a broad market index. But 40 years ago, there would have likely been a larger gap between the gross returns of the index and the net returns that even a skilled investor would likely achieve. So even if gross returns going forward are less than gross returns historically, it is quite possible that net returns going forward are in the same ballpark as (or maybe even better than) net returns achieved historically.

[bnwest]

Today S&P 500 is properly valued and has been for last 18 years.

See Fair Market Value

[Opponent Process]

I'd still like to see another 1000 shaved off the Dow to shake a few more people out. going under 10K again should provide enough capitulation. right now, too many people still think the stock market is a good investment.

[drjdpowell]

[gkaplan]

Wasn't the Dow in the 9,000s some time in 2002?
_________________
Gordon

[cjking]

[bolt]

[Opponent Process]

[avalpert]

[djw]

If the S&P 500 has been OVERvalued for the PAST 18 years, doesn't that imply that it could just as easily be UNDERvalued for the NEXT 18 years? How does q imply that the S&P 500 has nowhere to go but up from here? On the other hand, since we have no way of knowing whether it will go up or down from here, this analysis does seem to suggest that this is as good a time as any to put money into an S&P 500 Index Fund (although if it goes down another 10% from here that would be an even BETTER time...)

[mas]

Sorry, I have trouble giving this much credibility given the following quote from the book in March 2000, and listed prominently on their web site.:

[cjking]

The very phrase overvalued I used in my first post implies there are two measures of value, one given by the stock-market and another calculated some other way, which is the one I call fair value, but I'm happy to call it something else here.

I assume the majority in these forums will reject the idea that there can be any other meaningful measure of value other than current prices. I accept there's probably not much point arguing about it. I used to hold similar views. Even though I found the explanation of how the 1987 crash could be reconciled with efficient markets theory difficult to swallow, when I re-read "A Random Walk down Wall Street" shortly after it occurred, it wasn't until I read "Valuing Wall Street" in early 2000 that I felt I had intellectual justification for leaving the fold.

With regard to predicting the stock-market, the authors are economists and their official considered position is that they do not believe they can predict what will happen in the short term and medium term. I realise this is at odds with the jacket quote, I guess the publishers marketing department got them to stretch their warning that markets were overvalued into an uncharacteristically specific claim. Their actual claim is that "q" (and 10-year CAPE) predict "hindsight value", the average annual return experienced by a group 40 of investors with holding periods of 1 to 40 years from the prediction date.

There are some very impressive charts in their book showing the high degree of correlation between prediction and subsequent reality. (Most of the time the lines overlay each other, though the fact that they are 100 year charts may make them look more impressive than they would if you zoomed into shorter periods.) I've downloaded Prof. Schillers data from his web site and combined it with their "q" data, and managed to reproduce their results.

They say they do not contradict efficient markets theory, because a prediction for a 30 year plus period does not constrain the stock market doing what it likes over shorter period, say five to fifteen years. But they did say the market was at an unprecedented level of overvaluation at the end of the 1990's, and (notwithstanding the details of that Dow prediction) as far as I'm concerned they were completely vindicated very shortly after publication. (Though the complete absence of a crash since publication, had things panned out that way, would not have been a disproof of their position.)

With regard to whether "fair value" is helpful, the use I make of it is to be 0% exposed to equities when I predict the future return is only going to be 2%, or even when I predict the future return is going to 6.7%, but I have an alternate asset class which I'm happy to be 100% invested in, and which I believe has a future real return of 8.7%. (Since you ask, British "REIT's", which have crashed mightily over the past year. I am a British investor.)

[docneil88]

From http://www.smithers.co.uk/faqs.php :

[Verde]

[baw703916]

The problem I have with that analysis is that applying that logic, an investor should not have invested in the S&P 500 in, say, 1992, because it was overvalued. But the last 16 years have given pretty good returns, overall (granted, all of the returns happened in the first 8 years).

So is the take-home point that one should only invest when the market is not overvalued according to "q" or whatever metric one uses? Or, that in spite of the current gloomy mood among some prognosticators, this is a good time to buy stocks if you have the money and it fits in with your financial picture?

The first possibility I disgree with, the second one I strongly support.

Best wishes,
Brad
_________________
I don't foresee any Black Swans appearing in the future

[ken250]

cj, you're not alone...I use actively managed value funds.

[spam]

[baw703916]

There are various international real estate ETFs:

RWX (SSGA)
WPS (ishares)
DRW (Wisdomtree)

None are specifically focused on Britain, but all have the UK as a large percentage of holdings (along with Australia, Hong Kong, and Japan)
_________________
I don't foresee any Black Swans appearing in the future

[Russell]

[cjking]

[cjking]

OK, I've had time now to look in detail at how the SP500 has performed during the period of overvaluation. (Prof Schillers data is still 6 months behind, so I had to fill in my spreadsheet with provisional numbers.) The first number to pop out nearly made me laugh out loud - it's going to please the buy-and-holders.

If you had sold at the end of February 1991 when q went up to 1.03 (market 3% "overvalued") and stayed out until the end of June 2008, to be better off you would have had to find real returns elsewhere to better

... wait for it ....

6.7% per year!

I suppose I shouldn't have been surprised. In a 17 year period in which the market was at fair value at both the beginning and the end, it shouldn't matter what happened in-between, so it shouldn't be surprising that returns are exactly the same as the very long term average for the stock market.

In other news...

If you'd got out at the end of February 1993 when the 50% overvaluation threshold was first crossed, you would have had to find real returns of 6.2% per year elsewhere.

If you had got out at the end of October 1996 when the 100% overvaluation threshold was crossed, you would have needed to find real returns elsewhere of 4.3% per year.

Had you for whatever reason quit shares at the end of August 2000, when overvaluation peaked at 294% (shares nearly triple fair value) you would have avoided returns of -2.8% per year over the next 8 years.

If the threshold for selling in the "Valuing Wall Street" suggested strategy were raised to 100%, the revised strategy would still have got you out before the 2000 crash and the 1929 one, but you would have remained fully invested through the early seventies crash. (The 50% threshold would have bailed you out of that one as well.)

I guess the lesson I take from this is that it vindicates the new strategy I adopted at the beginning of this year, in which I switch asset classes not on the basis of the overvaluation of shares considered in isolation, but on the basis of what the level of overvaluation leads me to predict about share returns relative to what I expect from other assets.

[dumbmoney]

[yobria]

[dumbmoney]

[yobria]

[dumbmoney]

[cjking]

[baw703916]

[sopogah]

Humans talk logically but act emotionally. That applies to the matkets too.

S&P500 fair value will be decided after the panic is over. (wait a minute; me panic, nooo ).

May be next year this time, we will have a better picture. Unil then I am not planning to increase or decrease my 15% stock allocation.

Good luck to everybody.
_________________
_________________________________

"Money is a good servant, but a bad master"

"The fewer our wants, the nearer we resemble the Gods". Socrates

[UKbloke]

[donocash]

"Valuing Wall Street" is essentially a book explaining and justifying "q", it's not a part of its core mission to be an investment guide. Having said that, they do devote a chapter to what you might do to implement what you've learned, and one strategy they suggest is selling when the stock-market becomes 50% overvalued then staying out until it hits fair value again. The 50% level was fairly arbitrary, they examined other levels and found it didn't make that much difference what you chose."



Eyeballing your chart, this advice doesn't makes sense to me. Periods of pronounced overvaluation (1905, 1929, 1965, 2000) are all followed by plunges into pronounced undervaluation territory. In no case did the q drift lazily from pronounced overvaluation back to fair value. The q always plunged sharply into deep undervaluation. Wouldn't you want to wait for the q to indicate deep undervaluation for this to work as a timing device?

[grumel]

So you know the replacement costs for the s&p 500. That makes you pretty unique.

[cjking]

[docneil88]

From an amazon.com review of Valuing Wall Street:

[cjking]

[cjking]

[kenschmidt]

Well I for one have enjoyed your posts on this thread. I think you bring up some interesting points. Of course, as always, the difficulty becomes for me at least - "is the information actionable"?

I think it is to a limited extent. You have developed a mechanical strategy for utilizing Tobias' q ratio that works for you. Stepping back a little, I think it is apparent that at least some here have made similar decisions to that which your mechanical system points - i.e., as expected returns of an asset class go down, you start looking around for alternatives.

The "overvalued" calls for US Equities started in 1996. By 2000, they were very strong. Anyone posting on the old Morningstar board back then would agree I think. Many posters, myself included, moved into additional asset classes under the premise of diversification - but, thinking about it, what was the dynamic? US large cap equities likely had substandard/poor future returns while other asset classes such as REIT, small value, international, emerging markets, etc., had better expected returns. I moved into a lot of these asset classes in 2000 and don't regret it. I didn't abandon US equities, but adjustments were made. The last 8 years for me have been pretty good.

What I don't buy into is the concept of this - "when the q ratio shows the market is overvalued, I get out". I think that's a bad idea. Markets are too unpredictable for those kinds of bets. What I do think has merit is this - when market fundamentals indicate that perhaps future returns are poor, you consider your alternatives and plan and adjust accordingly.

An analogy that comes to mind - I recently read a book about card counting and black jack. When the ratio of face cards to 2-9 goes above a certain level, the player has the advantage. It is a slight advantage, so the bet should be increased - but not too much because the actual results are unpredictable and a you don't want a string of unexpected (bad) hands to wipe you out completely before the statistical advantage comes to fruition.

Using the q ratio, the "card count" has just gone to positive. You could still lose your shirt, but things are looking better than they have in awhile from a fundamentals perspective.

Best regards,
Ken

[tower]

July 7, 2008
CJ King. Nice post.
Russell. Your cite made me want to update the Harney-Tower study in light of the new data from Smithers & Co. Andrew Smithers kindly emailed me the data behind the chart.
PSteinx. Here is an alternative to calculating fair value, although it does not deal with most of your objections.

When I use all of the Smithers-Wright data from 1900 on to estimate the seven year real rate of return as a function of ln(q) I get as of today (July 7, 2008) the expected seven year real rate of return from investing in the S&P500 is 5.8%/year over the next seven years, where all rates of return are real continuously compounded geometric rates.

I also estimated an alternative, explaining the seven year real rate of return as a function of ln(q) and a distributed lag of past real returns, where the lag goes up to 12 years. And I assume that past high returns make the investors more optimistic, so the impact of a high past return is positive. Since the return so far this year has been highly negative I would not be surprised if the expected return is low. That is in fact the case. It is 2.975%/year over the next seven years.

I used Microsoft Excel Solver, which enables me to run a regression and constrain coefficients to be positive.

Thus if you believe in momentum (which has its foundation in behavioral finance-and does not come out of a maximizing framework), the returns look less good than if you do not believe in momentum.

What happens when I predict one year real returns? When I redo the analysis, with no momentum, the expected return is 6.07%. When I add momentum, it falls to 0.93% over the next year.

Thus, the return we expect depends strongly on whether investors are spooked by the recent drop in the stock market. Such spooking has happened in the past.

A caveat: I have checked these figures, but not double checked them. I hope to find time to write something up formally over the summer.

Ed

[3Factor]

[tower]

The return to the S&P500 in 1987 was +2.03% according to Yahoo. So the crash was quickly reversed.

[cjking]

[stratton]

[cjking]

I'm reviving this thread just to officially record my view that, with hindsight, it was premature.

A couple of weeks ago, Smithers and Co. published a revised version of "q", which differed from the original version by something like 35% in its assessment of market overvaluation. Apparently the problem was something to do with mark-to-market accounting. The good news is that the revised version agrees with CAPE, so I no longer have to worry about the discrepancy.

According to my latest caculation based on the revised version, fair value for the S&P 500 is currently 880, so today is the day my original post above should have been made.

http://www.smithers.co.uk/page.php?id=34