Stock Market Expectations
Which view of the stock market to you take?
On the one hand there is the fundamental interest rate, inflation approach. It says that the market overreacted to the two shocks of Greece and Deepwater Horizon and is now cheap.
The S&P 500 PE on trailing operating earnings is now 16.5, and according to the traditional relationship with interest rates and inflation that is well below fair value and thus creates a buying opportunity, at least in the short run.
Alternatively, you can take the money supply approach that says the Greece and Deepwater Horizon shocks were just the first shocks and that because of weak money supply growth more shocks of this nature should be expected.
Note, that the first drop in the PE early this year was simply the massive write-offs of 2009 rolling out of the trailing earnings data. Consequently, that PE drop was expected and was not a problem. But now with the spread between unit labor cost and prices at near record levels strong earnings growth should continue despite the point that a strong dollar is restrictive on the economy and will will dampen reported foreign earnings — up to half of the
S&P 500 operating earnings.
You pays your money and make your bets.
What do you think?
In the comments there was a discussion of the frequency of the market PE, and my point that there is no central tendency for the PE to converge on the average PE of about 14– 14.5 to be exact.
Here is the market PE since 1871.
Here is a histogram or frequency distribution of the market PE since 1871.
If you combine the individual distribution into ranges of about 2 PE points — below 12,
12 -13, 14-15, 15-16, etc., etc., this is the distribution you get.
As I said in the comments if you pick any random time the PE is just as likely to be at any point between 10 and 20 as at the average of some 14 to 15.
Maybe in advance stat course you could pick this apart, but for all practical purposes and for general analytical purposes this clearly appears to support my point. Maybe if commentators actually ever looked at the data before they told me I was wrong I could accept their criticism or disagreement. But for someone to make up numbers out of thin air and proceed to tell me
I am wrong is not acceptable.
I think there are more than two views and, for each view, there are many approaches for justifying it.
Personally, I view the stock market as a highly manipulated sham, and I plan to bet against it when I believe that reality cannot help but overcome the manipulation. I’m not there yet, but I’m getting close.
I don’t believe in either view. The market does not react to any measure of value, news, shocks, or any exogenous factor, nor is it rational. Check data on market movement and earnings. Movements always lead, so earnings can only be a driver if they operate through a time warp.
By the way, if you take a longer view than back to 1960, 16.5 P/E is in no way undervalued. It’s at the high end of the long avg band- something like 12 to 16. Bottom is a long way down. Further, trailing earnings are not a fair judge of value. You need future earnings, which are unknowable. Now, P/E is in the midst of a long process of unwinding from a historic high to what will be a multi-decade low before the next great buying opportunity. After 1960, it took 20 years. From an even higher high, it’s not hard to believe the bottom is still a decade or more off.
BTW, with lax accounting rules that were only honored in the breach, earnings were boldly overstated in 2000, so the real P/E was far higher than indicated. Are current earnings reports credible? I dunno.
And where are future earnings going to come from? Productivity growth in recent years has been unsustainably high, so that well is tapped out. The job situation is dismal and not getting better very quickly. Deficit hawks are killing any recovery opportunities. Real estate and construction won’t recover in a big way for years.
The market is a barometer of social mood, and the mood is contractive, negative, and war-like; not expansive and optimistic.
I’ll get interested when the P/E drops below 7, and then shows signs of turning.
“In former times analysts and investors paid considerable attention to the average earnings over a fairly long period in the past – usually from seven to ten years. This average figure was useful for ironing out the frequent ups and downs of the business cycle, and it was thought to give a better idea of the company’s earning power than the results of the latest year alone.” Benjamin Graham
I think that Graham would say that you should be looking at at least the past three years earnings not just the past year. I couldn’t track down a recent 3 year PE for the S&P, but eyeballing it, I’m guessing around 18-20. I think that Graham would say, mildy speculative. There may well be individual stocks or even sectors that have attractive PE ratios. But the market as a whole? Probably not the best idea.
I also think that since Graham’s day, earnings estimates have probably become on average less honest. I suspect that for a variety of reasons, earnings most quarters are modestly overstated and that corrections are made via slightly bumping the massive writeoffs during financial crises — if true, that’s another argument for using a longer time span that includes writeoff periods to compute PE.
Earnings estimates are indeed “manipulated” mostly by analysts. They are usually understated, so when a company reports things look better than “expected”, then they can win investment banking business.
It is very important not to look at PE in a vacuum. We are coming off a big recession so PE will be inflated becuase earnings are down, and many companies have/had no reported profits. If earnings do revert to a mean pre-recession then teh PE probably comes down. I think what is spooking teh market right now, is that the economy starts to back off recovery mode, and earnings go down. There are some attractive valuations on a company-by-company basis.
Given that Greece is a part of the problems cited, perhaps we should go to Delphi and restart the oracle there. Have to find a pretty young girl, sit her over the crack where the vapors rise and then ask her questions. For predicting the future this is about as good as many of the other predictions made. Or recall J.P. Morgan whos prediction on the market is that it will fluctuate. For a lower energy method get some loose tea and read the leaves, the web has stories on how to do this.
It seems to me that the peek of the growth chart did not correspond to a particularly large post recession growth rate. The reduction and the trend of the indicator has got to be troubling.
buy fire pit
It’s also of interest to look at Shiller’s 10 year trailing PE ratios. Go to http://www.econ.yale.edu/~shiller/data.htm and click on “long term stock, bond, interest rate, and consumption data”. Depending on your browser and the state of your karma, that’ll get you a spreadsheet. Scroll over to the far right columns and find the 10 year PE ratios, then scroll up and observe that the PE ratios for the last decade are wildly out of line with the older data. The last entry is 2009 which has a very low S&P500 price and thus a low PE …. compared to the immediately preceding years. But still kind of high when compared to older data.
Anyway, it’s all kind of frightening if you believe that the PEs will eventually revert to long term averages.
I agree with VtCodger and JzB. Long term the PE of every stock market mean reverts, and since the NYSE has spent so long at very high valuations, we can expect to spend a long time at lower valuations before the next secular bull market. When that may occur is anyone’s guess, but Vitaly Katznelson has a pretty convincing argument that we may not be back in a secular bull until 2020 or so. Until then we will be in a “range-bound” market, where mini bulls and mini bears will gradually wear down the PE to a much lower valuation level, without any sustained upward trend in the DOW.
Active Value Investing is a great book, BTW!
I think a PE of 16x for S&P is way overvalued given all the fear in the market and the uncertainty regarding earnings. This is a way abnormal earning environment and we are mid credit crisis. Only when central gov pull short term liquidity and raise interest rates will anyone get a glimpse of where investor sentiment really is. Moreover, I agree with Jazzbumpa that the PE ratio range is 12-16. Based on all the forgoing, S&P is fairly valued when trading at 8-10 X 2009 earnings.
There is no central tendency for the stock market PE to revert to its long term average. If you do a histogram of the market PE — using Shiller’s data — since the early 1880s you will find that the number of observations in each data range –below 10,10-12, 12-14, 14-16, 16-18, 18-20, and above 20– is very similar.
Hey Spencer, I’m a lousy mathematician but I’m somewhat skeptical that a histogram is a valid tool for determining if regression exists (Only Gaussian distributions have means to regress to? I thought they told me something different). On the other hand what do I know?
But I am a good enough mathematician to be pretty sure that if you are going to do histograms you probably need counting intervals that are the same size which certainly rules out “above 20” and doesn’t look quite right for the other intervals either
FWIW, here are the histogram counts for what I (perhaps naively) feel to be equal sized intervals defined by integer(10*log(PE))
5=0, 6=0, 7=3, 8=3, 9=12, 10=22, 11=22, 12=25, 13=34 14=18, 15=8, 16=2, 17=2, 18=0, 19=0
That’s sort of bell shaped. The mean looks to be around 18.0 give or take a few. It’d take me longer than I want to spend to compute the mean properly working from the logarithm of the mean. Like I said, I’m a lousy mathematician. …. And there appears to be a good deal of autocorrelation between values. If I ever knew how to approach that, I forgot right after my statistics final in 1958.
I never though a histogram had anything to do with regressions.
I always sort of went along with what I found in wiipedia:
In statistics, a histogram is a graphical display of tabular frequencies, shown as adjacent rectangles. Each rectangle is erected over an interval, with an area equal to the frequency of the interval. The height of a rectangle is also equal to the frequency density of the interval, i.e. the frequency divided by the width of the interval. The total area of the histogram is equal to the number of data. A histogram may also be based on the relative frequencies instead. It then shows what proportion of cases fall into each of several categories (a form of data binning), and the total area then equals
In other words if I had 100 observations and 20 were between 10 & 12, 20 between 12 & 14, 20 between 14 & 16, etc,, I would have a series of rectangles of equal height and wight that would send a message that there is no central tendency for the observations to tend towards any single observation.
I’m afraid I do not understand what point you are trying to make.
My point is that if you pick any random time the market PE is just as likely to be about any observation between 10 and 20 as the average value of around 15.
what is your point?
****My point is that if you pick any random time the market PE is just as likely to be about any observation between 10 and 20 as the average value of around 15.
what is your point?***
That what you assert is not true because your interval sizes are inappropriate and do not have meaningful sizes. And that when the intervalsy are made roughly equal, the probabilities of a value falling in a given interval are not equally distributed. Surely, you are not asserting that the interval “above 20” is in any way comperable in size to the interval 10-12?
If you wish to continue to believe that you analysis is correct, feel free …. But it probably isn’t. Whether mine is correct is certainly open to question.
VTCodger — see the charts and the comments I added to the original post.