Morningstar Expected Future Returns Article

[Animorph]

Some great charts in here for stocks and bonds:

A Noble Lie

"Finally, it's because the level of predictability is fairly low for the stock market. The cyclically adjusted price/earnings ratio, or Shiller P/E, one of the best valuation signals, will give you only a ballpark range of returns over the next five to 10 years. Exhibit 1 shows a scatter plot of 10-year forward real returns for the S&P 500 versus its starting cyclically adjusted earnings yield (the inverse of Shiller P/E), using data from 1926 to 2013. As you can see, the relationship is very noisy. The signal becomes more useful at the extremes, suggesting you should aggressively alter your stock allocation only on rare occasions.
"

[CaliforniaMan]

Interesting chart but for the most part I don't see how this helps in actual practice. With so much variation, and so few points at the extremes vs. the mass in the middle, it would be better to show this in quadrants rather than as a regression. Buy during big panics is one message you can take from this. Other than that, even during periods of time where the earnings yield is around where it is today, there are still a lot more up periods than down. It seems to me if you want to do allocation adjustment 'timing?' you would need to take into account this ratio along with bond yields and inflation at that time. And there is precious little data to do this kind of analysis, a few times of really high inflation, and a few times of really low bond yields. It still seems to me there is not as much predictability as we would like to feel there is. In general though, I don't like plots like this, they can be misleading because it gives a sense of predictability that is probably not there. Often outliers are just that, very special cases. Maybe the next time we see P/E 5.0 Rome really will be sacked. All this makes my brain tired, maybe it is just better to ignore, go outside, enjoy life and take it as it comes....

[NW-Bound]

The linked article opens with "Why it's OK to sin a little and "market-time"--provided you do it in a contrarian fashion." Heh heh heh...

Having been through a few market crashes together already, we will all have to agree that this is really tough to do. If people around us are running scared, we have to wonder if they know something we don't. And occasionally, that is true as shown by history. Most Jews did not know what was happening to them in WWII. Taleb in his Black Swan book describes how many Lebanese did not know how their country which was known as the "Switzerland of the East" - Beirut itself was called "Paris of the Middle East" -- could descend into sectarian violence as we know today.

Anyway, back on the stock market and its P/E, I wonder what the chart in the OP would look like if we exclude the period of 1980-present. The reason is that the market experienced a P/E expansion during the recent 30-year period. Is that the "new normal", or are we going to revert to single-digit P/E's of earlier years? My guess is that we will not, but then what do I know?

[Animorph]

Quote:

Originally Posted by CaliforniaMan Interesting chart but for the most part I don't see how this helps in actual practice. With so much variation, and so few points at the extremes vs. the mass in the middle, it would be better to show this in quadrants rather than as a regression. Buy during big panics is one message you can take from this. Other than that, even during periods of time where the earnings yield is around where it is today, there are still a lot more up periods than down. It seems to me if you want to do allocation adjustment 'timing?' you would need to take into account this ratio along with bond yields and inflation at that time. And there is precious little data to do this kind of analysis, a few times of really high inflation, and a few times of really low bond yields. It still seems to me there is not as much predictability as we would like to feel there is. In general though, I don't like plots like this, they can be misleading because it gives a sense of predictability that is probably not there. Often outliers are just that, very special cases. Maybe the next time we see P/E 5.0 Rome really will be sacked. All this makes my brain tired, maybe it is just better to ignore, go outside, enjoy life and take it as it comes....

My take away was kind of the same. To some extent this is a refutation of the expectation of low future returns. Sure, we're on the crummier part of the chart, near 4% or so I think, but it's not an investment returns death sentence. There is still a wide range of 10-year returns that have started at this Shiller P/E.

The article also shows there is better reason to assume Treasury returns will be subdued from this starting rate. But just to the extent of current low rates, not significantly negative returns. Doesn't seem like cash would be any better.

[CaliforniaMan]

Quote:

Originally Posted by NW-Bound Anyway, back on the stock market and its P/E, I wonder what the chart in the OP would look like if we exclude the period of 1980-present. The reason is that the market experienced a P/E expansion during the recent 30-year period. Is that the "new normal", or are we going to revert to single-digit P/E's of earlier years? My guess is that we will not, but then what do I know?

You make a really good point, I too wonder about this P/E expansion trend. Are we fooling ourselves that things "are different this time". But then again, P/E ratios are higher during periods of low inflation, and low during high inflation. Treasury rates also have to play a part, as do the quality of earnings (based on real growth and real productivity, rather than rising prices), and also the investors 'expectations' of these.

Link regarding this here:
Why are P/E ratios generally higher during times of low inflation?

Seems like we have kind of a multivariate question here as... "Will the REAL P/E ratio please stand up". What would happen to the Shiller P/E Ratio taking these into account? I mean adjusting PE for inflation and treasury rates and then looking at the difference between this "expected" or "normalized" P/E vs next decade's return. Would it be better? worse?

Anybody have an idea if this kind of evaluation has been done before or if it even makes any sense?

[Lsbcal]

Some were warning about PE10 last year when it was around 21. So far 2013 returns look pretty decent. So now it's 25. Nothing has convinced me that it will predict 2014 returns with any confidence level.

The chart in the OP is about 10 year returns. Not next year or the year after necessarily. Yet too many bring up PE10 as if it's a real useful indicator for the current market conditions. It is not.

[NW-Bound]

Quote:

Originally Posted by CaliforniaMan You make a really good point, I too wonder about this P/E expansion trend. Are we fooling ourselves that things "are different this time". But then again, P/E ratios are higher during periods of low inflation, and low during high inflation. Treasury rates also have to play a part, as do the quality of earnings (based on real growth and real productivity, rather than rising prices), and also the investors 'expectations' of these. Link regarding this here: Why are P/E ratios generally higher during times of low inflation?

You would think that P/E ratio, or rather its inverse E/P, would track inflation rate, as that makes just too much sense. Alas, I remember reading that it was not sufficient to explain the P/E expansion in the 1980-2000 decades.

So, I looked at historical data again for myself and the following is what I found.

From 1948 to 1965, the US enjoyed a prosperous post-WWII period of low inflation. The cumulative inflation from Jan 1948 to Jan 1965 was a mere 31.65%, which works out to an annualized value of 1.6%/yr over this 17-year period. The stock market return over this period was 6.8X after inflation adjustment, which is 11.9% annualized real return. Just wonderful!

And what was the P/E during that period? See the linked chart below from www-dot-multpl-dot-com. It was low, way too low for the low inflation of 1.6% computed above.

[Lsbcal]

Quote:

Originally Posted by NW-Bound ... And what was the P/E during that period? See the linked chart below from www-dot-multpl-dot-com. It was low, way too low for the low inflation of 1.6% computed above. ...

Interesting thought. I think yields in the 1950's were high too. In 1950 the SP500 was something like 8.7% (if I'm interpreting right from the Shiller data).

People in 1950 had just been through the 1930's Depression Era and WW2. Can we blame them for not paying much for earnings? Once I did a calculation and found that about 25,000 people per day died in the WW2 years. People were truly gun shy.

[NW-Bound]

The 11.9% return I showed above was the total return, which included the high dividends.

Well, when I looked closer at the P/E chart, I saw that there was a P/E expansion then too. The P/E was as low as 6 in 1948, and increased steadily up to the 20-ish range in the early 60s. I guess it took several years before investors got comfortable with equities after WWII.

People started to pay higher for stocks in 1980-2000. Some pundits said that was because more middle-class people started to buy stocks directly for their investment account and their 401k, and the equity risk premium has been reduced.

So, is the above phenomenon going to be reversed? Mean reversion must assert itself for Shiller's PE10 thesis to work!

[galeno]

My favorite value indicators when we used individual stocks for our equity allocation are: PE, Price2Sales, and Price2Book.

The Schiller PE10 is a poor value indicator for selecting individual stocks. But it is accurate for predicting long term equity returns for broad based indices.

Using the Schiller PE 10, Bill Bernstein's latest predictions, and my own voodoo, our 60/40 indexed port's expected annual net real return = 2.3-3.3%.