Philosophical Economics: Introducing the Total Return EPS

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Philosophical Economics: Introducing the Total Return EPS

Post by siamond »

Another fascinating (and very timely given all the fuss on the latest expected returns forecast from Dr Bernstein) piece from my favorite writer. How to put some order in the ever-changing combo of dividends, earnings and stock-buybacks. A terrific read, a very insightful idea, and definitely something I'll play with in my own Excel models...

http://www.philosophicaleconomics.com/2015/03/treps
blueleaf
Posts: 134
Joined: Fri Dec 27, 2013 7:30 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by blueleaf »

Thanks for posting. I love reading this guy's work. I have no idea if his forecast is right, of course, but he's so thoughtful and crafty with the data.

It definitely makes sense that eps should grow faster if you allocate more earnings to buybacks and investment and less to dividends. Bernstein's estimates just assume a constant historical eps growth rate, which would seem to require that the increase in net buybacks have no impact at all on long term eps. Seems like a weird assumption.
Startled Cat
Posts: 709
Joined: Thu Apr 03, 2008 8:54 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by Startled Cat »

I saw this piece earlier today and I'm glad someone posted a link to it here for discussion. Philosophical Economics has some first rate content. One post in particular that comes to mind is a discussion of how changes in accounting standards affect historical comparisons of CAPE: http://www.philosophicaleconomics.com/2013/12/shiller/

There is also a related post titled "Why is CAPE so high?": http://www.philosophicaleconomics.com/2014/08/capehigh/
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by lack_ey »

Great article there.

Some of the adjustments there are arguably a little bit wishy washy, but I think overall it's a very interesting treatment, better than many others. It certainly deserves more views. I am eagerly awaiting further writing about the topics skipped as mentioned here:
It turns out that we can use Total Return EPS to do all sorts of interesting things: accurately predict future growth based on position relative to trend, decompose and visualize historical returns in terms of their contributing components, estimate profit margins during periods in the late 19th and early 20th century when the data necessary to calculate them was not available, construct new-and-improved Shiller CAPEs that allow for valid comparisons across history and across countries, and many more. But those would be too much to discuss in one piece, so I’m going to save them for later.
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

I reconstituted the math from the article, this is quite easy. Using the ex post knowledge of the mean historical CAPE in the share count math is dubious, but changing it to an average of previously known values leads to similar results, so ok, not such a big deal.

I have to say that I find hard to understand why the author uses a mean valuation mechanism in this 'share count' math though, reinvested dividends would use the current price, so why would this be different if dividends are eliminated in favor of a 100% stock buyback?

Can't wait for the next piece...

(edited: removed erroneous statement; Excel error on my part!)
Last edited by siamond on Fri Mar 20, 2015 6:02 pm, edited 1 time in total.
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by SimpleGift »

This is another creative and insightful article by Mr. Livermore — but it troubles me that he assumes mean reversion of BOTH earnings growth and CAPE 10 back their historical averages. There's a very good case for mean reversion of corporate earnings, I believe, back toward a long-term trend growth rate of about 5.7% per year (chart below). To my mind, this follows naturally from the continued dynamism of capitalism, plus the normal business cycle, that we've seen for over a century — at least in the United States.

Image
Source: Philosophical Economics

But Mr. Livermore loses me when also assumes mean reversion of CAPE — as if this were a natural occurrence and there is some "normal" level of price/earnings at which stocks are destined to trade. I realize that in this current piece he is critiquing a market prediction made by Robert Shiller, who does believe in mean reversion of CAPE. But, as Mr. Livermore himself pointed out in a previous piece, CAPE has historically only fallen below average due to 1) war, 2) high inflation, or 3) a financial crisis (chart below).

Today, we have the opposite of these three fearsome factors and we are instead experiencing relative peace, low inflation and financial stability worldwide. Certainly an unprecedented catastrophe could come along (a global epidemic or environmental collapse) — but absent these, I don't see a good reason why CAPE must mean revert back to its historical average anytime soon.

Image
Source: Philosophical Economics
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

Simplegift wrote:Today, we have the opposite of these three fearsome factors and we are instead experiencing relative peace, low inflation and financial stability worldwide. Certainly an unprecedented catastrophe could come along (a global epidemic or environmental collapse) — but absent these, I don't see a good reason why CAPE must mean revert back to its historical average anytime soon.
I agree that this specific issue is really unclear. On one hand, one has to acknowledge that the CAPE mean varies quite significantly, as any rolling-average chart would easily show, and I am not quite sure that there is a SINGLE historical-mean to revert to. Without even speaking of changes in accounting rules. On the other hand, force is to acknowledge that 1/CAPE (or PE30 or whatever) is a pretty decent predictor of expected returns. And well, historical returns did seem to return to the mean, although I'm not quite sure this will stay true.

Anyhoo, as I explained in my previous post, Mr Livermore's use of the mean CAPE in his math is really NOT core & central to the general idea of the Total-EPS, which I personally find simple & beautiful. Eliminating the impact of the vagaries of the dividends payout ratio in all sorts of interesting math (expected returns and more) seems really welcome.
lack_ey
Posts: 6701
Joined: Wed Nov 19, 2014 10:55 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by lack_ey »

How good has CAPE been in determining future returns in other countries? Has there been mean reversion abroad? Several countries have good data at least 50 years, I think.


In any case, using mean CAPE as it is in the math produces a nice result. Maybe it would be fairer to use a rolling average instead, but it's hard to say and if it doesn't make that huge of a difference then using the mean I suppose may be less controversial to the Shiller bear types.
garlandwhizzer
Posts: 3565
Joined: Fri Aug 06, 2010 3:42 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by garlandwhizzer »

Simplegift wrote:

But Mr. Livermore loses me when also assumes mean reversion of CAPE — as if this were a natural occurrence and there is some "normal" level of price/earnings at which stocks are destined to trade. I realize that in this current piece he is critiquing a market prediction made by Robert Shiller, who does believe in mean reversion of CAPE. But, as Mr. Livermore himself pointed out in a previous piece, CAPE has historically only fallen below average due to 1) war, 2) high inflation, or 3) a financial crisis (chart below).

Today, we have the opposite of these three fearsome factors and we are instead experiencing relative peace, low inflation and financial stability worldwide. Certainly an unprecedented catastrophe could come along (a global epidemic or environmental collapse) — but absent these, I don't see a good reason why CAPE must mean revert back to its historical average anytime soon.
1+

Garland Whizzer
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

lack_ey wrote:How good has CAPE been in determining future returns in other countries? Has there been mean reversion abroad? Several countries have good data at least 50 years, I think. In any case, using mean CAPE as it is in the math produces a nice result. Maybe it would be fairer to use a rolling average instead, but it's hard to say and if it doesn't make that huge of a difference then using the mean I suppose may be less controversial to the Shiller bear types.
If you can find a public source with good PE data for 50 years or more, by all means, please share. From the little I've found, this may only exist for the UK, and not available for free, nor a modicum price. Mr Shiller did us a really big favor. Oh, and even 50 years is too short of a time period to show something truly meaningful about return-to-the-mean...

Yes, as I said in a previous post, I modified Mr Livermore's math to use a (long) rolling average and this didn't seem to create a significant impact. I still don't understand why the CAPE appears at all in this 'share count' math though. If I understand well, the author was trying to better extract the trend from (adjusted) earning's growth by avoiding valuation side-effects, but this doesn't seem quite right, this could introduce a skew on its own. Better use the proper numbers (i.e. use the real stock price as a divider, not a PE-adjusted price), and then use various mathematical tools to analyze.
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

Here is the first follow-up article from Mr Livermore. Clearly, there will be more. A terrific read!

http://www.philosophicaleconomics.com/2015/03/decomp/
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by SimpleGift »

siamond wrote:Here is the first follow-up article from Mr Livermore. Clearly, there will be more. A terrific read!
Yes, it's a "must read" for those interested in the long history of the U.S. market. Thanks for the link.

One interesting observation that jumps out from Mr. Livermore's analysis is how "solid" the recent stock market returns have been since the 2008 financial crisis (at the right, chart below). As opposed to the speculative market run-up of the 2000 tech boom, which was built primarily on multiple expansion, the market returns of the past 5 years have been based on solid growth in earnings-per-share — no speculative froth in sight.

Decomposition of S&P 500 Total Returns, 10-year Rolling Averages, 1871-2015
Image
Source: Philosophical Economics
rnitz
Posts: 282
Joined: Thu May 10, 2007 12:47 pm
Location: San Diego

Re: Philosophical Economics: Introducing the Total Return EP

Post by rnitz »

Anyone have any idea of who this guy really is? For the last year or so I've always assumed the name Jesse Livermore is a pseudonym (the great bear of wall street during the great depression). I don't want to out him, but I always guessed he was a finance professor at some university. A great read - I love his blog/site.
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

rnitz, no clue who he is, I've been wondering as well. A fascinating fellow for sure.

I ran some more numbers. I was hoping that the 1/PE10 (or PE30) metric, with the terms "P" and "E" redefined as the author of the article suggested (i.e. include the dividends components as 'virtual' stock buyback), would prove a better forecaster of expected returns than the regular 1/PE. Well, not quite. It's actually a bit inferior for the overall US history, and to add insult to injury, it is significant inferior for the past 50 years. Which baffles me a bit. Then I tried a DDM-derivative (div yield = 0; total-earnings-growth; modified CAPE math) and this proved very disappointing as well. Similar outcome when using the CAPE-adjusted logic or when using the actual market prices in the share-count math. Maybe the author will bring more clarity to the topic in a follow-up article.

I then reflected on the decomposition the author went through. Although this is terrific creative thinking, I still have a lot of troubles to accept this CAPE-adjusted math in the Total-Return EPS & Price indexes. I understand what he's trying to do, but this seems to introduce a very artificial skew in the numbers, and I have troubles believing the decomposition math due to that. Still... this author has a great knack for triggering new interesting lines of thought.
User avatar
Topic Author
siamond
Posts: 6010
Joined: Mon May 28, 2012 5:50 am

Re: Philosophical Economics: Introducing the Total Return EP

Post by siamond »

The third opus of this interesting experiment... A new & improved CAPE with dividends factored in, as expected. Although I confess a certain confusion after reading this piece!
http://www.philosophicaleconomics.com/2015/03/payout/
User avatar
SimpleGift
Posts: 4477
Joined: Tue Feb 08, 2011 2:45 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by SimpleGift »

In this most recent article, it's a bit difficult to distinguish what really matters to ordinary investors and what is just "inside baseball," with relevance only to true aficionados of the Shiller CAPE. But, personally, what I found most interesting were his insights on the historical growth trend of earnings-per-share:
Jesse Livermore wrote:Now, to be clear, the secular decline in the dividend payout ratio seen across the span of U.S. market history has not substantially affected the accuracy of the original Shiller CAPE.  However, it has substantially affected the trend growth rate of EPS.  So, though it may not be imperative that we use the Total Return version of the CAPE when measuring valuation, it is absolutely imperative that we use the Total Return version of EPS when analyzing earnings trends and projecting out future earnings growth.
...(SNIP)...
As the chart below shows, real EPS growth over the last two decades–on both a regular and a Total Return basis–has been meaningfully above the respective historical averages, driven by substantial expansion in profit margins.  Recall that high growth produces a high CAPE, all else equal.
What's been responsible for the striking increase in trend real earnings growth since about 1992? Globalization? The digital revolution? Low interest rates? The decline of real wages? All of the these factors?
Last edited by SimpleGift on Mon Mar 23, 2015 8:58 am, edited 1 time in total.
Startled Cat
Posts: 709
Joined: Thu Apr 03, 2008 8:54 pm

Re: Philosophical Economics: Introducing the Total Return EP

Post by Startled Cat »

Enjoyed the most recent piece. The big takeaway for me was the discussion of how modeling reinvestment of dividends at actual market prices would make the valuation metric dependent on the prior pricing of the market. This makes sense for a total return series, but it would be troubling for a valuation metric to depend on past valuations. Now the concept of reinvesting dividends at a "fair price" finally makes sense, though it's still an imperfect approximation.

It was also interesting that the metric uses Shiller CAPE as an input to determine the "fair price". I wonder if it would be practical to define the Total Return EPS CAPE recursively rather than building it on top of Shiller CAPE as an input. Starting with Shiller CAPE and iterating with successive approximations of the Total Return EPS CAPE would probably accomplish that, but I wonder if there's a more elegant way. As the author points out, it doesn't make any significant difference, but I just found myself curious about this aspect.
Post Reply