My Photo
Note: Jeff does not accept guest blog posts on A Dash of Insight.

For inquiries regarding advertising and republication, contact

Follow Jeff on Twitter!

Enter your email address:

Delivered by FeedBurner


  • Seeking Alpha
    Seeking Alpha Certified
  • AllTopSites
    Alltop, all the top stories
  • iStockAnalyst
Talk Markets
Forexpros Contributor
Copyright 2005-2014
All Rights Reserved

« Understanding Economic Progress | Main | Talking Stocks, Blogging and System in an Interview with Tadas Viskanta »

May 17, 2011



Inkerton --

You are completely correct. I have a price target in mind for each stock. If it hits the target, or even closely approaches, the risk reward changes and I sell. I am not making a ten-year forecast about any stock or about the market as a whole. None of my programs are buy and hold forever.

I have made many friends from the blog with contacts that involved email, calls, and in-person meetings. I never reveal anyone's name or what they say in email without permission, and any conversation is by mutual agreement. I don't mind anyone choosing to be anonymous in the comments.

With this in mind, it does not seem right for a commenter to use multiple names.

Thanks for helping me to clarify these points:)



I'm not meeting your challenge, or attempting to, but I would argue there is a lot of middle ground between Hussman and you. For example, I have come to the conclusion that you are generally right about the status of the market, and I like your measure of risk, using the St. Louise Fed Stress Index. I also agree completely about the minimal impact of QE2.

However, I also think, for example, that from today's price level, CAT is absolutely terrible right now as a long-term investment (where long-term is defined as a holding period of more than five years, likely much longer). Sure, CAT is firing on all cylinders now, but if you look at data, as you do, there will likely be another recession with five to seven years, and when that happens, highly cyclical CAT will trade vastly lower than it trades today. I am happy to give you my phone number so you can call me and mock me if CAT does not trade much lower than today's price within the next seven years, but I truly think it is a phone call you will never be able to make.

So, when Hussman is talking about "expected ten-year returns," that's the kind of thing I take him to be talking about, in part. Now, one of his problems is he applies that model to the entire market, even apparently to far less cyclical companies within it. He also is obsessed with and draws incorrect conclusions from Fed actions, because of his political ideology, and even gets into dubious discussions about the legality of Fed actions (another thing I like about you is you really do try to exclude politics).

So in short, I love this site, but I just wish you would make even more clear that when you are saying CAT is great, for example, you are only making short-term calls (one-year or less, maybe two years at most), and that nothing you are saying about CAT is meant to indicate that, from today's price, it is going to outperform the market over the next decade. I don't think you are saying that, and I don't think it would be true if you were.

Also, side note, I'm not posting my real name, as I would rarely post my real name on any website. It seems unfair for you to draw any conclusions when people do not post their real names publicly, particularly if (as in my case) it is readily ascertainable from the email address they provide you when posting.

Mike C


Almost kind of silly for me to point this out....but you are the one who brought it up. If you look off to the side with recent comments, few if any people use their real names in commenting. Somehow, I doubt there is a guy out there with the legal name Proteus. OK, you "got me". I posted a few comments under a nom de plume. So what.

So first let us note that you did not meet the challenge, OK?


You agree that the method I have cited is best for one-year forward earnings.

Not necessarily. It is a simple fact of the matter (not opinion, FACT) that analyst estimates are horribly off near earnings peaks and troughs. Whether there is a "better" method for forward estimates, I do not know. I haven't spent much time on it. Maybe it is the best we've got in the absence of a crystal ball. When I was working as a buy-side analyst, there were a few instances where my personal estimates were closer to the mark then the consensus sell-side number, but that was only in cases where I had spent a tremendous amount of time on that individual company.

Is this important? Everyone in the business looks at year-ahead earnings.

See, this is the sort of assertion that is demonstrably false. Jeremy Grantham probably manages more money then 10x you and I put together, and it is clear from his writing that he doesn't pay attention to year-ahead earnings forecasts in his investment decision making. Look, reasonable people can disagree about these things, but it is a mischaracterization to paint things with an "everyone in this business" does this or that. That is just a rhetorical trick to try and make one position seem like gospel.

I am happy to discuss what I actually write about, but I am not going to join you in veering off topic in the comments.

This seems to be your standard tactic with any comment I make now. Accuse me of "veering off topic" and then simply ignore the questions/issues I raise. You mentioned Hussman & Shiller in your note and then "invited comment". I respond on that point, and you accuse me of going off-topic. I don't really care anymore about the Hussman and Shiller valuation approaches. You have addressed them, and different professional investors will come to different conclusions about their validity has valuation tools.

I have NO interest whatsoever in discussing Hussman or Shiller valuation methods. I was simply responding to YOUR INVITATION to point out they don't forecast one-year earnings so it is a complete non-sequitur to "issue a challenge" about their "methods" for one-year earnings forecasts.

Everyone is trying to build a better mousetrap. I respect that. Me too. Hussman has tweaked his approach. David Merkel has done what I think might be some astounding work that he has shared (he is a model for combining detailed data analysis with some humility about conclusions).

In terms of QE2 versus fundamentals driving the 9-10 stock rally from the summer, the answer would be it is a complicated question. Cullen Roche over at Pragmatic Capitalist thinks much of the rally has been QE driven, so does Jeremy Grantham, so does Richard Koo. There is a deep bench that thinks it has been a factor. Some people think it has been a non-factor such as yourself or Bob Doll. I really don't know. I'm reminded of the Bertrand Russell quote about who is certain and who has doubts.

For the sake of the current discussion, the blend worked well during the period discussed--- real time performance -- not some "bad times to invest" back test.

Are you saying that a 1-year performance of a particular analytical model has any validity as to whether it would work over say a 10-30 year going forward basis?


Hi Jeff,

Interesting blog, I think few people realize how many leaps of faith they routinely make in the realm of cause and effect.

One comment I would have in the fed policy = improved fundamentals vs fed policy = stock gains debate is that what we don;t know is what affect fed policy has had on fundamentals. I'm sure one could make a case that they have a link, the question is how strong and what might change.


Mike C --

It is nice to see you using your own name again instead of a fake one. This helps readers put your comments in the context of what you have written before.

So first let us note that you did not meet the challenge, OK? You agree that the method I have cited is best for one-year forward earnings.

Is this important? Everyone in the business looks at year-ahead earnings. In fact, they usually look at shorter periods. You are free to consider alternative approaches, but it is certainly not a red herring.

I am happy to discuss what I actually write about, but I am not going to join you in veering off topic in the comments. I have written many articles about the complete Hussman and Shiller approaches, easily found through the search box. These include discussion of methodological errors like using too much irrelevant data (the Eisenhower and Taft eras) to create the illusion of science. There were no forward earnings estimates in 1950.

To emphasize, this is not an article about valuation alone. I have carefully included valuation, risk, and economic potential. All three are important. As it happens, I am working on blending these together in a longer history. I'll have more later this summer.

For the sake of the current discussion, the blend worked well during the period discussed--- real time performance -- not some "bad times to invest" back test.


Mike C

"I continue to invite any loyalist for another method -- particularly those espoused by Hussman or Shiller -- to show that they can predict earnings one year in advance with greater precision."

OK....since you invited

You've repeatedly stated this. As far as I know, neither Hussman, Shiller, or any "loyalist" has made claims about predicting earnings one year in advance so this is red herring. I think what Hussman refers to is whether specific metrics robustly predict forward returns. The question/issue isn't whether 1-year forward estimates are generally accurate. The question is can they be used in a long data series over a very long period of time to estimate 1-year ahead returns.

As Hussman states in this week's commentary:

"Analysts who claim that stocks are "cheap" when a market advance is mature and profit margins are elevated should be expected to demonstrate that their approach is historically reliable. Specifically, there should be long-term evidence (since at least the 1950's, to capture a substantial amount of variation) that their methodology provides highly predictive information about the subsequent performance of the stock market.

OK, so 1-year forward estimates are generally accurate. For sake of argument let's grant that (although they are generally off the mark at earnings peaks and troughs). Can you take those 1-year estimates, plug them into some valuation or return forecasting model, and get a highly accurate predictive model?

The comments to this entry are closed.