Expensive Misconceptions

The investment world abounds with research reports.  Intelligent and educated people generally benefit from careful study and accrued knowledge.

While it seems unfair, the investment world is different.  A little knowledge can be a dangerous thing!

There are many examples of this.  I have been stalled on this important topic because I was trying to do a comprehensive analysis.  It is often better to just get started!  I will start  with some of the most egregious and costly temptations for consumers of financial information.  I welcome more nominations to the list.  This is a great topic for us all to share ideas.

Shifting Indicators

This happens when the “rules” for interpreting data change to fit the per-conceived conclusion.  One recent example by bears related to the divergence of small cap stocks.  When the Russell 2K stocks were leading, the market was “frothy.”  When they lagged, it was a warning divergence.

Other indicators like sentiment, the Baltic Dry Index, Hindenburg omens, etc. are cited only when they fit.

Bullish analysts do the same.  If the monthly report does not fit the story, just look at non-seasonally adjusted data. year-over-year, or something else.  Many reports are susceptible to various spins.  The only solution for this is to know the agenda of the source.  This is rarely provided.

A persuasive chart

Please consider this chart, which is offered weekly as evidence that long-term investors have little to gain in the next decade while facing a lot of risk.


I have a simple question for you:  Could you step up in front of a group of people and explain this chart?  If not, why do you believe it?  A smart and influential guy presents something that you cannot really evaluate.  Why is this a sound basis for your decisions?

It is unchallenged because of the lack of peer review in the investment world.  It is challenging to explain the errors, partly because so few could appreciate the arguments.

A plausible story

So many investment arguments depend upon simple analogies that are immediately convincing.  The frog in the pot story (even though it is not true) is one example.  These are stories that enable us to imagine an outcome without any real data.

  • Stall speed for the economy.  Graphic but wrong.  Economic expansions generally do not stall out, despite the intuitive appeal.
  • The aged bull market.  This is another argument that appeals to intuition but has no supporting evidence.  Whenever there is a streak that exceeds normal history — a hitter in baseball, a basketball team winning many games in a row — there is a temptation to say that this must be ending soon.  In fact, a winning team or player is actually just as likely to continue the streak.  Bull markets and economic cycles that have longer-than-average length are no more likely to end soon. (Nice survival analysis from SF Fed).

Your intuition and the confident-sounding talking heads are both wrong.  It is plausible spin.  You can take a profitable contrarian position by betting on further economic recovery.

The insightful investor fights spin with data and analysis.


12 thoughts on “Expensive Misconceptions”

  1. The chart is pretty easy to explain. Do you believe that hussman is falsifying the data?

    If not, then what’s wrong with the chart

      1. It’s pretty self explanatory, can you go into more details as to what you are confused by?

        There are years at the bottom.
        Red line is following 12 year returns.
        Blue line is (Total Market Cap – finance)/ nominal GDP.

        Blue line units are on the left y axis in inverted log scale.
        Red line units are in terms of yield.

        It seems plausible. It sounds like you have other ideas and I would love to see a critique, but misunderstanding the position you are attacking makes for a poor argument.

        1. Michal — Your description of the blue line is not accurate. Even if it were, explaining a piece of research requires describing the relationship, not just the units on the axes.

          I am not confused, and therefore I can see what is wrong with the chart. My point is that people willingly accept such evidence without a real test.

          So try this. The Blue line is supposed to predict the red line, right? We have a prediction in place (although it has been a moving target since the research has changed the chart in the past). What market result in the coming year would make the Blue line prediction correct?

          And BTW, I do not attack straw men. This is an important issue.

          Thanks for joining in, and I hope you answer my question 🙂


          1. Jeff, I’m curious to hear your critique of the chart, so I’ll try to answer what would make the blue line prediction correct over the next year. Returns from Apr 23, 2004 – Apr 23, 2005 (~ +1.1%) will drop out of the interval, and we’ll have Apr 23, 2016 – Apr 23, 2017 added to the 12-year interval. I would estimate that the blue line falls from 7.5% to 5% annual return in the next year, so the 12-year non-compounding interval return would fall by 30%. So I think that means the next 12-month returns would need to be -28.5%.

            There are periods of the chart where the blue and red line diverged a good bit in the short-term, but still provided useful guidance in the long term. If I’m looking for a 12-year prediction, isn’t it misusing the chart to say, “what does the chart say will happen in the next year?”

            Do you think much better of “The Single Greatest Predictor of Future Stock Market Returns” on the Philosophical Economics blog? http://www.philosophicaleconomics.com/2013/12/the-single-greatest-predictor-of-future-stock-market-returns/

          2. Jon — You have the right idea. Let’s think about the implications.

            Please note that you say the chart “still provided useful guidance in the long term.” This is a natural way to phrase it, but it is not accurate. Neither the chart, the equation, nor any forecasts existed for most of the period described. It consists of a regression model of selected past data, engineered to provide the best fit on that history. At some point we need to draw a line and start looking at what happened after the development of the model.

            One way to evaluate the post-development time is to observe as each year is added. A little research will show you that this chart used to be about a ten-year forecast. Why the change? If we do not get the big down year in 2016 or 2017, will we quit seeing the chart? Will it become a fifteen year prediction? Will the regression equation change?

            Finally, anyone who reads this source knows that there is an ongoing prediction of a market crash. The language has become more aggressive over time, most recently including the word “obscene.” No one is looking at this as a twelve-year forecast. Some investment advisors even show this to clients as a way of selling a non-stock product. I do not have any position that I expect to hold twelve years from now. I plan to adjust as circumstances change.

            Thanks for joining in. This is just a summary of my concerns. At some point I’ll write in more detail. I really did want people to see that they are putting too much confidence in impressive-looking charts.

            (Thanks for the link. I have cited Philosophical Economics a number of times, but not the article in question. It is an interesting approach with initial plausibility. I like the idea that the author began with a logical model and hypothesis and then tested it. I’ll give it more thought).


          3. You’re right, about the description, for some reason I kept seeing gnp instead of gva. So yeah this is a more difficult concept.

            >> What market result in the coming year would make the Blue line prediction correct?

            No result in the next year could make the blue line prediction correct because it’s a 12 year prediction.

            The predictions that it does make.

            1. In 12 years the total return on the s&p 500 will be around 3%.

            2. It predicts that if you invested in 2005 you will have seen a total return of 6%, getting the raw data would be nice here, because I’m eye balling it.

            I found this yield calculator, but am unsure of it’s accuracy.

            if you invested in april 2007 then you would have gotten an 11 year return of 7.328% (dividends reinvested).

            So, The prediction is that the next year will be worse than the previous 11 years.

          4. Michal — That is the right idea. I’ll say more in the response to Jon.

            Thanks for helping me to elaborate on this.


          5. Jeff, thanks for the detailed response. I did not realize that Hussman has been modifying the time interval. That is clearly dishonest intellectually. And I agree entirely that he has been dogmatic in his predictions, and seems willing to massage the data to fit the narrative.
            As always, thanks for your thoughts and insights!
            I’ll continue to observe the model proposed by Philosophical Economics with interest.


    1. If you read philosophical economic’s next piece http://www.philosophicaleconomics.com/2013/12/valuation-and-returns-adventures-in-curve-fitting/ you see that the predictor there is a little tongue-in-cheek, basically just to say that these charts are not as informative as you think.

      Also, if I remember correctly, Jeff, you did link to the philosophical economics article! I remember because it was the first time I saw that blog and began following it. Keep up the good work and have a nice weekend off!

  2. Returns over the next x numbers of years arguments are a little obscure. Although, Buffett did famously forecast around year 2000 low returns over the next 10 years. For this chart with say, returns over next 5 or 10 years, not 12 I’d guess the correlation wouldn’t be so pretty. Instructive to think about, thanks.

Comments are closed.