Source: Vanguard


Over the weekend, I set the Closer Look at Mythology graphic to post early this morning.

By sheer coincidence, Meb Faber pointed us to the above chart from Vanguard looking at the correlation between traditional valuation metrics and subsequent investment returns. Given the serendipity of how similar the messages were in both, I thought it was worth exploring further.

What I find worth especially discussing is the simple observation that so many of the traditional metrics — the assumed truths of Wall Street — fail to withstand close scrutiny as having forecasting value. These include such varied metrics dividend yields, economic growth, Fed Model, profit margins, and past stock returns.

A few simple criticisms of the Vanguard piece are in order; yes, they are talking their book — forecasting is folly, asset allocation into broad indices is the best bet for investors — but that does not undercut their analysis. Perhaps a more significant complaint is a peeve of mine about single variable analysis — that looking at any one metric alone to explain complex systems (such as investment outcomes) is doomed from the start.

Regardless, this paper is worth reading. Keep it in mind the next time you see someone trotted out on TV to claim that stocks are a great/terrible buy RIGHT NOW because of any single one variable . . .


Single vs. Multiple Variable Analysis in Market Forecasts  (May 4th, 2005)

Forecasting stock returns: What signals matter, and what do they say now?
Vanguard research October 2012

Category: Investing, Markets, Really, really bad calls, Technical Analysis, Valuation

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

22 Responses to “Everything You Know About Investing is Wrong . . .”

  1. wcvarones says:

    It would be nice to see how those predictors worked out out of sample. Say, apply them to Japan the last 20 years.

    You really think we have a 40% chance of real returns greater than 8% (fig. 6b) the next 10 years? With interest rates at 0%? That implies an obscene equity risk premium. Not bloody likely.

  2. Jack Damn says:

    [...] forecasting is folly [...]

    Honest question here:

    I believe you have written the above statement before on your blog, though I don’t have any links to prove that, and while I too believe forecasting is fatal, my feeling is that you must use some forecasting when choosing stocks/funds/ETFs for your clients.

    Is that correct?

    I don’t have any clients and I’m not running “big money” … I’m just a bottom feeding retail swing trader technician, but I’d have to think that if I was one of your clients or if I was handling the type of money I imagine you’re handling, there must be some forecasting in the portfolio you suggest to your clients or else why would I, as your client, trust you when you say, “Buy Apple here.”. Why would I do it? You’d have to give me some reasonable explanation of the potential (potential, not guaranteed) future return of your recommendation.

    To do so, it seems to me, you’d have to inject forecasting into your stock/ETF/fund selection. Am I splitting hairs here? I only ask because I see statements like “forecasting is folly” plastered all over Twitter/Facebook/Stocktwits, etc…, but I often wonder how people handle that when they have clients or run huge sums of money.

    Can’t be pure Quant or pure charting … I mean I guess it could be a pure math driven portfolio, but even though the right side of the chart hasn’t formed, it is still a bit of alchemy and forecasting.

    Thanks for any thoughts.


  3. You may be referring to The Folly of Forecasting and other such sentiments.

    As to buy or sell decisions, rather than specific forecasts, they are (or should be) quantifiable probability analyses. When these factors are present, what does the distribution of potential outcomes look like? Stated another way, based on [these assortment of factors], statistically, does owning stocks/bonds/commodities here generate are high/fair/low returns for relative to the high/fair/low risks assumed?

    I would also suggest seeing James O’Shaugnessy’s What Works on Wall Street — it looks at multiple metrics in combination across different time periods and eras. Excellent explanatory as to a quant approach versus predictions

  4. streeteye says:

    Gold mine… for one thing, one year returns don’t correlate well with anything at all… for another, ‘consensus’ doesn’t correlate with anything… and for all the yowling on TV, one thing that sticks out is positive correlation of government debt/GDP. Stocks have done well over 10-year periods where the government has stimulated the economy and the debt has soared.

  5. [...] the original post: Everything You Know About Investing is Wrong . . . | The Big Picture Comments [...]

  6. Jack Damn says:

    Excellent. Thanks Barry. Appreciate the reply and book recommendation. I see there’s a Kindle Edition. I’ll snag that this morning.

  7. rd says:

    I think one of the biggest issues is that investing and trading get confused with each other. Very few people in the financial sector are investors because most of their time frames are too short. The average person should be an investor, but not a trader. For example, I will often go a couple of years or more with no portfolio changes other than regular dollar-cost averaging into retirement accounts and occasional rebalancing. The average Wall Street person could not even conceive of doing this. I won’t end up a billionaire but I am also highly unlikely to be bankrupt or blow up the world’s financial system. My investing timeframe is my estimated remaining lifespan, not just the upcoming monthly and quarterly reports.

    I track a lot of these indicators but things like Shiller’s CAPE and dividend yield are critical to me for understanding what my current investing stance should be based on market valuations. Similarly, big things that are hard to change overnight like GDP/debt ratio and demographic trends are likely to be good indicators as to whether or not an economy is robust or fragile over the next decade or two. Meanwhile, interest rates are a good indicator of what the “risk-free” return is that potential stock market returns should be compared to.

    So the average person could do very well simply investing in Vanguard LifeStrategy Funds and shifting between the Aggressive Growth (bottom CAPE quintile), Moderate Growth (middle 3 CAPE quintiles), and Conservative Growth (top CAPE quintile) ones based on the Shiller CAPE valuations. A handful of shifts in their lifetime and they would largely miss out on the major bears and generally participate in the major bulls.

    The last five years would be a bit of a conundrum due to the unprecedented control over the bond and money markets that the Fed and ECB have exerted but just hanging out in Moderate or Conservative Growth resulted in steady returns with muted risk.

  8. ChuckC says:

    “that looking at any one metric alone to explain complex systems ”

    And yet . . . Global warming is caused by Co2. And the ‘forecast’ 10, 20 50 years from now is catastrophe . . . . . .


    BR: Who ssaid CO2 is the sole cause of AGW? (This is yet another example of poor rhetoric).

    Scientists have named Fossil fuel (Coal, Diesel, Nat Gas) burning power plants, emissions from burning gasoline for transportation, Methane emissions from animals and agriculture, Deforestation (especially tropical forests) and farmland Increase in usage of chemical fertilizers on croplands. But dont trust my memory:

    Here is National Geographic:

    “There are several greenhouse gases responsible for warming, and humans emit them in a variety of ways. Most come from the combustion of fossil fuels in cars, factories and electricity production. The gas responsible for the most warming is carbon dioxide, also called CO2. Other contributors include methane released from landfills and agriculture (especially from the digestive systems of grazing animals), nitrous oxide from fertilizers, gases used for refrigeration and industrial processes, and the loss of forests that would otherwise store CO2.

    Different greenhouse gases have very different heat-trapping abilities. Some of them can even trap more heat than CO2. A molecule of methane produces more than 20 times the warming of a molecule of CO2. Nitrous oxide is 300 times more powerful than CO2. Other gases, such as chlorofluorocarbons (which have been banned in much of the world because they also degrade the ozone layer), have heat-trapping potential thousands of times greater than CO2. But because their concentrations are much lower than CO2, none of these gases adds as much warmth to the atmosphere as CO2 does.

    In order to understand the effects of all the gases together, scientists tend to talk about all greenhouse gases in terms of the equivalent amount of CO2. Since 1990, yearly emissions have gone up by about 6 billion metric tons of “carbon dioxide equivalent” worldwide, more than a 20 percent increase.”

    Go sell stupid elsewhere. We ain’t buying.

  9. [...] Barry: Everything you know about investing is wrong.  (TBP) [...]

  10. rallip3 says:

    One of the problems with forecasting US stock returns is that this is forecasting a subset: US stocks are a subset of World stocks which are in turn a subset of all investible assets (real estate, collectibles, commodities, derivatives, cash, fixed interest, etc …). This means that even if we could forecast the real return on the entire set of world investible assets, the allocation of that performance between the competing asset classes remains variable and unpredictable.
    A fashionable common-sense approach to forecasting world all-asset returns would be to consider world demographics and world rates of accumulation and consumption of capital. As far as I can see, these are not considered in the Vanguard study.

  11. b_thunder says:

    The only 2 rules that matter are:
    1. Buy low, sell high
    2. Don’t lose money

    All other “rules of thumb” taken by themselves are useless. They only work when the “rest of the crowd” follows those rules, and you can front-run “the crowd.”

  12. Francois says:

    There is an excellent interview of Meb Faber by Michael Covel dated 06/19/2012 for those who might be interested.

    The answer to your question is “No”; I am not affiliated in any way shape or form with Michael Covel, except as a listener of his podcasts.

  13. foss says:

    If only they had looked at trading on inside information…

    I bet that one has a slightly higher R squared :)

    After all, that’s one of the ways the big boys do it, since they know all of the above is true.

  14. Pantmaker says:

    JD- the “forecasting is folly” mantra is industry speak for “you have no business trying to manage your own money…put my firm on direct deposit and leave the non-forecasting of you investment decisions up to someone who can get paid properly for it.”

  15. I don’t speak on behalf of the industry, only on behalf of myself:

    Every speech I give to main street investors begins with the following: “The best investing advice you will ever receive is that you should dollar cost average each month into 3 or 4 broad (low cost) indices. History teaches us that most of you will ignore this excellent advice in pursuit of glory and or entertainment.”

    For much of the public, that is what they should do (but don’t).

    For people who have a) Behavioral / Discipline concerns; b) complex tax circumstances; c) more complicated inter-generational wealth transfer issues — they need some help.

    That is the value add good practitioners provide. I cannot speak to what anyone else is doing.

  16. wally says:

    Perhaps I’ve just missed it, but I didn’t see a description of the calculation of real returns in that paper. I ask because for instance, the 1-year predictable return from a dividend stock, let’s say AT&T for example, is quite high. The calculation must therefore assume buying and selling to capture return of gains… which is always a timing issue, whether short or long term. But any timing method must always assume some sort of escalating structure (ie: getting out of one stock high and into another stock low) in order to increase an actual portfolio beyond the general market (only dividends provide an internal method of increase of holdings). So, in that sense, Vanguard is really talking their book, which is that a long-term 6.8 percent return is really the only reliable expectation.

  17. So it seems that PER again is the undisputed winner in forecasting. Very much in line with “What works on Wall Street”.

    What would happen if we couple PER readings with a trend filter (200 MA and/or Dow theory)?

    Answer: too nice and simple to be true.



    BR: Price Earnings Ratio ? The best performer was 10 year P/E, followed by P/E ratio — but both of them stunk.

  18. faulkner says:

    Re: Single variable analysis. System 1 (Kahneman) naturally associates two experiences, images, information, etc. One form of this is “After it, therefore because of it.” This is manifest in System 1 beliefs like “Everything happens for a reason.” When System 2 emerges, this mental process takes the form of “What is the reason?” with the search subsequently singular. It takes attention to the form of one’s thinking to catch this, and ask better questions.

    The only thing worse is having multiple (disparate) variables without any overarching conceptual framework and evidence process to know how to relate and weight them relative to each other. At least with a single variable you can know when you’re wrong – even if most don’t. With multiple variables, the overwhelming response is to keep tweaking.

  19. Jonathan1 says:

    So, on behalf of the Little Guy, who is already dollar cost averaging into a diversified portfolio of Vanguard funds just to save for college and retirement, not to try to get rich quick a la SAC, does using the 10 month moving average alone still make sense as a sell signal?


    BR: The historical data suggests its a high probability exit strategy. (Note there are better entries than the 10 month)

    See this: A Quantitative Approach to Tactical Asset Allocation

  20. Greg0658 says:

    BR that advice at 11:05am I thought was solid ..
    but I’m still only going to > spend what I earn – I don’t think I can do anymore for the economy ..
    save in this system ? I did that in my other life (by law) – if its still there ? well :-| it will legally return when I get (again) to that moved goal post

    -I don’t buy illegal drugs and send my cash underground for that world growth
    -USA efforts in the world – I voted with best intentions – can do little else
    -Xmas gift giving for the 40% of the 70% economy – I’ll do whats required (I think)
    -I was asked to ride along to a capital stuff improvement seminar (I’m going) (its not today)

  21. eurostoxx says:

    aside from adding variables to ones analysis, a model capturing all the effects of the stock market (or any other market) would likely need various regime switching signals and models. Nothing works all the time.

    Also I dont see what is so bad with a 0.43 R2 for 10 year PE. still seems statistically significant to me

  22. [...] What factors matter for forecasting stock market returns.  (Vanguard via Big Picture) [...]