Skip to content

Regression to the Mean and other difficult Truthes

May 4, 2012

Have you ever wondered whether your stock picking skills add any value? Or whether you should just put all your money in an index fund and dedicate your energy to another pursuit? If you haven’t thought about it, you should stop reading this now, sell all your individual stocks and put your money in an index fund, because if you think you are a Great Investor then you probably aren’t! I wonder every day whether I’m adding any value to my personal wealth by investing in individual stocks. After five year’s of intensive reading and reflecting on investing, as well as practicing the art, I’m still not sure. I plan on being ‘still not sure’ in 10 or 20 years. In fact, I firmly believe that the more you believe you know something, the less likely you are to be right. I’m less sure today than 5 years ago that there is a clear path to becoming a superior investor. I’m not even sure that Warren Buffet is a great investor. It’s possible that Lady Luck simply holds all the cards; that nothing we do can make us for sure a better investor, or at least, that nothing we do can guarantee a superior investing performance. We may be a Great Investor and still have a mediocre track record. It’s perhaps a hard thing to contemplate, the thought that we aren’t really masters of our own destiny, that we are where we are, not because of our own aptitudes, capabilities and force of will but because we are lucky, or not. Our need for narrative has us weave a plausible ‘causal’ relationship between investment performance and ability of the portfolio manager; better-than-average investment performance equals better-than-average portfolio manager. We know that this is not true over a year or two, and perhaps it’s not true even over 40 years. Even after adjusting for risk (if that is truly possible) the relationship between performance and ability is tenuous at best. Perhaps there is just no relationship at all. If we’re lucky at most we can cling to Pasteur’s statement that “chance favors the prepared mind”.

If you have been reading Daniel Kahneman’s latest great book “Thinking, Fast and Slow” you’ll know where I’m coming from. If not, you should pick it up right away. It helps one reflect on this issue of luck vs. skill. It helps put into perspective our innate sense that experts are truly experts in the way we think. It helps us understand and accept how potentially large a role chance may play in our destiny yet how we are hard-wired to prefer our own self-aggrandizing narrative to the simpler explanation of  luck.

Enough theorizing. Let’s get concrete. We know that, by definition, the sum of the performance of all portfolio managers is simply the performance of the market, and thus portfolio managers in total add no value. In fact, they detract value since we have to pay them for their ‘services’ and there are transaction fees. Yet knowing this doesn’t  change our behavior. Those with funds to invest in equities still spend much time and energy searching for a portfolio manager that will outperform the market. Even if such a manager exists what makes them believe that they have the tools to identify such a manager? I have no idea.  I myself would have no idea of how to go about such a search. We ‘know’ that past performance doesn’t indicate future performance (after all, you can read this in every mutual fund prospectus) and rationally accept it as true but still act as if it weren’t. We may even be sophisticated enough to accept the concept of regression to the mean (managers that outperform in one year will likely underperform the next and vice versa), but here, too, we rarely use this knowledge to our benefit. In fact, we would have a hard time implementing any strategy related to regression to the mean as regards portfolio managers; Can you imagine selecting a portfolio manager based on his/her miserable performance over the past couple of years? We’re attracted to success; we want to believe that a string of years of outperformance is due to skill, not luck, and that the string will continue. We’re programmed to look for the causal relationship.

So, what about selecting individual stocks when we are our own portfolio manager? Can an individual investor picking individual stocks ‘beat the market’ because of skill? Same story. It’s hard to be sure. There’s too much white noise in the statistics. Even if you show me 1, 2, 5 or more years of outperformance, if I had to bet whether it was luck or skill I would reluctantly bet it was luck. After all, with enough people managing their own portfolio somebody’s going to be lucky 1000 times in a row. It’s simple probability.  The question really is, therefore, how can we harness this knowledge to improve our individual investment performance. I think the first step is to accept the concept that one can improve the probability of a superior outcome without a superior outcome actually happening, in the same way that you can improve the probability of a payoff in Blackjack by understanding the probabilities of the game, without actually coming out ahead. Luck can, and does, go against you. Once you accept that, you can build an investment strategy that has a probability of a better outcome, though you still have to accept that it is only a probability. I know this implies that you shouldn’t use your performance to judge your success which is rather difficult concept for all us stock hounds. But get used to it!

The next step is looking at what has worked historically in the market. A good place to start might be Tweedy Brown’s aptly titled paper “What Has Worked in Investing”. There are a number of other papers discussing the performance of low (value) vs. high (growth) P/E stocks and likewise for P/B stocks. Recently Greenbackd has had a series on ratios filters producing outperformance. Could it be this simple? Was Joel Greenblatt right that a formula can perform as well as any human stock picker? If so, what makes a formula better than the human mind? For insight into this, again I recommend Kahneman’s book “Thinking Fast and Slow”.

As I was contemplating this formulaic approach to investing, the name Walter Schloss came to mind. He seems to embody this kind of semi-mechanical approach, well really just a very disciplined approach, applying it rigorously with little variation for over 50 years, and we know what his returns were. Yes, it sounds a bit boring. No big bets. No big contrarian posture, no big story, just a plodding kind of strategy that requires more discipline than intuition. Come to think of it, doesn’t Buffett do pretty much the same thing? Most everything goes into the ‘too hard’ pile. Only simple companies with easily understandable businesses. They have to pass a few key tests. The big difference with Buffett is how he structured his investments, using a public company for captive capital and the insurance business for its float.

I know you’re thinking..boring! I don’t want to invest like that. Perhaps that’s just the reason it works. There is little human intuition involved. Joel Greenblatt, I think you’re on to something; a system that reduces to the minimum human foibles. Greenblatt’s magic formula is out there. Why isn’t everybody using it? The answer to that is the answer to why it works. We are human and our hubris leads us to believe we can make decisions in real time better than any formula. Too bad we’re wrong.

For the past year I’ve been trying to reduce the number of positions in my portfolio and concentrate my investments in the few ‘best’ opportunities I’ve identified. But given my musings above, it may be time to rethink that portfolio strategy and develop a more mechanical approach to building a portfolio. Something to reflect on over the coming months.

  1. I have been having a very similar “internal dialogue” of late. I tend to use a lot of quant screens for idea generation and then I look into the qualitative aspects once I know for sure the quant box is ticked. I have toyed with the idea of allocating 10-30% percent of my portfolio to a Magic Formula or Piotroski Score basket but I have yet to take that leap.
    I totally agree with you that there is a substantial amount of arrogance in the assumption you can do better than the 20-30% CAGR that these screens have historically offered.

  2. xmfconnor permalink

    Good, reflective post. However, I’d note two things:

    1) While you are correct PM’s do not “add value” in the sense that the summation of their performance is the market, that does not mean they do not add value, in the sense of utility. In other words, PM’s can allocate funds more efficiently to target their investors risk tolerance.

    2) Have you read Buffett’s “coin flip” essay? While Buffett and many other famed value investors may be due to luck, it is very improbably that the greatest sources of excess return (over the very long run) come from a very similar investing style– disciplined value investing. Not only that, but many of them are connected to each other (Buffett, Graham, Tweedy Browne, Columbia Business School, etc.)

  3. Thanks for your comment xmfconnor. I’m not sure I completely agree with you about portfolio managers. What is the utility that we are supposed to pay for? If you think of risk as the efficient market theorists do, then reducing risk (volatility) is tantamount to reducing return, and vice versa. I think 2% is a bit too much to pay for creating a portfolio matched to the risk (return) tolerance of the investor. It would be more efficient for the investor to simply use a combination of treasury blls and index funds and pay a fraction of a percent. If, on the other hand, you view risk as a value investor does, i.e. volatility is irrelvant and risk is simply the inverse of the divergence of price from intrinsic value, I hardly believe portfolio managers in aggregate have added value there.

    As to your second point I agree that the evidence seems to point to value investing as a method thatproduces superior returns. However, we have to define “value investing”. Portfolio managers that adhere (by their own definition) to value investing criteria achieve different results. We hear about the successful ones. Who’s to say there aren’t just as many unsuccessful ones? My point is why do you think that you (anyone) can implement a value investing style that will be successful? Many have tried to copy Buffett and failed. Could it be simply due to the fact that luck has played a very large part in Buffett’s success? The question needs to be asked and explored; one has to be careful not to fall into the trap of believing that success equates to skill.

  4. Alpha Vulture permalink

    Quote: For the past year I’ve been trying to reduce the number of positions in my portfolio and concentrate my investments in the few ‘best’ opportunities I’ve identified. But given my musings above, it may be time to rethink that portfolio strategy and develop a more mechanical approach to building a portfolio. Something to reflect on over the coming months.

    Maybe you should also worry about changing your portfolio strategy frequently. Maybe value investing does work, but not if your time horizon is not long enough.

  5. You’re perfectly right that from a scientific approach, changing one’s investment strategy frequently makes it difficult to assess whether one’s strategy is working or not, as the feedback loop in investing is so very long. But should that stop one from questioning whether the one’s basic investing strategy is right or wrong (or just needs to be tweeked)? I don’t think so.

    For me ‘blind faith’ in anything is a road I don’t want to travel. That doesn’t mean I don’t subscribe to value investing whole heartedly. I do. But look at all the different successful approaches there are to value investing. Which one is better? Actually I think the question each of us should ask is which one is better for me. As soon as I find that approach I’ll stop blogging.

  6. Jay, great post its something I ask myself all the time.

    A friend and I recently finished a study on the European markets testing various strategies over the past 12 years. The best 10 strategies all had returns of more than 730% compared to just over 30% for the market.

    Because of taxes and dealing costs the 730% is not achievable but still a great deal better than most investors ever saw.

    Here is the Google books link:

  7. Thanks Tim, I’ll definitely check it out. I was looking for some summer beach reading.

  8. peter permalink

    In regards to the greenbackd article, I have looked at these mechanical investing type ratios before and I will caution you to really examine the data before you make a decision. What I found was that there are a number of simple investment strategies that have beat the market over the past 3 or 4 decades but, most of them have stopped beating it over the past 5 or 6 years. They had sufficient margin over the average in the first 35 years that they still beat it when you look at the past 40. So you can see how just looking at a simple average is misleading. It is possible they will continue to outperform in the future but then again, with all the algorithmic trading going on, perhaps the market has just changed? Just a thought.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: