What do we really know to be true about the Stock Market? That is, what do we know for sure to be true, as opposed to what we believe to be true? What we present as fact should not be based on belief (or on results of badly designed studies) – it’s about having mathematical confirmation, or, lack of confirmation, which is just as important. If we don’t know something is true, we should not assume it, and much about the Markets that we think is true is at best true for a short period of time, and at worst – not true at all.
The crash of 2008 resulted in most portfolios taking a beating. In response, many advisers announced that they would drop buy and hold and would turn instead to market timing and active trading. This after years of research showing that active management underperforms the indexes in the long run. If anything, this showed how unprepared the investment professionals are when it is time to face reality, and how little faith they have in their skills and knowledge if they are able to throw the baby out with the bath water for no reason other than having failed to realize that they’ve been operating under a false premise.
If examined closely, much of what is currently taken for granted is at the very least doubtful – this much we know based on solid research (please see below for references). What are the implications? For starters, numerical tools are all suspect – models are worse than wrong because they are never accurate enough, and errors can overwhelm any ‘measurements’, especially when trying to make predictions about the future. Much of today’s research can not be verified to be correct except by using past data, and many times the researchers are misusing the statistical tools to derive the result they are looking for, as well use questionable techinques which are often simply wrong . It takes some practice and know-how to spot these studies. In and of themselves, the studies are consistent, yet most of the time the assumptions are violated, which invalidates their conclusions . Markets are far from Gaussian (20+ sigma events are much more common), and ‘outliers’ are the ones that can make or break a portfolio. Stocks can not be assumed to go up – it may be possible to have long periods when stocks are declining, or not much is happening (mean reversion can not be taken for granted) . Holding stocks for a long time is no guarantee either (time diversification is a fallacy) . History can not be used to predict the future – the generator itself can be changing (in a multifractal world) [5,6]. Are markets efficient? Not at all – Mandelbrot has shown that markets have memory, yet quantifying this is still an unfinished task.
The problem is, when dealing with the Stock Market, even doing the wrong thing can sometimes be profitable, and sometimes not . There is no approach that works all the time, and given the market’s volatility, it only takes one mistake for an approach to fail so spectacularly as to wipe out all of the profits obtained previously, especially when using leverage . Now we are back to buy and hold. The implication for advisers is to use much more conservative approaches with client portfolios. For those who like index funds, how much fun is it trying to hold S&P500 index through its gyrations? There is no guarantee that it will deliver even over the ‘long term’ – 10, 20 or 30 years. To obtain more consistent returns, much more attention has to be paid to bonds and their quality, and much less attention to the markets and what anybody else is doing. An adviser can add value by not doing all the things which are guaranteed to be wrong at some point and for which clients will pay dearly. It is just fine to underperform the market, if more consistent returns are desired with much less risk. Avoid trying to make predictions based on inconclusive evidence – it is better to stick to what you truly know, and to avoid what you truly do not know. Active management is not a panacea – it is a disaster waiting to happen, because it is based on the premise that a skilled investor can predict the direction of the markets, which is just another fallacy .
Some of the past and present ‘heretical’ research:
 Backtesting, flawed studies (many studies done incorrectly).
Amit Goyal and Ivo Welch “A Comprehensive Look at The Empirical Performance of Equity Premium Prediction”: link
 Mechanical filters and trading strategies (no consistent evidence in the past decade, more studies needed).
The Profitability of Technical Analysis: A Review: link
 Mean reversion (no consistent evidence).
The Long-Term Risks of Global Stock Markets, Philippe Jorion: link
 Fallacy of time diversification.
 Power law/non-Gaussian, fat tails.
 Multifractal Model of Asset Returns.
Mandelbrot, Fisher & Calvet 1997, “A Multifractal Model of Asset Returns”: link
 Black Sholes is wrong.
 Can history be used to predict the future?
Analysis can not predict earnings accurately. Even if you were able to predict earnings, profiting may be impossible (see page 75) of “Investment Management: Portfolio Diversification, Risk, and Timing–Fact and Fiction”: link