One of the most quoted sayings in investing goes like this: “If you take care of the downside, the upside will take care of itself.”
This sounds so simple. People say they control risks by considering the downside in every investment and they make decisions based on the so-called risk-reward ratio – risk being the downside and reward being the upside. Very few investors appropriately implement the idea of taking care of the downside. Even many of the gurus listed on this forum failed to heed this piece of timeless wisdom.
Thera are a few reasons which make it difficult to implement an effective risk control system.
First of all, the definition of downside can be vague. Some use the phrase “worst case” and some use the phrase “potential loss.” But either way for us to define downside, we have to define risk first. Howard Marks (Trades, Portfolio) has the best definition of risk I have ever seen –“risk means uncertainty about which outcome will occur and about the possibility of loss when unfavorable ones do.” If we take a step further, what Marks implies is there are essentially a number of elements involved:
- Assessment of the range of outcomes.
- Assessment of the probability of each outcome and which outcome is most likely to occur.
- Assessment of the probability of loss when unfavorable outcomes occur.
- Assessment of the magnitude of loss with each unfavorable outcome.
While I like Marks' definition, it’s missing a very important element – time horizon. The probability of a certain event varies enormously with different time horizons. Sometimes the longer the time horizon, the less likely an event will be to happen. Sometimes the shorter the time horizon, the less likely an event will be to happen.
Combining Marks' definition with the added variable of holding period, perhaps we can come up with the following definition of downside: the maximum magnitude of loss when the most unfavorable outcome occurs, given a certain holding period.
But even this definition has drawbacks – what about probabilities? Should we use a 1% likelihood of an unfavorable event? 10%? Or 50%? There is no definitive answer, but it’s a call we have to make. Making things more difficult is the fact that knowing the probabilities doesn’t mean you know what’s going to happen. In investing, very frequently less likely events happen and likely events fail to happen.
Investors often fail to recognize the inherent unknowability and unpredictability of a business and hence, the unanalyzability of the downside risk. They often overestimate their ability to analyze a business. To quote Mark Twin – “It’s not what you don't know that gets you into trouble. It's what you know for certain that just ain't true.” I’ve seen so many times that catastrophic loss happened because people thought they knew something, acted with high conviction and turned out to be wrong; that's how you lose a lot of money. The common factor is failure to recognize the edge of one’s circle of competence. Many currently frequently discussed examples fall into this category – Horsehead Holdings (ZINC, Financial), Valeant (VRX, Financial), Exco Resources (XCO, Financial), Avon Products (AVP, Financial). How many retail investors lost money piggybacking Mohnish Pabrai (Trades, Portfolio), Bill Ackman (Trades, Portfolio) and the Sequoia Fund?
The human brain is wired to take mental shortcuts and to prefer quantification over unquantification. In investing, this is manifested in the phenomenon of blindly applying a historical low multiple to a poorly analyzed fundamental metric. I cannot even count the times that I’ve seen an argument for worst case that goes like this: "If we apply a historical low P/E multiple to normalized earnings of x, we get a worst case of y." This is a weak argument, but most people find it reasonable because on appearance it makes sense. However, if we go a step further and think about the hypotheses involved, we can easily tell that the most important elements are missing. We have to ask why not use a fundamental that reflects an unfavorable outcome. We have to define normalized earnings, and we have to know whether this time is different. If we get any of them wrong, our worst case is bamboozled.
Most people have a bias for buying, meaning that they find a seemingly cheap stock through quantifiable screeners such as low P/E or low P/B and then they find confirmation that the stock is actually cheap so they can buy it. Their goal is to avoid missing out. Don’t get me wrong, the risk of missing out, or what Warren Buffett calls the mistake of thumb sucking, is real. But for most investors, the thought of missing out triggers deprival superreaction, which then causes us to act in a biased way – that is to use optimistic assumption in both worst case and sell point assessment and to find confirming evidences to justify our biases. This is a common process error. But to fix this error is no simple task.
Superior investors automatically consider all of the above in their risk assessment. It’s not supposed to be easy. As you can tell, a lot of judgment is needed, even after the painstaking amount of work one has to do and after the rigorous, diligent and often mentally challenging analytical process. But to not consider the aforementioned factors would not be prudent at least, and could lead to disastrous results at worst.
