In the many reviews I have received of our new web-site, one question that kept repeating itself was - what does the term dumb-bell strategy mean?. As real estate on the main page of the web site was scarce, I decided to write a piece on this and provide a link to this article for those who want an elaboration of what the dumb-bell strategy is.

Before I proceed, I would direct the reader to Nassim Taleb’s masterpiece Anti-Fragile p161 to p167 where in the dumbbell strategy has been explained with far greater eloquence than I can ever hope to achieve. Nevertheless, I will attempt to explain how I employ this strategy.

The traditional response to uncertainty in the world of finance is to first estimate risk using estimated variances in asset prices and come up with estimates around the probability of a given loss. With this estimated probability, it is then believed that one can optimize return for a given level of risk. Unfortunately this completely ignores the difference between risk and uncertainty. To get a little theoretical, the estimated variances that are used to calculate risk, are themselves subject to variance!! The probabilities calculated are subject to error and this error is significantly magnified when we extrapolate our estimates to calculate the probability of rare extreme loss events (small probabilities) which can be catastrophic in their consequences.

What then can be a response? How do we go about structuring portfolios such that we are protected against extreme loss. One answer is the dumb-bell strategy. i.e. first protect your downside. Allocate a portion of your portfolio to ultra-low risk assets that will stay robust against any market wide mayhem . For the remainder of the portfolio embrace risk whole heartedly by searching and investing in risky assets which can give a huge upside and if they fail you will be at-least protected to the extent you have stayed invested in your ultra-low risk assets. The dumb-bell strategy is a heuristic and simple rule of thumb that provides a practical solution to the problem of uncertainty. The idea is to understand that one cannot predict extreme events, but at least one can protect oneself by knowing what will be the maximum loss if such an event were to occur.

So how do we employ the dumb-bell strategy? We either have (a) ultra low risk investments such as liquid funds, tax free bonds, g-secs or (b) diversified index based funds or (c) basket of high conviction, extremely risky stock ideas. For clients with a low return benchmark such as inflation + 2%, allocation is a split between (a) and (b) and for higher risk clients who have a market benchmark, allocation is a split between (a), (b) and (c). You can see the idea is to either be invested in low risk fixed income products or take low cost index linked exposure to the market or jump in and embrace risk to the extent you want to. There is no middle path.

With his characteristic wry humor, Nassim Taleb’s says it well in p162 of Antifragile that “in risky matters instead of having all members of the staff on an airplane being “cautiously optimistic”, or something in the middle, I prefer the flight attendants to be maximally optimistic and the pilot to be maximally pessimistic or, better, paranoid”