Pilots have to undergo recurrent training to sharpen their skills every year. After numerous routine procedures, on the final day there are tests of the pilot’s ability to deal with emergencies. One of these tests the pilot’s ability to recognize and survive potentially fatal wind shear on landing. Since wind shear is a rare but devastating phenomenon, the practice sessions, thankfully, happen in a full-motion simulator.
Wind shear is a phenomenon that basically results in a sharp change in both the velocity and direction of winds in a very short distance. On landings this can be particularly dangerous. One moment the aircraft feels a headwind and increased lift, and an instant later it turns into a tailwind (see picture below). The conditions that make wind shear likely typically occur in the vicinity of a thunderstorm with downdrafts. A pilot who has reduced power and speed, as he would normally do when landing, can find it impossible to fight the downward forces from the shear and this can result in tragic consequences (Delta flight 191 into Dallas in 1985 is a tragic reminder of this peril).
To avoid wind shear’s devastating consequences, tests in the simulator teach pilots to recognize and react in a decidedly counterintuitive manner. While on a normal landing a pilot would pull power back and reduce airspeed, in a wind shear condition power goes to full to maintain airspeed and lift! A few times in my own practice I did the “normal” thing and instantly was penalized and pulled to the ground as the simulator screen went red. I was reminded repeatedly, against my normal impulses, to maintain forward momentum, since any loss of speed results in loss of lift, and loss of lift results in the inability of the aircraft to fight the downward forces of the shearing wind.
Similarly, in a recent paper, I wrote about the need for new thinking when faced with the bimodal outcomes we are currently witnessing in markets and economic conditions (Vineer Bhansali, “Asset Allocation and Risk-Management in a Bimodal World,” PIMCO Viewpoints, December 2011). The two major conclusions were that first, asset allocation needs to be more defensive in nature when faced with the probability of multiple equilibria, and second, explicit tail hedges that look expensive in a normal world may indeed turn out to be cheap if the unimodal morphs into the bimodal.
There is a third, important conclusion I want to address here: when faced with bimodal outcomes, momentum as a risk factor becomes potent, and cost-efficient exposure to momentum becomes critical to proper portfolio construction. This, like the recipe for wind shear avoidance by pilots, is counterintuitive thinking. (See V. Bhansali, “Market Crises – Can the Physics of Phase Transitions and Symmetry Breaking Tell Us Anything Useful?,” Journal of Investment Management, 2009, and for a brilliant explication of the underpinnings of the phenomena, see Mohamed A. El-Erian and A. Michael Spence, “Systemic Risk, Multiple Equilibria and Market Dynamics – What You Need to Know and Why,” PIMCO Viewpoint, March 2012).
Investing in momentum is not something that we are used to doing since we have grown up in a world of unimodal, normal distributions. In a unimodal, normal distribution, markets spend most of their time in the middle of the probability distribution, since the thin tails at either end imply that mean-reversion forces the probability to be highest close to the middle. In a bimodal world, the middle of the distribution is where the markets will spend the least of their time, and they will traverse between either end of the probability distribution. The dynamics of this traversal are exactly the opposite of mean reversion, and are accompanied by momentum in markets. In other words, the same mean-reverting forces that create the normal distribution are replaced by trending forces that create bimodal distributions.
Figure 2 shows the correlation of various hedge fund strategies with the VIX (the Chicago Board Options Exchange Volatility Index which measures the volatility of S&P 500 options). VIX is a good measure of the turbulence or wind shear in the equity markets, so negative correlation with VIX (as exhibited by most of the strategies shown below) shows that market turbulence may result in destructive forces on typical mean-reversion based portfolios. On the other hand, the momentum strategy (Managed Futures in Figure 2) and the short-bias strategy both exhibit zero to positive correlation with this turbulence, with the correlation rising in periods where the turbulence is especially bad. While underweighting relative to the index helps cushion against market downdrafts, it may be expensive in terms of opportunity costs to not hold equities for an extended period of time. On the other hand a small allocation (3%-5%) to momentum-based strategies may provide significant risk mitigation benefits. This is because of the asymmetry of the momentum factor returns: it tends to pay off large when the bounds of mean-reversion break, and effectively helps keep the portfolio from slamming into the ground.
And for an illustration of how momentum can be a source of positive returns, see Figure 3. It shows that the Barra momentum style factor has demonstrated significant cumulative positive returns over more than 15 years. (Barra is a brand of MSCI. Barra factors are represented by MSCI indexes and are meant to reflect target exposure to factors such as momentum, earnings or leverage.)
Obtaining exposure to momentum is not that hard, but it forces one to think very differently about markets and investing. The momentum strategy consists of (1) buying what is going up and selling what is going down, (2) buying more of what is going up more and selling more of what is going down more (called pyramiding), and (3) stopping the buying and selling at predetermined position sizes for risk management. This is the perfect antithesis of value investing where “buy low sell high” is the mantra, yet there is incredible value in this style of investing if (1) done cost efficiently and (2) in a controlled fashion.
In a set of illuminating papers written over a decade ago, academics Fung and Hsieh demonstrated that this strategy is theoretically and empirically the same as buying an “option straddle” (a call and a put), and hence offers exposure to rising market volatility. The potential benefit of the strategy is that by following a limited set of transparent rules it can avoid high option premiums that one would pay in a straddle. When implied volatilities are high, such dynamic strategies can indeed become cost efficient.
Of course the implementation is not free – if markets do not trend but mean-revert repeatedly, the “whipsaw” can cause the strategy to buy high and sell low continuously, creating losses that can add up. Empirically, these whipsaw effects have been relatively low compared to the potential for attractive gains as in 2008. Other researchers (C.S. Asness, T.J. Moskowitz and L.H. Pedersen, “Value and Momentum Everywhere,” NBER, 2008) have shown that momentum is not limited to any one market, but is actually “everywhere,” and thus the strategy is best implemented across all asset classes (stock indexes, bonds, commodities and currencies). In addition, the momentum factor tends to do better when there are periodic bouts of illiquidity as is typically the case when risk-aversion rises. Since all of these markets have liquid futures contracts, the momentum strategy can actually be implemented very cost efficiently as a collection of long and short positions in futures contracts.
In a world of zero interest rates the potency of bonds (or duration) to provide diversification-based hedges is very limited (e.g., at a 2% yield the maximum capital gain if 10-year yields fall to 1% is only about 15%). In this world of low, pegged interest rates, an investor who is going to take risk needs other means to make the portfolio more inured to unforeseen shocks and market storms. In my opinion, the only hope for investors to stay in the air long enough to avoid the forces of wind shear that accompany these uncertain prospects is to look at effective alternative beta strategies, such as momentum, that can be implemented efficiently. Way back in in 2003, when we first started to implement tail hedging strategies for our clients, the phrase “tail hedging” was an outlier, and indeed the practice was not part of the investment vernacular as it has become today. If we are correct in our forecast of a bimodal world, the value of being invested in the momentum risk factor will likely prove to be just as relevant as tail hedging has been over the last five years.