Leverage is such a dirty word in finance. Images of greed, gluttony and eventual doom spring to mind whenever it’s mentioned. Stories of high flying, and high failing 3x daily-leveraged ETFs. The vol-pocalypse of February 2018. Other miscellaneous disasters all involving leverage are all brought up.
What is leverage though, exactly? As a simplification think of it as a scaling factor for daily returns. So if you have daily returns r, then l*r would give you the total returns, where l is the leverage factor. When l = 1, we get 1*r i.e., simply r back, thus no leverage.
Do you see the problem here? Leverage is a mathematical concept. It’s just everything on the number line between 0 (holding cash) to infinity (can’t do that, we’d go bankrupt instantly) that is not the number 1. Well, not really infinity but some crypto swaps do try with 100x leverage on offer 🙂
Some researchers have tried to debunk the myth of leverage being bad. Tony Cooper has a short understandable piece here. Lots of systematic fund managers that use volatility targeting use leverage (Rob Carver’s Leveraged Trading is a good introduction book). But it’s usually a word that’s avoided, instead referred to as ‘volatility targeting’ as we just did above.
How can leverage be useful? Let’s zoom out a bit, and think about buying and holding an asset. Let’s continue and say we have a given risk profile defined by a volatility / standard deviation we can withstand. Then the only time we should ‘buy and hold’ an asset at 1x leverage (meaning 100% in) should be if its volatility never changes and is exactly equal to our given risk threshold. That never happens in reality. For the SP500 (we’ll use SPY), the volatility jumps around dramatically over time. We have super calm time periods like 2017 where annualized volatility rolling over 3 months is around 5% and super crazy ones like October 2008 where it’s around 75%. Anyone who feels like both these periods feel ‘equally’ risky probably hasn’t been investing very long.
Enter, volatility targeting.
Let’s do an experiment. We say our risk tolerance is equal to SPY’s volatility during the whole time since inception which is 18.9% annualized (from May 1993 to May 2020). We don’t want to go crazy levering up, so we cap our max leverage at 3*. Now, let’s see what happens if we dynamically target our required volatility every day based on the trailing 3 month (63 trading days) volatility**
Just this simple trick, applied on a single ETF, yields around 10-20% better metrics on all measures. Since SPY’s inception, you would’ve almost doubled your money following something like this.
Imagine what would happen if we applied this along with other such concepts across asset classes and used better portfolio construction techniques to boot.
We’ll explore all that and more in the coming weeks and months!
*We’ll use the 3 month treasury as the borrow rate here for leverage values greater than 1.
**Volatility has this nice property of being sticky in the short term, which we’re utilizing here. Also, it doesn’t really matter which parameter exactly you’re using, 1-6 month trailing volatility measures all work similarly.