Are investors willing to pay for easy access to leverage? In the April 2020 version of their draft paper entitled “Embedded Leverage”, Andrea Frazzini and Lasse Pedersen investigate the relationship between the leverage of a financial asset (absolute percentage price change per one percent change in the underlying) and its return. They consider equity index options and individual stock options for different maturities and levels of moneyness, and leveraged exchange-traded funds (ETF). They consider three ways to test whether securities with more embedded leverage offer lower (monthly) returns: (1) portfolios sorted on leverage; (2) long-short factors that bet against leverage; and, (3) regression analysis. They consider alphas based on four (market, size, book-to-market, momentum) and five (plus market volatility) risk factors. Using groomed daily data for options on around 3,300 individual stocks and 12 stock indexes during 1996 through 2018, and daily data for seven 2X leveraged ETFs for seven major U.S. stock indexes during 2006 through 2018, they find that:
- Equity (index) options typically exhibit leverage of 2 to 20 (3 to 40) across the range of maturity and moneyness. Options that have shorter maturities and are further out-of-the-money have higher leverage.
- Within each class (index options, individual stock options and leveraged ETFs):
- High-leverage securities (for example, short-maturity options and far out-of-the-money options) generate systematically lower average gross returns and alphas than low-leverage assets.
- Portfolios that are long (short) low-leverage (high-leverage) securities, hedged to be insensitive to movements in underlying assets, generate large average gross returns and Sharpe ratios in excess of one.
In summary, evidence indicates that investors are willing to pay a premium for easy access to leverage, thereby depressing future returns for securities conveying it.
Cautions regarding findings include:
- Reported returns are gross, not net. Portfolios of options, in particular, would be costly to maintain due to turnover from expirations and large bid-ask spreads.
- Leveraging may amplify wildness in return distributions, such that conventional statistics assuming a normal distribution (such as Sharpe ratio) lose meaning as predictors.
