The authors examine the performance of leveraged and inverse leveraged exchange-traded funds (ETFs) to determine the extent to which leveraged ETFs succeed in their objective of delivering daily returns that are multiples (or inverse multiples) of such common benchmark indexes as the S&P 500 Index. Practitioners seeking to hedge their portfolios, or those seeking to take advantage of the leverage inherent in these products to speculate on the future direction of benchmarks, will find this research helpful.

Leveraged and inverse leveraged ETFs provide a tradable mechanism for achieving exposures that would not otherwise be possible. Practitioners may use leveraged ETFs to profit from their views on the direction of major benchmarks or to hedge portfolio exposures in the short term, but before practitioners can use leveraged ETFs, they must understand the performance characteristics and the causes of deviations from expected results.

The authors provide an important service to practitioners by demonstrating that leveraged ETFs achieve positive abnormal returns and that inverse and inverse leveraged ETFs provide negative abnormal returns. Their research also reveals that transaction costs drive the underperformance of inverse and inverse leveraged ETFs.

The authors study all 98 US equity benchmark–targeted leveraged and inverse leveraged ETFs identified by Morningstar as of 1 July 2012. The sample period is from June 2006 to 29 June 2012. Price and share data are from Datastream.

The authors develop a baseline regression model in which the daily return to a leveraged ETF is a function of the daily return of the benchmark index and the leverage multiple. If each ETF produces daily returns that track the target exposure multiple, then the ETF returns should be entirely explained by this model and its annual performance should reflect the fund’s expense ratio (typically, 90−95 bps). But the authors note that leveraged funds show positive abnormal results, whereas inverse and inverse leveraged funds demonstrate negative abnormal results. For example, for a leveraged multiple of −3, the excess return is −8.27%; for a multiplier of +3, the excess return is +3.24%.

To determine the cause of these results, the authors examine “mirror pairs” of funds with the same sponsor and leveraged multiplier but with the opposite direction of exposure. By extending their baseline regression model to include the effects of transaction costs, the authors conclude that the rebalancing needs created by even moderate levels of index volatility bring transaction costs to a level that explains the underperformance.

Although I appreciate the potential value of leveraged and inverse leveraged ETFs as an addition to the practitioner’s toolkit, it cannot be stressed enough that leveraged ETFs significantly raise the costs of being wrong. As practitioners, we hope to be right more than we are wrong, but we must acknowledge that, sometimes, we will be wrong. Therefore, understanding the downside of using leveraged ETFs is essential.