Summary Leveraged ETFs suffer from volatility drag, causing them to underperform their benchmarks, which can be exploited by shorting leveraged ETFs while buying the underlying asset. Backtesting since July 2023 shows a strategy of 67% BTC and 33% short BITX yielding a 34.03% annualized return with lower drawdowns. Adjusting the allocation to 75% BTC and 25% short BITX increases annualized returns to 45.73%, but with tail risk. Introduction Leveraged ETFs always have one problem: they are extremely volatile and unable to produce their desired returns due to volatility drag. In this article, I will introduce a strategy that uses volatility drag to achieve a return of more than 30% at low risk. What is Volatility Drag and how can we profit from it? Volatility Drag means that a leveraged ETF will underperform its benchmark due to volatility in its underlying. Take the following extreme example: Bitcoin falls by 20% in one day and rises by 25% in the next one. For Bitcoin, this would mean that it had a return of 0% within two days. However, a leveraged ETF would lose 40% on day 1 and gain 50% on day two, yielding a return of -10%. The higher the volatility of the underlying, the higher the volatility drag. Figure 1 illustrates this by showing you the performance of 2x Bitcoin Strategy ETF ( BATS: BITX ) and BTC-USD . As you can see, BITX, which was supposed to have twice the performance of BTC, even underperformed its underlying. Furthermore, it shows a much higher standard deviation and maximum drawdown. The thesis of this article is that investors can profit from volatility drag by shorting leveraged ETFs while buying the underlying asset. Ideally, this should be done by including the leverage in the asset allocation process. Since BITX aims to have twice the daily performance, and thereby volatility as Bitcoin, a portfolio of 67% BTC and 33% BITX should be balanced as the 33% in BITX will have the approximately same volatility as the 67% in BTC. I use Bitcoin for this example as the asset is known to be very volatile. In theory, the return of the strategy should go up with the volatility of the underlying. Figure - Performance Metrics of BITX and Bitcoin (Composer ) Backtest Results Since July 2023 (the earliest data point for BITX), the strategy has shown an annualized return of 34.03%. As Figure 2 shows, the strategy also produces positive alpha to a buy-and-hold approach in BTC-USD. Its maximum month-over-month drawdown is 10.39%, which is much lower than that of BTC-USD. While volatility is half that of Bitcoin, the strategy also has a negative beta of 0.34. This means that one could allocate even less to BITX and more to Bitcoin to achieve a 0 beta portfolio. The reason this makes sense is that although BTC should be perfectly hedged by BITX on the first day, asset allocation shifts on day 2 due to the asset's inverse performance. While the 67%, 33% allocation perfectly hedges at the start of using the strategy, a 0 beta portfolio would focus more on making the portfolio neutral on a continuos basis. Figure 2 - Alpha, Beta and Standard Deviation of the Strategy (67% long Bitcoin / 33% short BITX) and Bitcoin (Self calculated in R, Data from Yahoo Finance ) When shifting the allocation from 67% in BTC and 33% in BITX to a 75% / 25% asset allocation, the strategy's annualized return increases to 45.73%. Additionally, the strategy's beta now is at 0.156 (Figure 3) while the alpha deteriorates to only 0.0265, meaning that risk-adjusted performance gets lower. This can also be seen in the standard deviation rising to 23.85%. However, as the maximum month-over-month drawdown is at less than 5%, most of the volatility is likely caused by upward moves (otherwise the drawdown would be higher), suggesting that the strategy's true risk is overstated by the volatility measurement. A factor in favor of the first strategy can be seen when looking at a worst-case scenario of Bitcoin dropping 50% in a day (I know it's unlikely but it still is a possibility). In this case, the first strategy would lose only the gains it made since the last rebalancing due to the strategy's asset allocation shifting over time with BITX underperforming Bitcoin. The second strategy, however, could lose 12.25% more than the gains it made (the amount of losses it has to make until it reaches the asset allocation of 67% BTC and 33% BITX where the daily risk is perfectly hedged), making tail risks a bigger concern. Nevertheless, even a 12.25% drop is very good in comparison to what would happen to Bitcoin and shows the biggest advantage of this strategy. Figure 3 - Alpha, Beta and Standard Deviation of the second Strategy (75% long Bitcoin / 25% short BITX) and Bitcoin (Self calculated in R, Data from Yahoo Finance ) Risks Although I have talked about tail risk already, right now, there are also other, more severe risks for the strategy. The biggest one, in my opinion, is that the asset allocation between BITX and Bitcoin will shift over time, leaving you less hedged in some situations. While this only leaves investors at the potential loss of their current gains (+12.25% for strategy 2) This problem can easily be solved by rebalancing the strategy once every year or even semi-annually to take profits so to speak. Lower rebalancing intervals would likely increase transaction costs, while higher ones endanger the gains already made. In the end, the rebalancing intervals are marginal and should be subject to the trade-off between the two factors already named. The second large risk is that volatility in Bitcoin drops over time. This might be a valid concern as crypto reserves might give bigger stability to the market. Nevertheless, investors can simply swap to other, higher volatility coins in those cases to get the desired return. One should only make sure to pick coins that are liquid enough. Conclusion While volatility drag is terrible for investors who buy leveraged ETFs, one can use it to achieve better risk-adjusted returns than Bitcoin itself does. Overall, both strategies have their advantages. While I prefer the second strategy due to a higher return combined with a lower maximum drawdown, the first strategy has a better hedge against tail events. Nevertheless, both strategies should be suitable for more conservative investors, giving them great returns at very low risk.
