Paper: arXiv 2301.03186

Authors: Hayden Brown

Abstract

Daily leveraged exchange traded funds amplify gains and losses of their underlying benchmark indexes on a daily basis. The result of going long in a daily leveraged ETF for more than one day is less clear. Here, bounds are given for the log-returns of a leveraged ETF when going long for more than just one day. The bounds are quadratic in the daily log-returns of the underlying benchmark index, and they are used to find sufficient conditions for outperformance and underperformance of a leveraged ETF in relation to its underlying benchmark index. Results show that if the underlying benchmark index drops 10+% over the course of 63 consecutive trading days, and the standard deviation of the benchmark index’s daily log-returns is no more than .015, then going long in a -3x leveraged ETF during that period gives a log-return of at least 1.5 times the log-return of a short position in the underlying benchmark index. Results also show promise for a 2x daily leveraged S&P 500 ETF. If the average annual log-return of the S&P 500 index continues to be at least .0658, as it has been in the past, and the standard deviation of daily S&P 500 log-returns is under .0125, then a 2x daily leveraged S&P 500 ETF will perform at least as well as the S&P 500 index in the long-run.

Complexity vs Empirical Score

  • Math Complexity: 5.5/10
  • Empirical Rigor: 4.5/10
  • Quadrant: Lab Rats — theoretically deep, empirically untested

Why this score: The paper provides useful mathematical bounds on the long-term performance and volatility decay of daily leveraged ETFs using quadratic approximations. While mathematically sound and clearly presented, the empirical validation relies on basic historical assumptions (e.g., historical standard deviation and mean thresholds) rather than rigorous backtesting across market regimes.

Research Flowchart

  flowchart TD
    A["Research Goal: Estimate long-term returns & performance bounds of daily leveraged ETFs"] --> B["Inputs: Daily log-returns, volatility (std dev), & leverage multiples (-3x, 2x)"]
    B --> C["Derive Quadratic Bounds for leveraged ETF log-returns over multi-day horizons"]
    C --> D["Analyze Conditions for Outperformance & Underperformance"]
    D --> E["Finding 1: -3x ETF yields ≥1.5x short index return if index drops 10%+ over 63 days (std dev ≤ 0.015)"]
    D --> F["Finding 2: 2x S&P 500 ETF outperforms index long-term if annual return ≥ 0.0658 & std dev < 0.0125"]