Volatility Premium Strategies

1. CSI 1000 Index Option Volatility Selling

Systematic short put + call strategy on CSI 1000 index options, filtered by IV rank

Sell OTM put and call options on the CSI 1000 index when the 20-day IV rank exceeds a threshold. The strategy collects option premium (theta decay) while managing risk through IV-rank filtering, gamma limits, and position sizing. Backtested on 68,554 option observations covering 1,022 contract series from 2022–2025.

Strategy Logic

• Sell OTM puts (Delta ~-0.25) and calls (Delta ~0.25) when IV rank > 50th percentile
• Hold to expiry or roll at 80% profit
• Per-trade gamma ≤ ¥2M, portfolio gamma ≤ ¥200M
• No delta hedging — directional exposure accepted

Figure 1: Strategy NAV curve with CSI 1000 index overlay (2022–2025)

Figure 2: Implied volatility distribution and IV rank time series

Figure 3: Monthly return heatmap

Figure 4: Drawdown profile

Figure 5: Trade-level P&L distribution

Figure 6: Portfolio Greeks evolution (Delta, Gamma, Vega, Theta)

2. HS300 Option Volatility Arbitrage

Delta-hedged short volatility on CSI 300 index options — replication of institutional vol carry strategy

Replication of an institutional-grade volatility carry strategy on HS300 index options. The strategy sells OTM calls at four delta levels (20/25/30/35%), gamma-equal weighted with daily vega target of 0.125 vol, and delta-hedges with IF futures intraday (4× per day). The core alpha comes from the persistent spread between implied and realized volatility (~2.1% on average).

Strategy Construction

• Sell: 4 OTM call strikes (Δ=20%, 25%, 30%, 35%), gamma-equal weighted
• Vega budget: Daily total vega = 0.125 × implied vol
• Contract selection: Nearest month if ≥5 days to expiry, otherwise next month
• Delta hedge: IF futures, rebalanced at 10:00, 11:00, 13:30, 13:55
• Costs: Futures 2bp + Option IV×3% + Management 50bp/yr

P&L Decomposition

• Revenue: Theta − Gamma (IV² − RV² variance premium)
• Risk: Vega (IV spikes during market stress)
• Extras: Futures basis income + intraday trend from hedging

Our Replication (Variance Premium Model)

Pain Point & Solution: Momentum Overlay (Long Gamma Hedge)

The vol carry strategy's main vulnerability is extreme trending days (e.g., Sep 24, 2024: HS300 surged +22%). Short gamma loses heavily when the underlying makes large directional moves. CITIC's solution: overlay an intraday momentum strategy (long gamma) that profits from large moves, creating a natural hedge.

We replicate this with a daily-frequency long-gamma proxy: excess realized variance as momentum PnL. On extreme days, realized variance spikes → momentum profits offset vol carry losses.

Momentum Overlay Results (Our Computation)

Problem Month Fixes

The 0.5× overlay is optimal: Sharpe 2.14 (close to CITIC's 2.29), MaxDD reduced from -11.15% to -6.01%. Remaining gap vs CITIC (-3.52%) is due to daily vs intraday frequency — CITIC hedges 4× per day using tick-level data.

Yearly Returns vs CITIC Reference

Scenario Analysis (Vol Carry Only)

Figure 7: HS300 Vol Arb v2 — (A) NAV comparison: vol only vs +0.5×/+1.0× momentum, (B) Drawdown reduction, (C) Monthly heatmap with momentum, (D) Long-gamma payoff profile, (E) 924 event zoom, (F) Yearly returns vs CITIC. Data: HS300 index + QVIX (AKShare).

Figure 8: Base vol carry replication — NAV, IV vs RV, drawdown, and PnL decomposition

Figure 9: Monthly return heatmap — base vol carry (2020–2024)

3. European Phoenix Autocallable DCA (CAIE ETF)

Structured product NAV-ification — weekly DCA into European Phoenix autocallables on a vol-targeted index

The Calamos CAIE ETF (launched June 2025, SPR "Most Innovative Product" award) packages weekly dollar-cost averaging into European Phoenix autocallable notes linked to the MerQube US Vol Advantage Large Cap Index (MQUSLVA), which tracks S&P 500 futures with a 35% implied-volatility target. The vol-targeting mechanism creates an ideal underlying for structured products: high implied vol (→ high coupons) with controlled downside risk.

Vol Advantage Index (MQUSLVA)

• Underlying: S&P 500 futures
• Target volatility: 35% (weekly rebalancing using implied volatility)
• Deduction: 6% annual interest
• Key innovation: Uses implied vol (forward-looking) vs historical vol (backward-looking)
• Leverage: Low IV → high exposure; High IV → low exposure

Phoenix Autocallable Terms

Why This Works

• Vol targeting eliminates tail risk: When VIX spikes, the index de-levers automatically → underlying can't crash 60% to trigger knock-in
• Backtested knock-in rate: 0.00% (2007–2025)
• Backtested autocall rate: 97.22%
• Higher coupon than alternatives: NDX-linked only 3.2%; Worst-of basket 8.7%; MQUSLVA 15.8%
• Single underlying = simpler + more transparent vs traditional worst-of baskets

Our Replication — Real Computation

We reconstructed MQUSLVA from scratch using SPX (CRSP + OptionMetrics, WRDS) and VIX (CBOE, WRDS), then ran Monte Carlo pricing (200K paths) and a historical DCA simulation.

Stage 1: MQUSLVA Index Reconstruction

Leverage = 35% / VIX, rebalanced weekly. Applied to SPX excess returns with 6% annual deduction. Data: 5,198 trading days (2005–2025).

Stage 2: Phoenix Monte Carlo (200K paths)

Note: Our MC uses GBM with historical MQUSLVA parameters. JPM's higher autocall rate likely reflects their more precise vol-targeting mechanism and the actual MQUSLVA index's lower realized drawdowns.

Stage 3: Historical DCA Simulation (2020–2025)

Key finding: Zero knock-ins on actual historical MQUSLVA paths, confirming JPM's 0% knock-in backtest result. The vol-targeting mechanism successfully prevents the underlying from crashing through the 60% barrier.

Figure 9: Complete Phoenix DCA replication — (A) MQUSLVA reconstruction vs SPX, (B) VIX & leverage dynamics, (C) MC return distribution, (D) DCA portfolio NAV, (E) Notes composition, (F) Cumulative coupons. Data: SPX (CRSP/OptionMetrics WRDS), VIX (CBOE WRDS).

Coupon Comparison (from JPM)