ETF Options Research
Selling the volatility risk premium on ETFs, measured two ways on 35 years of data, with the one number the data do not contain made explicit.
Computing the tables (first load after a deploy can take half a minute)…
1. The premium is real: the variance-swap upper bound
A CBOE 30-day vol index (VIX for SPY, VXN for QQQ, GVZ for GLD, OVX for USO, and so on) is by construction the fair strike of a 30-day variance swap. Selling that swap each month and settling against realised variance needs no model, no hedging and no assumption about which option you sold. It is the cleanest measure of the premium, and the best any option seller can hope for.
Read the skew and the worst month before the Sharpe. At a 20% risk target the single worst month wipes out more than a year of premium, and it arrives in the one month everything else is also losing. Retail accounts cannot trade variance swaps; this row is the ceiling, not the strategy.
2. What you can trade: a hedged short straddle, and the haircut that decides it
Sell the 30-day at-the-money straddle, hedge the delta at every close with the ETF, settle at expiry, pay a bid-ask on the straddle and a cost on every hedge share. The straddle is priced by Black-Scholes at the index vol minus a haircut, because an index is a variance-swap rate that sits above the at-the-money implied vol a straddle actually fetches, by an amount that depends on the skew. The free data contain the index. They do not contain the at-the-money vol.
The whole verdict sits inside two vol points. One after-hours quote on the paper account showed SPY's at-the-money straddle 3.7 points below the VIX and USO's 7.9 points below OVX; if those held during market hours, the strategy would earn nothing after costs. That is why the desk captures the gap every afternoon instead of assuming it.
3. The gap, measured so far
No market-hours capture yet. The daily capture runs at 15:30 ET; this table fills in from the first one.
Until this table has months of history per ETF, no expected return is quoted for the strategy anywhere on this site.
What this research cannot tell you
- • The at-the-money haircut. Everything above is a sensitivity table until the daily capture has measured it per ETF.
- • Assignment, early exercise and pin risk at expiry: the simulation settles at intrinsic, the real book is rolled before the close on expiry day.
- • Margin. A short straddle book at this risk target needs a portfolio-margin account; a Reg-T account will be stopped by buying power long before the risk limits bite.
- • Before 2016 the realised side is the matching futures at the 16:00 ET bar, which correlates 0.87 to 0.98 with the ETF but is not the ETF.