Synthetic vs. Physical Covered Calls: Tracking Error and the Illiquidity Discount
October. The proxy basket — 60% small-cap industrials ETF, 40% two liquid single-names — rallies 4.2% into month-end. The 30Δ short calls go in the money. The desk rolls up and out, books a $38,000 loss on the call leg, and keeps the premium machine running. The micro-cap, the thing you actually own, finishes the month up 0.4% on volume so thin the spread never tightened below 3%.
A physical covered call on the micro-cap — if listed options existed — would have expired quietly out of the money. You would have kept the full premium and all of a flat stock. The synthetic version did the opposite: it paid out on a rally your stock never participated in. Same trade label, opposite month.
The physical trade you are not doing
A physical covered call is one instrument on one name. Long 100 shares, short one call. Below the strike you keep premium and shares; through the strike your upside caps — but against shares you hold, so the accounting closes. Assignment delivers stock you own. No residual leg.
Price the hypothetical for the $2M micro-cap: a 30Δ call, 30 DTE, on a name with realized vol near 45% would print roughly 3.5–4% of notional per month, $70,000–$80,000. That option does not exist. No listing, no market-maker, no bid. The physical payoff is the benchmark, not a trade.
The synthetic trade you are doing
The synthetic covered call writes the same 30Δ, 30 DTE call — but on the $1.7M proxy basket, sized to match the micro-cap’s dollar-beta ($2M × 1.35, scaled to the basket’s blended β of ~1.59). Collected premium runs ~$64,000/month against $2M of micro-cap notional, ~3.2% per month, compounding to roughly 36.5% APY net of the put-spread hedge cost.
Mechanics match. Economics do not. The short call caps upside on the proxy while your real exposure — the micro-cap — keeps whatever path it takes. When the two legs move together, the synthetic behaves like a physical call with small slippage. When they diverge, it caps gains you never had and leaves losses you actually took uncovered.
Basis: you sold volatility on the wrong name
The core difference is basis — R_microcap − R_proxy — and the short call converts that basis from background noise into realized P&L.
A physical covered call is short volatility on your own inventory. Time decay works for you against a price you hold. A synthetic covered call is short volatility on somebody else’s inventory. You collect implied vol on liquid industrials (ETF leg ~30% IV, single-names 40–45%) while carrying the realized path of an illiquid name that does not share their order flow.
Three consequences follow:
- Upside given away without offset. Proxy rallies 4%+, micro-cap flat. The short call loses $30,000–$40,000 on the liquid leg against no corresponding gain.
- Downside kept without compensation. Micro-cap drops 8% on a guidance cut; proxy flat. You keep $64,000 against a $160,000 loss the hedge ratio assumed would not arrive alone.
- Volatility mismatch. Proxy IV overstates micro-cap systematic vol in calm months (you over-earn) and understates it in stress, when the micro-cap gaps on air while the ETF stays orderly.
Timing: monthly rolls against stale prices
The 30 DTE monthly roll assumes both legs observe the same calendar. They do not, and the proxy side reprices continuously while the micro-cap reprices in lurches — stale closes, 3–5% spreads, prints hours old.
- Roll decisions key off proxy momentum. A late-month proxy rally forces a defensive roll priced at elevated IV — protection against a move your inventory never made.
- Assignment risk is one-sided. Proxy calls assigned on a Friday close deliver liquid shares you flatten Monday. The micro-cap never transacts, so assignment is pure friction with no offsetting release.
- Rebalance lag compounds it. Loadings are re-estimated monthly; between prints the basket notional is wrong by definition. A physical call has no sizing estimate — one call per 100 shares, fixed.
Expect 2–4 months a year where timing mismatch dominates the call leg — one idiosyncratic event, one flow-driven divergence, both months where the roll happens on the proxy’s schedule against the micro-cap’s silence.
Physical vs. synthetic: payoff comparison
Four scenarios, one-month window, $2M micro-cap notional. Physical column assumes the hypothetical listed call at ~3.75% premium. Synthetic column uses the realized $64,000 (~3.2%) on the proxy leg:
| Scenario (1 month) | Physical covered call | Synthetic covered call on proxy |
|---|---|---|
| Both legs +5% | Capped at strike. Keep ~$75K, no residual. | Call loses ~$35K net; micro-cap +$100K retained. Net better than physical, accidentally. |
| Proxy +4%, micro-cap flat | Stock flat, call expires. Keep full ~$75K. | Call loses ~$38K on proxy rally; micro-cap +$8K. Net drag ≈ −$30K vs. physical. |
| Both legs −8% | Keep ~$75K vs. −$160K loss. Cushion only. | Keep $64K vs. −$160K loss. Roughly equivalent, smaller by ~$11K. |
| Micro-cap −10% idiosyncratic, proxy flat | Keep ~$75K vs. −$200K loss. | Keep $64K vs. −$200K loss; put spread pays nothing on a flat basket. Worst month. |
Read the table as the argument: the synthetic matches the physical only when both legs fall together. Everywhere else, tracking error decides whether you over- or under-perform — and the underperformance clusters in the months that hurt.
The 36.5% APY as payment for tracking error
Gross premium of $64,000/month against $2M notional is ~3.2% monthly. Net of the $11,500/month put-spread cost, $52,500/month compounds to ~36.5% APY. That number looks like yield. Treat it as compensation for carrying a 22% idiosyncratic residual and 8–12% annualized tracking error while short volatility on a proxy.
Decompose it:
- ~15–20 points of the APY is straight short-vol carry — the premium a physical call would earn in a normal IV regime.
- The remainder is an illiquidity discount — payment for cap-without-offset months and loss-without-hedge months. Annualized tracking error near 11% implies a monthly 1σ basis band of ±3.2%: ±$64,000 on $2M, the entire monthly premium, as routine noise. At 36.5% the APY clears that band with margin but not comfort.
Below ~20% APY — a sustained low-vol regime compressing 30Δ premium under $40,000/month — the ratio inverts and the tracking error is no longer priced. That is the standing instruction to cut premium-leg notional when sector IV drops under 18.
When the two are close enough, and when they are not
The synthetic is an acceptable stand-in under three conditions: trailing 60-day correlation above 0.70, annualized tracking error under 12%, realized beta within 0.15 of target. Inside those bands the payoff table collapses toward the joint-selloff row and the basis costs roughly what the premium pays.
Outside them — correlation under 0.60 two month-ends running, tracking error above 15% for 40 sessions, a single-month basis print beyond ±6% — the synthetic is not a covered call. It is short volatility on a relationship that stopped holding. Halve call notional at the first breach; stand it down at the second.