← All Lessons

The Variance Risk Premium

Implied vol exceeds realized vol on average. That gap is the variance risk premium, and it is the structural edge that pays short-vol traders for taking on the risk that buyers want to hedge.

18 min read · Lesson 11 of 24

The empirical fact

If you compute the IV of SPY 30-day options today, and compare it to the RV that SPY actually delivers over the next 30 days, on average the IV will be higher. Across decades of data, on hundreds of underlyings, the same pattern holds. Implied volatility runs structurally above realized volatility.

The gap is not a small one. For SPY, the long-run average IV is roughly 16-17%, while the long-run average 30-day RV is roughly 13-14%. The wedge of 3-4 vol points is the variance risk premium (VRP).

Fig. 11.1 Distribution of (IV − RV) for SPY 30-day options. The center of the distribution sits clearly to the right of zero. Negative observations exist (IV undershoots) but they are the minority.

Why the gap exists

Two main forces produce the VRP. Both are about who wants what.

First, insurance demand. Investors who hold long stock want to hedge tail risk. They buy puts. They are willing to overpay for that insurance, the way people overpay for car insurance or home insurance. The expected payout is less than the premium, but the buyer values the protection more than the actuarial cost.

Second, risk aversion among option sellers. The other side of the trade is the option seller, who takes on a position with bounded gain and unbounded loss (in the case of a short call) or bounded gain and large bounded loss (in the case of a short put). Sellers require compensation for taking on this asymmetric risk. They charge a premium above the fair statistical price.

These two forces compound. The result is a persistent wedge between what options cost and what the underlying actually delivers.

VRP is a return, not a free lunch

It would be tempting to read the VRP as a guaranteed profit for option sellers. It is not. The premium exists precisely because short volatility is a risky position. On average it pays. In the worst weeks, it pays brutally negative.

Selling vol is structurally similar to writing insurance. You collect steady small premiums for years, and then a hurricane hits and you pay out far more than you ever collected on a single event. Long-term, the math works. Short-term, the path can wipe you out.

The variance risk premium is real. It pays you for being there in the bad weeks, not in the good ones.

How the VRP is captured

Several approaches exist to harvest the VRP. They differ in mechanics but share the same underlying logic: be short premium when the premium is rich, manage the tail risk on the way.

Short straddles or strangles. Sell ATM or OTM options on both sides. Maximally exposed to spot but pure short-vol exposure.
Iron condors. Short strangle with long wings. Capped tail risk in exchange for lower premium captured.
Variance swaps. Direct OTC instrument that pays the spread between IV and RV. Used by institutional players.
VIX futures roll. Sell front-month VIX futures and roll them as they expire. Captures the term-structure premium that often coexists with the VRP.

When the VRP fails

The VRP is an average. The distribution has a left tail. In specific weeks (October 1987, August 2015, February 2018, March 2020), realized volatility blew through implied volatility by orders of magnitude. Short-premium positions held through those weeks suffered enormous losses, sometimes 100% of capital.

This means VRP harvesting is not a buy-and-hold strategy. It requires regime detection and risk management. Periods when the VRP is negative (IV cheap relative to RV) are usually the periods immediately preceding a vol spike. Selling premium in those periods is approximately the worst possible time to do it.

The Delta S framing

The VRP is what makes our entire research program coherent. We build systematic signals to detect when the premium is rich enough to harvest and when the regime suggests we should not be in the trade. The Z-score on (IV − RV) is the entry signal. Event dates and structural breaks define the exits. The VRP itself is the source of return; everything else is risk management.

Cross-sectional VRP

The VRP is not uniform across underlyings. SPY and QQQ tend to have richer premiums than individual stocks. Indices that experience occasional crashes carry more premium than indices that move smoothly. Single-name VRP can be dominated by event risk (earnings, M&A) rather than ambient demand for hedging.

This creates opportunities and traps. Some single-name options carry inflated premiums driven by retail speculation rather than rational hedging demand. The premium can stay rich indefinitely if the underlying is meme-driven. Others are persistently underpriced because no one is hedging them.

Term-structure VRP

The VRP varies across expiries. 30-day VRP is the most studied and most consistently positive. Very-short-dated options (0-3 days) sometimes carry negative VRP because the realized volatility includes specific intraday moves that the IV did not anticipate. Longer-dated options (90+ days) carry stable but smaller VRP per unit of time.

The richest part of the curve, on a per-day basis, is usually 30-45 days. This is where systematic short-premium programs concentrate.

What you carry forward

Implied vol exceeds realized vol on average. The gap is the variance risk premium.
The VRP exists because of insurance demand and risk aversion among sellers.
It is a real return, not a free lunch. The path includes severe drawdowns.
Capturing the VRP requires regime detection and risk management.
The richest VRP per unit of time tends to live in 30-45 day options on liquid indices.