The volatility risk premium
Why implied volatility persistently exceeds realized volatility, what selling that gap actually earns, and how professionals size a strategy whose losses arrive all at once.
By PB, options traderPublished 4 min read
Strip away the strategies, the Greeks, and the acronyms, and professional options trading keeps returning to one empirical fact: options, on average, cost more than they end up being worth. The implied volatility baked into prices has historically exceeded the volatility subsequently realized — not always, not everywhere, but persistently enough across decades and markets to constitute one of the best-documented regularities in finance. The gap is the volatility risk premium, and this lesson is about treating it as a professional would: as a real source of return with a real, concentrated cost.
The evidence, briefly and honestly
The cleanest measurements come from variance swaps and option strips. Carr and Wu’s study of synthetic variance swap rates found that for major US equity indexes, the variance priced into options systematically exceeded the variance that materialized — a seller of index variance was paid, on average, over their 1996–2003 sample, and the finding has extended robustly since. Bakshi and Kapadia reached the same conclusion from another angle: delta-hedged long option positions on the index — positions that isolate the volatility exposure — lost money on average, which is only possible if buyers were overpaying for volatility.
Strategy indexes tell the practitioner’s version. Cboe’s PutWrite index (PUT), which mechanically sells at-the-money S&P 500 puts against cash collateral month after month, has compounded over multiple decades at equity-like rates with lower volatility — and with drawdowns that arrive suddenly, in crises, when equities are also falling.
Two honesty clauses belong next to the evidence. First, the premium is measured on average: individual months are frequently negative, and single episodes (1987, 2008, the February 2018 volatility event, March 2020) erased years of accumulated carry for poorly structured sellers. Second, the premium is largest and most reliable in equity index options; in single names it is smaller, noisier, and sometimes absent.
Why the premium exists — and persists
A durable premium needs an economic engine, not just a backtest. The VRP has three reinforcing ones.
Insurance demand. Institutions holding trillions in equities have structural reasons to buy downside protection — mandates, regulation, career risk. Persistent hedging demand meets a limited pool of willing insurers, and the price of insurance settles above its actuarial cost. This is also why the premium concentrates in index puts and expresses itself as skew.
The character of the risk. The seller’s losses are not random noise; they cluster in market crashes, when the seller’s other assets, income, and access to capital are simultaneously impaired. Compensation for a risk that is correlated with bad times must exceed compensation for an equal-sized diversifiable risk. This is the same logic that prices the equity risk premium, applied one derivative higher.
Behavioral reinforcement. Lottery-like payoff preferences support demand for cheap-looking out-of-the-money options, and loss aversion supports overpaying for protection. These tilts are stable features of human decision-making, not temporary mispricings awaiting correction.
None of these engines is going away, which is why the premium survives being published. But note what the engines imply: the VRP is payment for a service — bearing crash risk. Anyone collecting it is running an insurance business, whether or not they think of it that way.
What collecting it actually looks like
The return stream of systematic premium selling has a signature shape: many small gains, occasional large losses, negative skewness at the strategy level. A short strangle program on an index might win in seventy-plus percent of months; the losing months decide whether the strategy has positive expectancy at all.
The instrument matters less than the exposure. Covered calls, cash-secured puts, credit spreads, iron condors, short straddles — all are short volatility, differing in how much tail they leave open and how much premium they surrender for protection. A defined-risk structure like a condor gives up carry relative to a naked strangle; that surrendered carry is the price of a survivable worst case:
Contract multiplier: 100 (US equity options)
- Net credit
- $200.00
- Max profit
- $200.00
- Max loss
- −$300.00
- Breakevens
- 93 / 107
| Leg | Side | Type | Strike | Premium | Qty |
|---|---|---|---|---|---|
| 1 | long | put | 90 | 0.8 | 1 |
| 2 | short | put | 95 | 1.8 | 1 |
| 3 | short | call | 105 | 1.8 | 1 |
| 4 | long | call | 110 | 0.8 | 1 |
Regime matters, too. The premium compresses when volatility is already low and everyone is selling it, and it inverts around stress: when the term structure flips into backwardation, realized volatility frequently exceeds implied for a stretch. Mechanical selling through every regime is a choice to be paid less exactly when risk is highest.
Sizing: the whole game
Because losses are rare, large, and clustered, position sizing does more work in a VRP strategy than in any directional one. The professional rules are unglamorous:
Size to the scenario, not the average. Before entry, price the position’s loss in a 2008- or 2020-shaped month — spot down hard, implied volatility doubling, margin requirements expanding — and cap total short-premium exposure so that scenario is an acceptable drawdown, not a terminal one. Keep the denominator honest: returns are earned on capital reserved, including the buying power the position consumes at its worst point, not on the premium collected. And prefer structures whose maximum loss is a contract term rather than a market outcome; the certainty is worth carry.
The one-sentence summary of two decades of research and several famous blowups: the volatility risk premium is real, and it is exactly as large as the discomfort of the person collecting it. Strategies that promise the premium without the discomfort are removing the reason it exists.
FAQ
- If the VRP is so well documented, why hasn't it been arbitraged away?
- Because collecting it requires bearing losses concentrated in exactly the states of the world where capital is scarcest. That is not a market inefficiency — it is payment for insurance. The premium compensates a risk that cannot be diversified away, so it persists the way equity and credit risk premia persist.
- Is selling options therefore a good strategy for individual traders?
- It can be a legitimate component, sized honestly. The failure mode is treating the win rate as the edge: strategies that win 90% of months can still have mediocre or negative expectancy after their losing months. Structure (defined risk), sizing, and regime awareness matter more than entry signals.
- Does the premium exist outside US equity indexes?
- Yes — research documents variance risk premia in international equity indexes, rates, currencies, and commodities, with varying sizes and reliability. Equity index options remain the deepest and most studied market for it.