Systematic Cash-Secured Put Writing

Overview

Put writing is the purest expression of the volatility risk premium in equities. You sell a put on the index, you hold cash equal to the notional you might be assigned, and you collect the premium. If the index stays above the strike, you keep everything. If it falls below, you effectively buy the index at the strike, which means you eat the loss below it.

Why does this pay on average? Because investors are structurally willing to overpay for downside protection. Pension funds, insurers and retail investors all want a floor under their equity exposure, and they are price insensitive about getting it. Someone has to be on the other side. That someone gets paid a premium above what the actual distribution of outcomes justifies.

That is a real, persistent, economically explainable edge. It is also, unambiguously, the business of selling insurance. And insurance companies go bankrupt.

Strategy logic

• Short leg: Sell one put per period at a rule-defined strike. Delta 0.20 or a strike 5 percent below spot are common anchors.
• Collateral: Hold cash or short-dated government bills equal to the full notional you could be assigned. This is what "cash secured" means and it is what separates this from a leveraged blow-up.
• Collateral yield: In a positive-rate environment, the interest on the collateral is a meaningful chunk of the total return. Do not forget to include it.
• Roll: At expiry, settle the old put and write a new one. If it expires in the money, take the loss, then write the next one.

The whole strategy is that last sentence. The hard part is not the mechanics. It is having the discipline and the balance sheet to keep writing the next put after a losing month, because the recovery is where the returns come from.

Parameters (knobs)

• Strike distance: At the money puts collect the most premium and lose most often. Deep out of the money puts collect little and lose rarely, but when they lose, they lose enormously relative to the premium collected. The risk to reward is roughly constant. What changes is the shape, not the expected value.
• Tenor: Shorter tenors collect premium more frequently and have faster time decay per day. They also have far more gap sensitivity and cost more in spreads over a year.
• Notional sizing: This is the only knob that really matters. Selling puts on 100 percent of your capital is a fully invested equity-like risk. Selling on 300 percent is a way to go to zero.
• Loss management: Holding to expiry is mechanically simple and avoids the trap of buying back panic-priced options at the worst moment. Rolling down and out delays the loss but can turn a bad month into a bad year.

Where it works and where it hurts

It works in essentially every regime except the ones that matter. Quiet markets, grinding bull markets, mildly choppy markets: the puts expire worthless and you bank the premium. A well-sized put writing program can produce a long, smooth, extremely attractive-looking equity curve.

Then February 2018 happens. Or March 2020. Or 1987. The index gaps down, implied volatility explodes, and the puts you sold for a small premium are suddenly worth many multiples of that. The mark to market loss lands all at once, margin requirements balloon at the same instant, and the people who were leveraged get liquidated at the bottom.

The pattern is the point. Short volatility strategies do not have a normal return distribution with occasional bad luck. They have a distribution that is engineered to produce many small gains and rare enormous losses. A backtest that stops before a crash will look magnificent and tell you nothing.

Backtest design checklist

• Real prices only. Model-priced backtests systematically understate what you actually pay to enter and exit, and they cannot capture the skew that makes out of the money puts expensive.
• Include the collateral yield explicitly. A lot of published put-writing performance is quietly boosted by the T-bill return. Show it as a separate line so you know how much of the edge is actually the option.
• Model margin, not just P&L. Simulate the broker's requirement through a crash. If the requirement exceeds your cash at any point, the strategy would have been force-closed, and the backtest that kept holding is fiction.
• No stop losses in the backtest unless you can justify the fill. Gaps mean the price you wanted is not the price you get. A stop at twice the premium received might fill at ten times it.
• Report the drawdown honestly. The headline should be the worst month and the recovery time, not the Sharpe ratio. Short volatility strategies have flattering Sharpe ratios almost by construction, which is exactly why the Sharpe ratio is the wrong lens.

Common failure modes

• Sizing by premium instead of notional. "I only collected 50 dollars, how much can I lose?" The answer is the full distance to zero on the notional. This misunderstanding has ended more accounts than any other in options.
• Leverage creep. The strategy works so reliably for so long that it seems safe to do more of it. Size grows quietly. Then the tail arrives at the new size.
• Capitulating at the worst moment. Buying back puts at the bottom of a crash locks in the loss and means you are not there for the volatility crush afterwards, which is when the strategy earns much of its money back.
• Confusing low volatility with low risk. The reported standard deviation is low. The actual risk is a rare enormous loss that standard deviation does not describe.

Our notes and suggestions

If you run this, run it small, run it cash-secured, and pair it with an explicit plan for the day it goes wrong. Many practitioners spend a fraction of the premium on far out of the money puts as a tail hedge, which gives up some of the return in exchange for surviving the event that would otherwise end the program. That trade-off is almost always worth it.

Be very honest with yourself about the shape of this return stream. It can produce five great years and then give all of it back in two weeks. If you cannot hold through that, or if your capital cannot, you should not be selling puts at all. What would change our mind about running it: the implied to realized volatility spread going persistently negative, or a margin regime where you cannot survive a 20 percent overnight gap.