August 30, 2026
The obligation is identical to a cash-secured put. Only the collateral changes, and that single difference is what separates a strategy whose worst case is owning a stock from one whose worst case is owing money.
Selling a $50 put obliges you to buy 100 shares at $50 if you're assigned. That is true whether or not you set aside the $5,000.
Cash-secured means the money is there, reserved, doing nothing. Uncovered — sold on margin, often called naked — means the broker holds a fraction as collateral and extends credit against the rest. The promise is unchanged. Only the funding is.
So the honest description isn't that margin puts are a riskier strategy. It's the same strategy, with a loan attached and the option of doing more of it.
Requirements vary by broker, but the common shape takes the greater of two calculations: something like 20% of the underlying's value minus the amount the option is out of the money, plus the premium; or a smaller flat percentage of the strike, plus the premium. There's a floor per contract underneath both.
A $50 put on a $52 stock might tie up somewhere around $800 to $1,000 rather than $5,000. Six to seven times the position for the same collateral.
If the stock falls to $44, the requirement recalculates on a lower stock price with the option now in the money, and it rises — possibly toward the full strike amount as it goes deeper.
Read that second line again, because it's the mechanism that does the damage. As the position moves against you, the collateral demanded goes up at the same time as the account's equity goes down. Two forces, same direction, same moment. A position that consumed a fifth of your buying power at entry can consume all of it after a move that isn't especially unusual.
When equity falls below the requirement, you post cash or the broker closes positions. Some brokers call first; many reserve the right not to, and the agreement you signed almost certainly says they can act without notice.
They choose what to close, when, and at what price. It happens during a decline, in the widened markets described in the spread piece, and typically in whatever order is most convenient for the firm rather than most favourable for you. Being liquidated is not selling at a bad price. It's someone else selling at a bad price on your behalf.
Brokers also raise house requirements unilaterally, and they do it on exactly the names and at exactly the times you'd least want them to. A position that was compliant on Friday can be non-compliant on Monday without the stock having moved at all.
A cash-secured put that goes badly leaves you owning a stock worth less than you paid. Unpleasant, bounded by the strike, and repairable by time if the company survives. The money was already spent the moment you set it aside.
An uncovered put that goes badly can leave you with a debit balance: assigned into shares you can't pay for, liquidated at whatever the market offered, and owing the difference. The loss is not bounded by what was in the account.
That's a different category of outcome, and it's the reason the tier system described in the broker approval piece puts uncovered writing at the top. Not because it's sophisticated. Because the loss has no floor at your account balance.
Leverage by itself is not the problem. One margin put on a $50 stock in a $60,000 account is a modest position however it's collateralized.
The problem is that buying power is a number, the number is now large, and it invites use. The seller who would have sold one cash-secured put sells six, because six is what fits, and six is not the same trade repeated — it's a single concentrated position that will be assigned all at once, in the same session, on the same news.
This is exactly why the sizing piece argues for sizing to the obligation rather than the buying power. On margin those two numbers diverge by a factor of five or six, and only one of them describes what happens if you're assigned.
Nearly every account destroyed by option selling was destroyed this way. Not by a clever strategy going wrong, but by an ordinary strategy sized against collateral instead of against the promise.
Larger accounts can qualify for a risk-based margining system that models the whole book rather than each position, and it generally produces lower requirements for genuinely hedged positions. Minimum equity thresholds are substantial, typically six figures.
It's a real improvement for someone running a diversified book with offsetting exposures. Applied to a concentrated set of short puts it amplifies the same failure, since lower requirements mean more contracts fit, and the model's assumptions about correlation are least reliable in the conditions where you need them most.
There's a legitimate version, and it has a specific shape: someone with a large, genuinely diversified book who would accept every assignment, who has the cash or the credit to take all of them simultaneously, and who is using margin for efficiency rather than to hold more positions than the account could fund.
The test is a single question, and it's the same one from when not to sell premium, applied to the whole book at once. If every short put I hold were assigned this Friday, could I pay for all of it?
If the answer is yes, margin is doing nothing but freeing up idle cash, which is a reasonable thing for it to do. If the answer is no, you aren't selling puts efficiently. You're short more stock than you can buy, and the market decides when that becomes a problem.