Buying on Margin: Why the Call Comes at 28%, Not 50%

American investors were carrying $1.502 trillion of margin debt at the end of June 2026, up from $1.226 trillion in December 2025.^[5] That is a 22.6% increase in six months, which tells you roughly what mood the market has been in. It also means a lot of people are about to learn something about leverage that they think they already know.

Here is the thing almost everyone gets wrong. Ask a reasonably informed investor what happens when you buy stock with 2x leverage and the stock falls, and they will tell you that a 50% decline wipes you out. That is arithmetically true and practically irrelevant. Your broker does not wait for your equity to reach zero. Under a typical house maintenance requirement of 30%, the call arrives at a 28.6% decline. Under 35%, at 23.1%.

A 23% decline is not a crash. It is a bad Tuesday in a bad quarter. I ran these numbers the first time after watching a friend get liquidated in a drawdown that, in hindsight, fully retraced within four months. He was right about the stock. He was wrong about the loan.

What buying on margin actually is

A margin account is a loan from your broker, secured by the securities sitting in that same account. You put up cash, the broker lends against it, and you buy more stock than your cash alone would cover. The Federal Reserve Board's Regulation T governs how much a broker-dealer may lend at the moment of purchase. Section 220.12(a) of the regulation sets the required margin for a margin equity security at “50 percent of the current market value of the security or the percentage set by the regulatory authority where the trade occurs, whichever is greater.”^[1] In plain terms: half down, half borrowed, at most.

So $50,000 of your money buys $100,000 of stock. You own $100,000 of collateral and you owe $50,000. Your equity is the difference.

Now notice the asymmetry, because it is the entire article. The collateral floats. The loan does not. When the stock falls 20%, the collateral is worth $80,000 and the loan is still exactly $50,000, so your equity has gone from $50,000 to $30,000. A 20% move in the asset produced a 40% move in your money. That is what leverage means, and that part everybody understands.

What matters more is that the ratio your broker watches, equity divided by market value, is falling twice as fast as the stock. That ratio is the one with a tripwire under it.

Two numbers do all the work: 50 and 25

Regulation T handles the front door. FINRA Rule 4210 handles what happens afterward. Rule 4210(c) requires customer equity of at least “25 percent of the current market value of all margin securities” carried long in the account.^[2] Fall below that and you are in a maintenance call.

Except you almost certainly will not be measured against 25%. Rule 4210(d)(1) tells member firms to formulate their own margin requirements and to review the need for higher ones on individual securities and individual accounts.^[2] FINRA is blunt about what firms then do: a firm “might set the maintenance margin at 30 or even 40 percent of the current market value of the securities in your account,” and “firms may increase the house requirements at any time and aren't required to provide you with advanced written notice.”^[4] The SEC's own investor bulletin gives the same range, describing house requirements as “typically between 30 to 40 percent, and sometimes higher depending on the type of stock purchased.”^[3]

Read that second quote again with a drawdown in mind. The requirement can be raised while your position is already falling, with no notice, precisely when volatility is high, because high volatility is one of the conditions that prompts firms to raise it. The number you underwrote your position against is not a constant. It is a variable the counterparty controls and revises in the direction that hurts you.

The formula, and the numbers it spits out

This is a two-line derivation and it is worth doing once by hand, because after you have done it the whole risk profile of a margin position becomes something you can compute in your head.

Let V be the current market value of the position and L the loan, which is fixed. Equity is V minus L. A maintenance call fires when equity divided by V drops below the maintenance rate m. Solve for V and you get the trigger price: V below L divided by (1 minus m). Express that as a percentage decline from where you bought, and you get a closed form that depends on nothing but your starting equity fraction and the maintenance rate.

def call_decline(initial_equity_fraction: float, maintenance: float) -> float:
    """Percentage decline from cost basis that triggers a maintenance call."""
    return 1 - (1 - initial_equity_fraction) / (1 - maintenance)


for m in (0.25, 0.30, 0.35, 0.40):
    print(f"maintenance {m:.0%} -> call at {call_decline(0.50, m):.1%} decline")

# maintenance 25% -> call at 33.3% decline
# maintenance 30% -> call at 28.6% decline
# maintenance 35% -> call at 23.1% decline
# maintenance 40% -> call at 16.7% decline

Sanity-check it against the S&P 500. A 23% peak-to-trough drawdown in the index is not rare. Individual stocks do that in a week on an earnings miss. If you want the fuller picture of how ordinary these moves are, the piece on what volatility actually measures is the companion to this one, because volatility is the input that decides how often you visit the tripwire.

The call is not a phone call

The phrase “margin call” carries a cinematic connotation: a ringing phone, a broker with bad news, a deadline. Drop it. FINRA says firms “don't have to issue a margin call before selling securities in your margin account to meet a margin call and may sell enough securities to completely pay off your margin loan, not just meet the margin call.”^[4] The SEC puts it the same way: even if your firm offers you time, “it can sell your securities without waiting for you to meet the margin call.”^[3] An extension is possible in exceptional circumstances, and firms “aren't required to do so.”^[4]

The part nobody tells you: curing a call costs a multiple of the call

Say you bought $100,000 with $50,000 down under a 30% house requirement. The stock falls 35%. Your position is worth $65,000, the loan is still $50,000, and your equity is $15,000. The requirement is 30% of $65,000, which is $19,500. You are short $4,500.

Deposit cash and you write a check for $4,500. Fine. But most people in a drawdown do not have idle cash, so they sell. And here the arithmetic turns nasty, because selling stock reduces the collateral base at the same time it reduces the loan. To restore compliance you have to liquidate the deficiency divided by the maintenance rate: $4,500 divided by 0.30, which is $15,000.

Roughly $3.33 of stock sold for every $1 of the call. At a 25% requirement it is $4 per dollar. That is the mechanism that turns an ordinary correction into a cascade. Forced sellers do not sell the shortfall, they sell a multiple of it, and every one of them is doing it into the same tape at the same time. If you want to see what happens when that dynamic runs in reverse against people who were short, the GameStop rally and its margin mechanics is the case study.

Interest is boring, charged daily, and priced backwards

Margin interest almost never causes the blowup, which is exactly why it gets ignored. It is charged daily on settled debit balances, accruing from the day the credit is extended, and billed on a tiered schedule. Fidelity's published rates, effective December 12, 2025, run from 11.825% on balances of $0 to $24,999 down to 10.075% on balances of $250,000 to $499,999, against a 10.575% base rate.^[7]

Look at the direction of that schedule. The more you borrow, the less you pay per dollar. Credit-risk pricing runs the other way everywhere else in finance, and it runs the other way here too in the sense that the collateral is liquid and marked continuously, so the lender is genuinely safer. It is still a little funny that the pricing signal points at “borrow more.”

On $50,000 at 10.375%, you are paying about $5,190 a year for the privilege. Your position needs to clear that hurdle before it has made you a dollar, which is the same hurdle that quietly eats the returns most active traders report. And because the interest is charged to the account rather than billed separately, an unpaid month adds to the debit balance, which nudges your call threshold slightly closer with no help from the market at all.

Margin versus a portfolio line of credit versus options

Margin is one of three ways to get leverage or liquidity out of a portfolio, and they fail differently.

A securities-backed line of creditis the private-bank cousin. FINRA describes a typical SBLOC as permitting you to borrow “from 50 to 95 percent of the value of the assets in your investment account,” with a hard catch: SBLOCs are non-purpose loans, so “you can't use the proceeds to purchase or trade securities.”^[6] Higher advance rates, cheaper money, broader use for a kitchen renovation or a tax bill. The catch is that they are demand loans, meaning “lenders may call the loan at any time,” and if you cannot post collateral the firm may sell your securities.^[6] Different paperwork, same ending.

Options are the structurally different one. A long call gives you leveraged upside with a maximum loss equal to the premium you paid, and no maintenance requirement, because there is no loan. You cannot get a call on a long call. What you pay for that is time: the position expires, and it decays whether or not you were right. The tradeoffs are laid out in the piece on how calls and puts actually differ. The honest framing is that margin gives you unlimited duration and a forced-exit risk, while a long option gives you a bounded loss and a deadline. Pick which failure you would rather own.

What to actually do with this

Go find your broker's house maintenance requirement. Not FINRA's 25%, not the number in a textbook, the one in your agreement, for the specific securities you hold. Then run the formula. You will get a single number: the percentage decline that ends your ability to hold the position. Write it down next to the ticker.

Most people who use margin have never computed that number, which means they are holding a position whose exit condition they cannot state. And it is not a fixed number, since the broker can move it, so treat it as a ceiling rather than a promise. That is also the argument for not concentrating a levered bet in one name, which is the same argument diversification makes in the unlevered case, only with a hard deadline attached.

Leverage is not the problem. Being wrong about the trigger is. Do the division before you borrow the money, not after your broker has already sold the stock for you.