Circuit Breakers: The Exact Percentage Drops That Halt the Whole Market

It’s 2:10 on a Thursday afternoon and the S&P 500 is down 6.4%. Your broker’s app is slow. Somebody on TV says the market is “about to hit a circuit breaker,” and you have no idea whether that means your positions get protected, frozen, or liquidated.

Here is what actually happens. Three thresholds stop the entire US equity market: a 7% decline, a 13% decline, and a 20% decline, each measured against the S&P 500’s official closing value from the prior session. The first two freeze all listed stocks and ETFs for fifteen minutes and can each fire only once a day. The third ends the session — no reopening, no auction, come back tomorrow. And there is a window near the close where two of those three thresholds simply stop existing.

What gets measured, and against what

The reference is yesterday’s official S&P 500 close. Not the futures print, not this morning’s open, not the intraday high. Exchanges compute the three levels before the bell, round to the nearest index point, and publish them, which means you can do the arithmetic yourself the night before with three multipliers: 0.93, 0.87, 0.80.

Say the index closed at 6,000. Level 1 sits at 5,580, a 420-point fall. Level 2 sits at 5,220, or 780 points. Level 3 sits at 4,800, a 1,200-point drop. Those point distances scale with the index, which is why “a 500-point drop halts the market” is a sentence with no meaning attached to it. At an index of 3,000, Level 1 is 210 points away. At 8,000, it’s 560. Same rule, and the point figure nearly triples.

Round numbers make this easy to check. Multiply your prior close by 0.93 and you have the level that matters most, because Level 1 is the one a bad day actually reaches.

The window where the rules quietly change

Between 9:30 a.m. and 3:25 p.m. Eastern, all three levels are live. After 3:25, Levels 1 and 2 switch off. For the last thirty-five minutes of the session, the S&P 500 can fall 12.9% and nothing interrupts it. The only backstop left is the 20% full-day close, and it stays available right up to the bell.

The logic is not that regulators stopped caring at 3:25. A halt near the close does more damage than the decline it interrupts: freeze trading at 3:40 and you strand every closing-auction order, every index fund’s end-of-day rebalance, and every option about to expire that afternoon. The rulemakers took a rougher final half hour in exchange for a functioning close.

What a 7% index drop does to your account

It does not take 7% off your balance, and the gap between those two numbers is where people get hurt.

Take a concrete position. You hold $80,000 of stock in a margin account, funded with $40,000 of your own money and $40,000 borrowed at the standard Reg T 50% initial requirement described on the SEC’s investor page on margin. Your holdings run a beta of roughly 1.3 — growth-tilted, nothing exotic. Prior close was 6,000.

The index touches 5,580 at 11:15 a.m. and everything stops. Your positions, at beta 1.3, are down about 9.1%. That’s $7,280 off an $80,000 book. Your equity — the part that’s yours — has gone from $40,000 to $32,720. You are down 18.2% while the headline says seven.

Trading resumes at 11:30, the selling continues, and by 1:40 the index hits 5,220. Down 13%. Your book is off 16.9%, or $13,520, and your equity is $26,480. That’s a 33.8% loss of your capital on a day the index lost thirteen. FINRA’s overview of investment risk makes the general point that leverage cuts both ways; the arithmetic here is what that sentence costs.

Now the worst case. The index reaches 4,800 at 3:12 p.m. and the market closes for the day. Your positions are down 26%, a $20,800 loss, leaving $19,200 of equity against $40,000 this morning. You’ve lost 52% of your capital, the market is shut, and there is nothing to do until 9:30 tomorrow.

Does that trigger a margin call?

Almost, and the “almost” is worth understanding because it’s narrower than it looks.

Most brokers set house maintenance margin near 30% of market value. FINRA’s floor is 25%, and brokers routinely go higher. With an initial equity fraction e and a maintenance requirement m, a position can fall by 1 − (1 − e)/(1 − m) before equity touches the line. At Reg T’s 50% and a 30% house requirement, that’s 1 − (0.50/0.70) = 28.6%.

Your position fell 26% on the Level 3 day. You cleared the line with two and a half points to spare. Check it directly: your holdings are worth $59,200 and your equity is $19,200, a ratio of 32.4% against a 30% requirement. No call — this time.

Divide the two numbers and you get the beta that would have changed the outcome: 28.6% ÷ 20% = 1.43. A portfolio at beta 1.0 sails through a Level 3 day at Reg T minimum equity. Beta 1.5 gets the call. You were at 1.3, which is why you finished the day merely devastated rather than forcibly liquidated. And if the next morning opens down another 2%, your position crosses 28.6% cumulative and the call arrives before you’ve had coffee.

How the thresholds got to 7/13/20

The current framework dates to 2013 and replaced a system built on Dow Jones point values, which is why so much stale advice survives in comment sections.

Two changes did the real work. The reference index moved from the Dow — thirty stocks, price-weighted — to the broad, cap-weighted S&P 500. And the thresholds came down hard: the first halt used to need a 10% Dow decline and now needs 7% on the S&P. A morning that would have traded straight through in 2005 gets paused today.

The point-based era explains the folklore. On a 1997 Dow near 7,900, a 350-point drop was 4.4% — a genuinely bad day. On a Dow near 45,000, 350 points is 0.8%, a lunch break. The fixed-point rule decayed into nonsense as the index compounded, which is exactly why it was scrapped.

Single stocks halt constantly, under different rules

Market-wide breakers are rare. Single-stock halts happen every week, and they run on a separate mechanism called Limit Up-Limit Down, keyed to each security’s own rolling five-minute average price rather than to yesterday’s close.

When a stock would print outside its band, quotes enter a limit state for fifteen seconds. If nothing resolves it, the security pauses for five minutes. On a turbulent day these fire dozens of times across the tape while the S&P 500 moves less than a percent. Nasdaq publishes the live halt list, and it’s worth opening once on an ugly morning purely to see how routine these are. Volatility pauses share that feed with news-pending halts, regulatory halts, and IPO-related halts, each carrying its own reason code.

Two things almost everyone gets wrong

“A halt means the market fell 7% from its high.” It’s 7% from the prior close, and those diverge fast. Return to the 6,000 close with Level 1 at 5,580. If the market gaps down and peaks at 5,940, the trigger is only 6.06% below the session high. If it gaps up to 6,120 before reversing, the fall from the high needs to be 8.82% — (6,120 − 5,580) ÷ 6,120. Identical rule, and a day that opens strong has to drop nearly nine percent from its peak to earn a pause.

“A halt protects me.” It suspends trading, which is a different thing. Suppose you bought a stock at $100 and set a stop at $95. The market breaks through Level 1, your stop converts to a market order, and in a fast tape the next available bid is $91.40. You lose $8.60 a share against the $5.00 you planned around — a 72% overshoot, and it happened before the pause, not because of it. Meanwhile an unfilled market sell order sits idle for fifteen minutes and reopens into whatever the auction produces. Cash on the sidelines buys at reopening prices, not halt prices. And an option expiring that afternoon keeps decaying on a clock you cannot trade against.

When this arithmetic is wrong

Beta is not stable in a crash. Correlations converge toward 1.0 when everything sells at once. A portfolio measured at 0.7 over three calm years can behave like 1.0 for one ugly afternoon, which makes every low-beta comfort in the example above too optimistic on precisely the day it matters.

House maintenance requirements move mid-crisis. Brokers raise them on volatile names with little notice, sometimes to 50% or 100%. The 28.6% cushion computed above assumes 30% stays at 30%. If your broker moves to 50% overnight, the allowed decline at Reg T minimum equity goes to zero and the call is immediate.

The 3:25 cutoff shifts on half days. Early-close sessions move the equivalent boundary. Check the exchange calendar instead of assuming.

Level 3 is not the end of the risk. The cash market closes, index futures keep trading overnight in some form, and the next open reprices everything. A 20% close followed by a gap-down morning is not a single 20% event.

These are exchange rules, not statute. The 7/13/20 framework has stood since 2013 as an SEC-approved rule, and rules get amended. Verify the current numbers with the exchange before betting anything on them.

What this does not tell you

This is arithmetic on published thresholds, so it says nothing about the probability of a halt on any given day — that depends on a volatility regime nobody here can forecast. It also takes no position on whether halts reduce or amplify volatility. That question is genuinely contested: one argument holds that a pause gives information time to spread, another that it pulls selling forward as traders race to exit before the freeze. The evidence is mixed enough that quoting one side as settled would be dishonest.

Futures are a separate rulebook entirely. Equity index futures have their own limit-down structure with overnight levels, and those frequently bind hours before the cash market opens.

And the beta is yours to supply. The mechanics above work for any number; whether your actual holdings sit at 0.8 or 1.5 is something only your own positions can answer.

FAQ

What percentage drop halts the stock market?

Seven percent triggers a fifteen-minute Level 1 halt, 13% a fifteen-minute Level 2, and 20% closes the market for the rest of the day. All three measure against the S&P 500’s prior official close. On a 6,000 close, that’s index levels of 5,580, 5,220 and 4,800.

How long does a market-wide trading halt last?

Level 1 and Level 2 each run exactly fifteen minutes. Level 3 lasts until the next trading day, so anywhere from twenty minutes to six and a half hours depending on when it fires. Since Levels 1 and 2 can each occur only once, the most pause time a session can produce short of a full close is thirty minutes.

Can the market halt twice in one day?

Yes. A single session can contain a Level 1 halt, then a Level 2 halt, then a Level 3 close — three stoppages. What it cannot contain is two Level 1 halts or two Level 2 halts.

Does a 7% drop halt the market at 3:45 p.m.?

No. Levels 1 and 2 are unavailable after 3:25 p.m. Eastern. In that final thirty-five-minute window only the 20% threshold is live, so a decline of anything up to 19.99% trades straight through to the bell.

Why do individual stocks halt when the index has barely moved?

They use Limit Up-Limit Down, which watches each security’s own five-minute average price rather than the index. A Tier 1 stock above $3 pauses on a 5% move outside its band; a Tier 2 stock needs 10%, and both bands double in the last twenty-five minutes. Hundreds fire on active days while the S&P 500 sits inside 1%.

How much would my margin account lose in a Level 3 halt?

Multiply 20% by your portfolio beta, then by your leverage multiplier — 2.0 for a position funded at Reg T’s 50% initial requirement. Beta 1.0 gives 40% of equity; beta 1.4 gives 56%. At a 30% house maintenance rate, the call lands once position value falls past 28.6%, which requires beta of about 1.43 or higher on a 20% day.

Three things to check before the next volatile session

Tonight’s S&P 500 close, so you can compute your own three levels with 0.93, 0.87 and 0.80. Your broker’s current house maintenance requirement, usually buried in the margin disclosure and usually above FINRA’s 25% floor. And the LULD tier and band that apply to what you actually own.

Market-wide breakers get the headlines. Single-stock pauses are the ones you’ll live through.

This article is general information, not financial advice. See our disclaimer.

Also worth reading: Money Market Funds Breaking the Buck: What a Drop Below $1 NAV Means

More in Risk · all topics · calculators · how this was checked