On April 7, 2025, the S&P 500 swung 8.5% in a single session, one of the widest intraday ranges since March 2020. It closed down just 0.2%. No circuit breaker fired. Five years earlier, in March 2020, the same index tripped the exact same 7% threshold four separate times in eight trading days, and each time the market stopped cold for exactly 15 minutes, no more, no less. The gap between those two episodes says more about how circuit breakers actually work than any textbook definition does: the pause isn’t triggered by how bad a day feels, it’s triggered by a single number calculated the night before, and the duration is fixed in advance down to the minute.

Circuit breakers are one of the few pieces of market plumbing that most traders have heard of but few can describe accurately. The rules are precise, the history behind them is stranger than the mechanics, and at least one major economy tried a version of this system and pulled it within four days because it made things worse, not better.

The Three Levels, and What Each One Actually Halts

The U.S. market-wide circuit breaker (MWCB) system is governed by NYSE Rule 7.12 and mirrored across NASDAQ and Cboe. It measures a single-day decline in the S&P 500 Index against the prior day’s closing price, and it has exactly three thresholds.

Two details trip up most people who’ve only heard the numbers secondhand. First, a Level 1 or Level 2 breach at or after 3:25 p.m. ET does not halt anything. The exchanges decided that with only 35 minutes left in the session, a pause creates more confusion than it prevents, so the market is simply allowed to close wherever it lands. Second, each level can only fire once per day. If the market drops 7% and recovers to a 6% loss by the close, that’s the end of it, Level 1 doesn’t reset and refire on a second dip below 7%.

Level 3 is the outlier. It’s not a pause, it’s a full stop. A 20% single-day decline in the S&P 500 has never happened under the modern rule, and the closest the U.S. market has come since the rule took effect is nowhere near it. For context on how the exchange defines a normal session in the first place, the structure of U.S. market hours sets the 9:30 to 4:00 window that all three levels operate inside.

If the S&P 500 closes at 5,000, Level 1 triggers at 4,650, Level 2 at 4,350, and Level 3 at 4,000. These numbers are recalculated every single trading day off the prior close, which is why the same percentage drop looks like a different point number on the ticker from one week to the next.

Born From a Single Day in 1987, Not a Slow Policy Process

Circuit breakers exist because of one specific afternoon. On October 19, 1987, now called Black Monday, the Dow Jones Industrial Average fell 508 points, a 22.6% loss, still the largest single-day percentage decline in the index’s history. The S&P 500 lost roughly 20.4% the same day. Around $500 billion in market value disappeared in a single session, and NYSE trading volume nearly doubled its prior record, overwhelming phone lines and order systems so badly that brokers couldn’t reach their own clients.

President Reagan appointed a Presidential Task Force on Market Mechanisms, chaired by Nicholas Brady, to figure out what happened. The Brady Report pointed to program trading and portfolio insurance strategies, automated systems that sold into the decline and accelerated it, as central culprits. Out of that report came the first circuit breaker rule, implemented in October 1988: a 250-point Dow drop triggered a one-hour halt, and a 400-point drop triggered a two-hour halt. Those were fixed point values, not percentages, which meant the thresholds became easier to hit as the Dow itself grew over the following decade.

The system got its first real test on October 27, 1997, when the Dow fell 554 points, a 7.18% drop, and NYSE halted trading early under the point-based rule then in force. That was the only activation for the next 23 years. Even the 2008 financial crisis, despite months of brutal single-day losses, never produced a decline sharp enough in one session to hit the threshold; the damage in 2008 was spread across weeks rather than concentrated in a single trading day the way 1987 and 2020 were.

In 2012, regulators rebuilt the system from scratch after the 2010 flash crash exposed how badly the old point-based Dow rule fit a market that had changed. They switched the reference index to the broader S&P 500, moved to the 7/13/20% percentage structure still in use, and required daily recalculation instead of quarterly. The revised rule took effect on April 8, 2013, and it’s the version that governs every halt discussed in this article.

March 2020: The Only Real Stress Test the Rule Has Passed

The 2013 rule sat dormant for seven years. Then COVID-19 hit, and the S&P 500 tripped Level 1 four times in eight trading days, an event density the market hadn’t seen since the rule’s predecessor was written in 1988.

Three of the four halts were triggered within the first several minutes of the session, before every listed security had even finished its opening auction. That detail mattered enough that a joint working group of exchanges and regulators reviewed it afterward, checking whether halting the whole market before individual stocks had properly opened created its own distortions. The group’s conclusion, published in a formal report, was that the mechanism worked as designed and didn’t need structural changes, even though the timing was unusually early in the session on three of the four days.

A circuit breaker doesn’t undo a loss, it only interrupts the process of realizing it. Every point the index shed before the 15-minute pause was still gone the moment trading resumed; the halt bought time to think, not room to recover.

None of the four March 2020 halts reached Level 2. The closest the S&P 500 came to a 20-point-based Level 3 shutdown for the day was nowhere close in percentage terms, even during the worst single session of that month, when the index fell 12% on March 16, well short of the 20% threshold that would have ended trading outright.

The Other Circuit Breaker: What Happens to a Single Stock

The market-wide system protects the index. A separate mechanism, the Limit Up-Limit Down rule (LULD), protects individual stocks, and it exists because of a completely different crisis: the Flash Crash of May 6, 2010.

That afternoon, a single algorithmic sell order worth roughly $4.1 billion in E-mini S&P 500 futures, placed by an asset manager, drained liquidity so fast that high-frequency market makers pulled their quotes entirely. Between about 2:32 p.m. and 2:45 p.m. ET, the Dow fell nearly 1,000 points, close to 9%, then recovered most of it within the same hour. Individual names traded at absurd prices during the gap: Accenture, which had closed the day before near $42, briefly traded at a single penny, while Apple’s stock, closed around $250 the prior session, momentarily changed hands near $100,000 a share, trades the exchanges later canceled entirely because with no real buyers or sellers left, whatever stub quotes remained in the system executed.

Regulators’ first fix was a crude single-stock circuit breaker that paused any S&P 500 name for five minutes if it moved 10% in five minutes. That was replaced in 2012 by LULD, which is more precise: it sets a price band above and below a stock’s average trading price over the preceding five minutes, and if the price tries to move outside that band and stays there for 15 seconds, trading pauses for five minutes.

These bands widen automatically during the first and last 25 minutes of the session, when prices are naturally more volatile around the opening and closing auctions. That’s also the stretch covered by pieces on the first hour of trading, when order imbalances and price discovery tend to be roughest.

The Crack in the Popular Theory: China’s Four-Day Experiment

The assumption behind every circuit breaker is that a forced pause calms a panic. China’s own experience directly contradicts that.

On January 1, 2016, the China Securities Regulatory Commission introduced a circuit breaker on the CSI 300 index: a 5% move triggered a 15-minute halt, and a 7% move shut the market for the rest of the day. On January 4, the very first trading day the rule was live, the index fell fast enough to trigger both stages, closing the market roughly 80 minutes early. On January 7, the fourth trading day, the CSI 300 dropped 7% within 29 minutes of the open, making it the shortest trading day in the exchange’s 25-year history at that point. Regulators suspended the entire mechanism that evening, four trading days after launching it, admitting it had contributed to the selling rather than contained it.

The theory researchers use to explain this is called the magnet effect. When investors know a fixed halt is coming at a specific number, some of them rush to sell before that number hits, worried they’ll be frozen in a losing position once the halt starts. That rush itself pushes the price toward the threshold faster, and the halt becomes partly self-fulfilling. Academic studies on other exchanges, including a 2004 analysis of the Egyptian market by researchers Tooma and Sourial, found the same pattern: circuit breakers can measurably increase investors’ risk aversion right as the market approaches the trigger level, which is the opposite of the calming effect the rule is designed to produce. The U.S. system has avoided this outcome partly because its first threshold, 7%, is wide enough that few sessions get anywhere near it under normal conditions, unlike China’s tighter 5% opening band.

Futures Never Really Sleep, and That Complicates Everything

The cash equity market closes at 4:00 p.m. ET, but S&P 500 futures on CME’s Globex platform trade nearly around the clock. CME runs its own price limit system on E-mini and Micro E-mini S&P 500 contracts, mirroring the 7/13/20% cash-market levels, but during non-U.S. trading hours the futures market applies a hard 7% limit in both directions, up or down, regardless of whether the cash-market rule would apply.

The gap between an overnight futures move and an actual circuit breaker showed up clearly around the April 2025 tariff shock. On the evening of April 6, before the S&P 500’s 8.5% intraday session the next day, futures fell sharply, roughly 3.7 to 3.9% on the S&P and about 3.3% on the Dow, but never reached even the 5% limit-down band, let alone 7%. Japan’s market told a different story that same Monday: the Nikkei 225 fell 7.8% during the session and actually tripped its own circuit breaker, a reminder that different exchanges set their tolerance for a single bad day very differently. Unlike futures or Tokyo-listed stocks, cryptocurrency markets have no equivalent limit system at all; the article on why bitcoin never closes covers a market structure with no scheduled close and no regulatory circuit breaker of any kind, which is one reason crypto flash crashes of 20 to 50% in a matter of seconds have become almost routine since 2017.

Edge Cases That Rarely Make the Headlines

A few mechanical details matter more than they seem to at first glance.

Options and index futures tied to a halted index don’t keep trading independently. When NYSE Rule 7.12 triggers a cash-market halt, CME’s own rules require a corresponding halt in S&P 500 futures and options, so the two markets stay synchronized rather than one racing ahead while the other is frozen.

When trading reopens after a Level 1 or Level 2 halt, the first 15 seconds run under a locked-quote rule: traders can enter new orders but can’t cancel or modify existing ones, a small friction meant to prevent a rush of order cancellations from destabilizing the reopening auction.

Quarterly options and futures expirations, known as triple witching, and Fed announcement days tend to produce the sharpest intraday swings outside of genuine crashes, which is why volatility around those dates sometimes gets confused with circuit-breaker territory even though it almost never approaches the 7% threshold. Traders tracking when volatility peaks during a normal week will notice the pattern clusters around scheduled events, not random panic, which is a structurally different kind of movement than the unscheduled shocks that have historically triggered MWCB halts.

How the Rest of the World Handles the Same Problem

European exchanges lean on stock-level volatility auctions rather than one blanket market-wide percentage rule, which means a single catastrophic day in Frankfurt or Paris doesn’t halt trading across the board the way a 20% S&P 500 drop would in New York. Japan sits somewhere in between: individual names carry their own daily price limits, but the index itself can also trip a halt, as it did during the April 2025 tariff selloff.

The Market Is Now Betting on Whether Its Own Circuit Breaker Will Fire

By late 2025, prediction markets had turned the circuit breaker itself into a tradable event. On Kalshi, a contract titled “Will a NYSE marketwide circuit breaker happen before 2026?” saw over 126,000 transactions since it opened in December 2024, with implied probability hovering around 11% heading toward year-end, down slightly from the day before. It resolved no. Traders were, in effect, placing real money on the odds of a mechanism whose entire purpose is to interrupt trading during a moment of panic, a small irony given that the rule exists specifically because the 1987 crash proved investors couldn’t be trusted to price extreme risk calmly in real time. Whether that contract becomes a recurring annual market or fades as a one-off curiosity, it marks the first time the circuit breaker has been priced as an asset in its own right rather than just a backstop nobody expects to use.