Level 2 Sits 6.45 Percent Under Level 1, Not Another Seven Percent

The market-wide circuit breaker is one of the few events in United States equity trading whose trigger is written down and published before the opening bell. Three thresholds, and everybody knows them: 7 percent, 13 percent, 20 percent. The thresholds are not where the reading goes wrong. The denominator is.

Consider two people watching the same tape on a bad afternoon. The first sees an index quote that reads 7.02 percent below yesterday's close and expects a halt. The second is watching after a Level 1 halt has already occurred and reopened, counting a further 7 percent down from the price at which trading resumed, and also expects a halt. One of those two numbers is measured the way the rule measures. The other is measured from a moving reference point that the rule does not use, and it is wrong by an amount that can be stated to the second decimal place. On a 6,000.00 prior close it is wrong by 30.60 index points in one direction and 54.60 index points in the other. This piece follows that single 7 percent all the way from the rule text to the point value the tape actually compares against.

One Fixed Number, Not a Rolling One

FINRA Rule 6121, Supplementary Material .02 carries the definition in one sentence. Paragraph (b) reads: “For purposes of this Rule, a Market Decline means a decline in the value of the S&P 500® Index between 9:30 a.m. and 4:00 p.m. on a trading day as compared to the closing value of the S&P 500® Index for the immediately preceding trading day.” Two facts are packed into that sentence, and both of them are the kind that a screen hides.

First, the comparison is against a single value: yesterday's closing value of the index. Not the opening price, not the reopening price after a halt, not the intraday high. Yesterday's close, one number, fixed for the whole session. Second, the window in which a Market Decline can exist runs from 9:30 a.m. to 4:00 p.m. A futures quote showing a steep overnight decline at 7:00 a.m. falls outside that window entirely, whatever the percentage on it.

Paragraph (a) of the same supplementary material adds the part that decides who is right when a screen and an exchange disagree. FINRA halts trading otherwise than on an exchange on a Market Decline “as determined by a primary listing market and publicly disseminated”, and the paragraph ends on a proviso, “except as otherwise provided below”. The determination belongs to the primary listing market. A terminal's percentage-change field is a derived display; it is not the instrument that trips the breaker.

The Ladder Arrives as Three Prices

Because the reference is a fixed number, the three percentages can be turned into three prices before the market opens, and they are. The SEC Investor Bulletin: New Measures to Address Market Volatility is explicit about the form the thresholds take: “These triggers are set by the markets at point levels that are calculated daily based on the prior day's closing price of the S&P 500 Index.” Point levels. Not a percentage evaluated continuously against whatever the market is doing, but three prices, published.

The NYSE Report of the Market-Wide Circuit Breaker Working Group records the plumbing and the clock. In its Section III.A: “At 7:00 a.m. and again at 8:00 a.m., the CTA SIP generates the three decline values (7% (breach Level 1), 13% (breach Level 2), and 20% (breach Level 3)), based on the index value of SPX at 4:00 p.m. the prior trading day.” Two generations, both before the open, both anchored to a 4:00 p.m. value from the previous session. By 9:30 a.m. the day's three trigger prices are settled facts.

Work the arithmetic on an illustrative prior close of 6,000.00 index points. The percentages are the rule's; the multiplication is ordinary. Level 1 is 0.93 times the close, or 5,580.00. Level 2 is 0.87 times the close, or 5,220.00. Level 3 is 0.80 times the close, or 4,800.00. Measured as distances from the prior close, the rungs sit 420.00 points, 780.00 points and 1,200.00 points down. Measured as the distance from each rung to the next, they sit 420.00 points, then 360.00 points, then 420.00 points apart.

What Another Seven Percent Costs After a Reopening

Now the second reader. A Level 1 halt has occurred, the fifteen-minute minimum has run, trading has resumed near 5,580.00, and the tape turns down again. The instinct is to measure the next halt as another 7 percent from here. The NYSE Market-Wide Circuit Breakers FAQ answers this directly, and the fact that an exchange put the question in its own FAQ is a fair sign of how often it is asked. The FAQ states: “Note that a Level 1 and Level 2 breach can be triggered only once per trading day.” It then works the example: “if the S&P 500 Index declines an additional 7% after the reopening following a Level 1 MWCB, another MWCB would not be triggered unless the Index declines by 13% (Level 2 breach).”

Turn that into prices. Another 7 percent from 5,580.00 would be 5,189.40. The rule's Level 2 is 5,220.00. The second halt therefore arrives 30.60 points before the reader expecting another 7 percent is looking for it. Expressed as a percentage from the reopening level, the distance from Level 1 to Level 2 is 360.00 divided by 5,580.00, which is 6.45 percent. Not 7 percent, and not 6 percent either.

The error flips sign one rung lower. Another 7 percent from Level 2 at 5,220.00 would be 4,854.60. Level 3 is 4,800.00. This time the expectation is too shallow: 54.60 points of decline remain after the imagined trigger, and the true distance from Level 2 to Level 3 is 420.00 divided by 5,220.00, which is 8.05 percent. A trader sizing a hedge to the assumption that Level 3 is 7 percent below Level 2 has the last 54.60 points unhedged.

The compounding version of the same mistake is worth stating because it is tidy. Three consecutive 7 percent declines multiply to 0.93 cubed, which is 0.804357, a total decline of 19.56 percent. On a 6,000.00 close that is 4,826.14, which is 26.14 points above Level 3 at 4,800.00. Three 7 percent legs do not reach the 20 percent rung. The ladder's rungs are 7, 13 and 20 measured from one place, and treating them as 7, 7 and 7 measured from wherever trading last resumed

The Same 420 Points, Two Percentages

There is a neater way to see why the ladder resists percentage arithmetic. On a 6,000.00 close, the distance from the prior close to Level 1 is 420.00 points. The distance from Level 2 to Level 3 is also 420.00 points. The same distance, twice, in the same ladder, on the same day. As a percentage it is 7.00 percent the first time and 8.05 percent the second, because the base under it has shrunk from 6,000.00 to 5,220.00. The 360.00 point gap in the middle behaves the same way: 6.00 percent of the close, 6.45 percent of Level 1.

This is the reason the three values are published as prices rather than left as percentages. A price is invariant to where the reader happens to be standing. A percentage is not, and a percentage quoted without its base is a number waiting to be misread. The same discipline applies one layer down in the single-stock regime, where the price bands are built from a reference price and a percentage parameter that have to be read together; that mechanism is worked through in the earlier piece on reading the reference price and percentage parameter.

After 3:25 p.m. Two of the Three Stop Working

The second reading problem is temporal rather than arithmetic. The same 5,580.00 that halts the market at 2:00 p.m. does nothing at 3:40 p.m. The rule text as filed is blunt about it, and one word off from everything published since. In the text of proposed NYSE Rule 80B carried in SEC Release No. 34-65427, Exhibit 5: “The Exchange will not halt trading if a Level 1 Market Decline or a Level 2 Market Decline occurs after 3:25 p.m.” The SEC bulletin draws the same line with the minute itself on the far side of it: “A market decline that triggers a Level 1 or Level 2 circuit breaker before 3:25 p.m. will halt market-wide trading for 15 minutes, while a similar market decline 'at or after' 3:25 p.m. will not halt market-wide trading.” The NYSE FAQ states the window from the other side: “Level 1 and Level 2 can be triggered between 9:30 a.m. and 3:25 p.m. Level 3 can be triggered at any time.” A decline printing exactly at 3:25 p.m. is inside the halt window on the proposed rule's “after” and outside it on the “at or after” the Commission and the exchange use; this piece follows the inclusive form.

Level 3 keeps working to the bell, and its consequence is categorically different from a fifteen-minute pause. The proposed Rule 80B text, which the operative NYSE rule now carries as Rule 7.12: “If a Level 3 Market Decline occurs at any time during the trading day, the Exchange shall halt trading in all stocks until the primary listing market opens the next trading day.” FINRA's paragraph (c)(2) mirrors the effect for off-exchange trading, halting it “for the remainder of the trading day”. So in the last thirty-five minutes of a session, two of the three published prices are inert and one of them ends the day. The measurement window in paragraph (b) still runs to 4:00 p.m., so 4,800.00 is live at 3:55 p.m.

The asymmetry has a practical edge for anyone who trades the close. Between 3:25 p.m. and 4:00 p.m. a decline through Level 1 or Level 2 leaves the closing auction intact, while a decline through Level 3 stops trading for the rest of the day. Those are two different states of the world reached by the same index falling past two of the three prices generated at 8:00 a.m. Single-stock halts behave on their own clock and their own triggers, which is a separate reconciliation and one covered in the earlier piece on what happens when a stock halts mid-session.

The Futures Carry a Seven of Their Own

One more seven is worth separating out, because it is the one most likely to be mistaken for the cash trigger before the open. The MWCB Working Group report records, in its Section II.B, a change made at CME: “Overnight Trading Hours ("OTH") Price limits in certain futures and options on futures were changed from 5% hard limits to 7% hard limits.” The report also sets out the recommendation behind it: “Consider changing the trigger for the limit-down state for the E-mini S&P 500 futures contract to 7% from 5%, to align with the Level 1 MWCB trigger of 7% for equities.”

Alignment of the number is not identity of the mechanism. The report's own word for the futures figure is a limit, and the word it uses for the equity figure is a trigger. A limit-down state and a market-wide trading halt are different outcomes, and the overnight percentage attaches to a futures contract rather than to the index value that paragraph (b) measures between 9:30 a.m. and 4:00 p.m. A pre-open screen reading 7 percent on the front-month contract is therefore adjacent to, and not the same as, the 5,580.00 implied by the decline values the SIP generated at 8:00 a.m. What reference price CME applies to those overnight limits was not checked here, and that gap is carried below rather than guessed at.

Concrete Framework

Write down the prior close first. Before the open, take yesterday's closing index value and multiply it by 0.93, 0.87 and 0.80. Those three prices are the day's entire circuit breaker state. Keep them as prices on the same screen as the index, not as percentages in the back of the head.

Stop computing percentage change from anything except that close. A reopening price, a session low, an opening print: none of them is the base. When a halt has already happened, the question is the distance from here to the next published price, not another 7 percent.

Precompute the two gaps once. From Level 1 the next rung is 6.45 percent away; from Level 2 the next rung is 8.05 percent away. Those two figures hold for any prior close, because they are ratios of the rule's own percentages. Memorising 6.45 and 8.05 removes the entire class of error described above.

Put a line at 3:25 p.m. in the trading plan. Before it, three prices are live. At or after it, one price is live and its consequence is the end of the session. Any hedge or stop whose logic depends on a pause of at least fifteen minutes arriving should be examined for what it does in the last thirty-five minutes, when that pause is unavailable at two of the three rungs.

Treat the pre-open futures percentage as a separate instrument. It shares a number with Level 1 by design, and it neither triggers nor forecasts the cash halt, because the Market Decline the rule defines begins at 9:30 a.m.

What Would Invalidate This

Several things, and they are worth listing precisely because the arithmetic above is only as good as the inputs.

The point values published each morning were treated here as exact multiples of the prior close. If the primary listing markets apply a rounding convention to those three values, then every price in this piece moves by a fraction of a point and the gap percentages move in the last decimal place. The retrievals used here returned no rounding instruction in the operative rule text. The exhibit's bracketed, deleted language did carry one — “Each Level shall be rounded to the nearest fifty points” — so the exactness assumed above rests on a deletion rather than on silence. A rounding convention published elsewhere by an exchange would refine, though probably not overturn, the figures.

The 6.45 percent and 8.05 percent figures depend on the thresholds staying at 7, 13 and 20 percent and on the reference staying the prior day's closing value. A rule filing that changed either would change both gaps. The same applies to the 3:25 p.m. boundary, which the Working Group report describes as having already moved once, from 2:30 p.m. The Level 3 consequence is already moving too: NYSE filed a change to Rule 7.12 on September 4, 2026, operative thirty days later, replacing the reopening-next-day language with a halt running “until 4:00 a.m. ET or later on the following trading day”. The arithmetic above is untouched; the sentence about Level 3 ending the session is not.

The reopening scenario assumes trading resumes somewhere near the Level 1 price. If a reopening auction prints materially away from 5,580.00 in the worked example, the 30.60 point figure changes while the 5,220.00 trigger does not, which is precisely the point. The invariant is the published price, not the gap in points.

Finally, the futures paragraph rests on an exchange report describing a change at another exchange. The CME rulebook chapter governing equity index price limits was not read for this piece, so the reference price those overnight limits attach to, and their behaviour during regular hours, remain unverified here. Anyone trading that alignment should read that chapter before relying on the number it shares with Level 1.