On Thursday, Goldman Sachs' phones were ringing off the hook.
Extremely high level of distress.
This is a quote from John Flood, Goldman Sachs’ top equity trader, in his intraday report to clients on Thursday. His phone, in his own words, “is exploding.”
The Nasdaq 100 index closed down 6.4% on Thursday, erasing approximately $2.3 trillion in market value—the largest single-day decline since March 2020. The S&P 500 fell 4.4%, and the Philadelphia Semiconductor Index plunged 12%. The "fear gauge" VIX surged 70%, NVIDIA wiped out all its gains for the year, and SanDisk dropped 22% in a single day.
The immediate trigger for this crash was the Democrats' sweeping victory in both chambers of Congress during the midterm elections. Goldman Sachs immediately warned that this "blue tsunami" would lead to a halt in approximately 45% of planned U.S. data center capacity expansions—undermining the core narrative supporting AI tech stock valuations overnight with the reality of politics.
However, the severity of the market downturn far exceeds what can be explained by the policy shock alone.
In a report released during trading on Thursday, Nomura Securities derivatives strategist Charlie McElligott offered another explanation—that much of this crash was triggered by the smartest players themselves.
McElligott described a market paradox: on Thursday, during U.S. equity trading hours, after Morgan Stanley calculated $52 billion in pre-close mechanical sell orders at noon—a fact known throughout Wall Street—everyone tried to sell early and buy back at lower levels, turning the front-runners themselves into drivers of the downward momentum. He warned that if the usual “low point around 2–3 PM followed by a rebound” failed to materialize, the front-runners would also become forces pushing prices lower. Indeed, the day ended in severe losses. Combined with $47 billion in leveraged ETFs, $61 billion in volatility-controlled funds, and dealer negative gamma, total mechanical selling pressure exceeded $100 billion, and it will persist into Friday.
Today in the Asia-Pacific market, the sharp declines in South Korea and Japan are answering this question.
A $100 billion time bomb—everyone on Wall Street knows about it.
To understand McElligott’s logic, first understand a mechanism.
There is a category of products on the market called leveraged ETFs, such as "2x Long Nasdaq." Before each day’s close, they must perform one essential task: rebalance their positions back to the target leverage ratio based on that day’s gains or losses. When the market plunges, this means they must sell large quantities of stocks before closing— the steeper the decline, the more they must sell— and this process is passive, mechanical, and unavoidable.
Another type is volatility control funds (Vol Control), which systematically reduce equity exposure when market volatility spikes—also mechanical, typically executed one to two days later.
On Thursday, these two forces combined to an enormous scale.
Morgan Stanley’s derivatives team calculated at noon local time on November 5 that, assuming prices remained at the level seen at 1 p.m., forced selling of leveraged ETFs before close would amount to approximately $52 billion. Of this, the semiconductor sector alone would account for $21 billion, and the Nasdaq-100 for another $15 billion.
Nomura’s McElligott provided slightly lower figures, but they were equally striking: leveraged ETFs amounted to approximately $47 billion, the second-largest single-day rebalancing in Nomura’s database, behind only the $51.3 billion on June 5 of this year. He added that this selling pressure was “particularly painful” because 86% of the assets were concentrated in technology and AI-related securities.
Adding approximately $61 billion in volatility-controlled fund sell orders and an additional $7 billion in selling pressure from dealer negative gamma for every 1% decline, the market's mechanical supply easily exceeds $100 billion.
The key characteristic of these numbers is that they can be precomputed.
After Morgan Stanley calculated the $52 billion figure at noon, in McElligott’s words, “No information stays secret on Wall Street for more than 45 minutes”—the number quickly became “common knowledge across Wall Street.”
The smart money's plan: sell early, wait for the low, and buy back.
What was the traders' reaction to this mechanical sell order?
McElligott said that active traders typically attempt to "front-run" when faced with this type of mechanical flow.
The logic is simple: since there will be a large wave of passive selling before close, I’ll sell now and buy back at a lower price after the sell-off is over.
Moreover, this strategy has been effective recently. McElligott noted in the report that during recent periods of “single-day large volatility,” the market often forms its daily low between 2:00 PM and 3:00 PM Eastern Time, after which front-runners and hedging positions begin to reverse, take profits, and close positions, driving a rebound before the close and resulting in a closing price higher than the extreme implied by the anticipated mechanical selling.
Last Thursday, McElligott just reminded the market that “it’s happening again”—at the time, leveraged ETFs faced sell orders totaling $35 billion, and the market ultimately formed the classic pattern of a “2-3 point low followed by a closing rebound.”
During the South Korean market crash this past June, the U.S. stock market followed a similar script.
On the day the April tariffs were suspended, this mechanism reversed: McElligott calculated the largest historical pre-close buy-to-rebalance of +$34.3 billion, as front-runners bought in advance, ultimately driving the market sharply higher.
So on Thursday, when Morgan Stanley’s $52 billion figure spread across Wall Street at noon, every trader who knew the number was thinking the same thing: sell now, buy back at the low point around 2–3 p.m.
The trap of reflexivity: The cavalry is the one selling.
Then McElligott wrote in the report the chilling concern: “What if that ‘hypothetical 2 PM to 3 PM low followed by a rebound’ never occurs?”
Clearly, the alarming issue now is that these numbers are so large that if we don’t see the almost taken-for-granted late-day rebound—where price typically bounces back from a low by a standard one or two percent—and instead fail to recover during the standard TWAP window from 2 to 3 PM, making a new low instead… after falsely signaling to the market that it would rebound in the final hour, then closing at the lowest point.
This is precisely the core of the reflexivity paradox.
When everyone knows the same script and acts on it simultaneously, the script itself becomes ineffective.
The $100 billion in mechanical sell orders that were originally scheduled to occur before the close were dispersed throughout the day by active traders who front-ran them—these front-runners themselves became the force driving the market lower.
Meanwhile, there is a one- to two-day lag in sell orders from volatility-controlled funds—the plunge on Thursday would trigger further selling by these funds on Friday. McElligott wrote that the volatility-controlled selling corresponding to the afternoon 3 p.m. decline amounted to approximately $61 billion; if the decline expanded to 10%, this figure would surge to $70 billion.
Even more deadly: when all the front-runners have already sold and are waiting for “that low”—who will buy?
McElligott used a meme in his report to depict the scenario: everyone is sitting on chairs, waiting for the same rebound—but there aren’t enough chairs.
ZeroHedge summarized the analysis as:
Sometimes the cavalry arrives at 3 p.m. Sometimes, the cavalry is the one selling.
On Thursday, the cavalry didn’t arrive. The market closed at its low for the day: the Nasdaq 100 fell 6.4%, and the S&P 500 fell 4.4%.
Panic unfolds: The trade of the year is wiped out in 48 hours
Before the reflexive mechanism detonates, another powder keg awaits its fuse in the market.
McElligott dubbed this year’s “Trade of the Year” on Wall Street the “Semiconductor-Energy Dumbbell Strategy”—long semiconductor stocks (benefiting from AI computing demand) alongside long energy stocks (driven by geopolitical risk premiums and inflation hedging). This strategy generated a cumulative return of 66% just weeks ago, with a Sharpe ratio of 3.3.
Then, over the past 48 hours, both legs were swept out simultaneously:
- Energy: Down 10.2% this week
- Semiconductors: Down 13.1% this week
The weekly Sharpe ratios are -17.2 and -10.1.
The collapse in energy prices stems from the market repricing of the situation in Iran—McElligott noted signals from the White House suggesting a potential softening on Iran, causing the war premium to be rapidly unwound, resulting in the largest weekly decline in Brent crude since the COVID-19 pandemic.
The collapse of semiconductor stocks stems from market concerns over data center regulation following the Democrats' sweep of Congress.
Both directions collapse simultaneously, and funds holding a barbell strategy face not just losses, but forced liquidation. The mindset described by McElligott—“protect this year’s gains and survive into next year”—is completely shattered at this moment.
Walleye Capital announced today the return of client funds and the dismissal of the head of its equity division, calling it the worst month since the start of the COVID-19 pandemic. Goldman Sachs trader John Flood described long-only institutions as "once again stuck," noting that massive cloud computing positions bought during yesterday's rebound were quickly sold off, with "no buyers yet in sight."
Abnormal trading volumes at the ETF level confirm the extent of panic: SPY volume is 160% above its 20-day average, QQQ is 100% above, and the semiconductor ETF SMH is 150% above.
According to Jai Malhi, Senior Equity Strategist at Vanda Research, retail investors' selling volume on Thursday is "set to be the largest retail sell-off since late July last year."
The bonds have been short-squeezed, and the other leg is also bleeding.
Another leg, overlooked, broke at the same time.
The CTA trend fund has accumulated a large short position in bonds this year, as part of the "short bonds/long inflation" trade and as an implicit complement to the barbell strategy.
The stock market crash triggered a rush of safe-haven buying into the bond market, causing U.S. Treasury prices to surge and forcing short sellers to cover their positions. McElligott’s calculations show that CTAs face a “massive short squeeze” on bond positions: if U.S. Treasury prices rise another 2%, it would trigger approximately $225 billion in global forced bond buying, including about $48.5 billion in U.S. 2-year Treasuries, $69.5 billion in Eurozone 2-year bonds, and $23.9 billion in Japanese 10-year bonds.
This means that the short squeeze in bonds itself is becoming a force driving bond prices higher and further pressuring long equity positions—a further reflexive cycle.
The crash isn't over: Friday's delayed selling pressure
Closing on Thursday does not mean the end of this mechanical selling pressure.
There is a lag in the execution of volatility-controlled funds. McElligott wrote that Thursday’s market decline will continue to trigger sell orders from these funds on Friday—“today’s decline is further loading into the queue.”
UBS’s estimate is more straightforward: its risk control rebalancing model shows that if the S&P 500 falls by 2%, systemic funds would sell between $95 billion and $105 billion.
The Asian market is already answering this question.
During early trading on Friday, the Korean Composite Index plunged 8.8%, falling below the 5,000-point threshold and triggering a circuit breaker— a level previously regarded by South Korean President Lee Jae-myung as a milestone in economic recovery, now swiftly breached. The Nikkei 225 dropped nearly 5.8%, falling below 60,000 points during trading and retreating more than 20% from its June high, officially entering a technical bear market.
U.S. stock futures showed a slight rebound during Asian morning trading, with Nasdaq futures briefly turning negative, indicating that some investors are beginning to buy the dip. However, UBS warned that systemic selling pressure has not yet been fully released.
Another variable on Friday: the U.S. October non-farm payrolls report will be released that day, with market consensus expecting an addition of 59,000 jobs and the unemployment rate remaining at 4.2%. The direction of the data will determine whether this correction finds a window for respite.