Nasdaq’s 5.5% Fall: Where Does November 5 Rank?
A consistent FRED calculation ranks November 5 first by points but 36th by percentage loss, with the ten worst sessions and their subsequent returns.
The Nasdaq Composite closed at 24,315 on November 5, 2026, down roughly 5.5% in a single session. How bad was that by historical standards? Recalculated from the daily closing series used here, it was the largest point loss, but the 36th-largest percentage decline.
Those answers are consistent. When an index rises from hundreds of points to tens of thousands, the same proportional loss produces a much larger point change. A record in points does not automatically mean a record in the share of an investor’s capital lost.
First fix the index, the date and the denominator
This article measures the Nasdaq Composite, not the Nasdaq-100 or QQQ, which tracks the latter. Their constituents and weights differ, so their daily moves and historical rankings differ too.
November 5 is the US trading date. The regular close was at 5 a.m. on November 6 in Taipei. The calculation uses Nasdaq’s NASDAQCOM series, distributed by the Federal Reserve Bank of St. Louis through FRED: 14,058 valid closing observations from February 5, 1971 through November 6, 2026, yielding 14,057 adjacent-session returns. Non-trading days and missing-value rows are removed before taking the previous valid close.Series and definitions
FRED records 25,736.510 for November 4 and 24,315.000 for November 5. Those observations imply a loss of 1,421.51 points, or 5.523320761%. Both changes are calculated from the same closing series to maintain a consistent basis.
Returns are ranked before rounding. That matters near the cutoff: November 5, 2008 lost about 5.5322%, slightly more than the latest 5.5233%. Both can become “down 5.5%” in a headline without being tied.
The ten worst sessions were not ten independent crises
The table uses the same FRED series for rankings and subsequent returns, with a data cutoff of November 6, 2026. Forward returns start at the event-day close. One month means the same date in the next calendar month; twelve months means the same date in the following year. If that calendar date does not exist, the month’s last day is used. If there is no closing observation on that date, the calculation takes the first available trading close afterward. These are not fixed 21- or 252-session windows, and the starting point is not the close before the crash.Daily source data
| Rank | US trading date | Daily change | 1 month later | 12 months later |
|---|---|---|---|---|
| 1 | 2020-03-16 | -12.32% | +23.58% | +95.11% |
| 2 | 1987-10-19 | -11.35% | -12.85% | +7.09% |
| 3 | 2000-04-14 | -9.67% | +8.62% | -42.51% |
| 4 | 2020-03-12 | -9.43% | +13.76% | +84.95% |
| 5 | 2008-09-29 | -9.14% | -16.46% | +7.07% |
| 6 | 1987-10-20 | -9.00% | -4.67% | +18.56% |
| 7 | 1987-10-26 | -9.00% | +5.88% | +28.87% |
| 8 | 2008-12-01 | -8.95% | +16.75% | +55.63% |
| 9 | 1998-08-31 | -8.56% | +12.98% | +82.71% |
| 10 | 2008-10-15 | -8.47% | -8.98% | +33.47% |
| 36 | 2026-11-05 | -5.52% | NA | NA |
NA means the observation window has not matured. It does not mean zero, and no forecast has been substituted. The November 2026 event has not yet reached either horizon. Returns measure index prices only, excluding reinvested dividends, fees, taxes and currency effects.
Three of the top ten sessions occurred in October 1987, two in March 2020 and three in 2008. These are multiple observations within a handful of episodes, not ten independent experiments. Counting how many subsequently rose cannot establish the probability that the next crash will be followed by a gain.
Data versions also matter. Using the currently downloaded FRED closes puts October 20, 1987 sixth and October 26 seventh. Other historical tables may reverse those dates because their recorded closes or rounding differ. A ranking is most useful when its source and calculation travel with it.
The crises behind the dates
These are the market conditions surrounding the sessions, rather than claims that any one announcement caused the entire decline.
The three entries from 1987 belong to the same crash episode. October 19 was Black Monday; October 20 and 26 fell within the continuing turmoil that month. The Federal Reserve’s retrospective describes interactions among falling prices, program trading and portfolio-insurance sales, margin demands and difficulty processing trading information. The Fed supplied liquidity and encouraged continued bank lending to preserve market functioning. The episode illustrates how funding and trading mechanisms can compound a sell-off, beyond changes in expected corporate profits.Federal Reserve retrospective
August 31, 1998 came amid global risk aversion after Russia’s debt crisis. The Fed’s contemporary review records Russia’s August 17 announcement of ruble devaluation and a moratorium on servicing official short-term debt. Markets subsequently became less willing to bear risk. Revelations about Long-Term Capital Management’s large positions and losses worsened conditions in mid-September. The August decline should therefore not be presented as occurring after the later public LTCM crisis: the episodes were connected, but their chronology matters.Federal Reserve report
April 14, 2000 fell within the dot-com bust. The Fed subsequently reported that after equity indexes reached March highs, concern about elevated valuations in some sectors and the prospect of higher interest rates contributed to a spring decline that was particularly sharp in the Nasdaq. That documents the broader valuation reset, rather than proving a single cause for April 14. A very large daily loss does not establish that a prolonged reassessment has finished.Federal Reserve report
The three entries from 2008 trace different dimensions of the same financial crisis. On September 29, the House rejected the troubled-asset rescue proposal by 205 votes to 228. That is a verifiable policy event on the day, not a calculation of how much of the market’s loss it caused.House vote, St. Louis Fed timeline
The October 15 Beige Book reported weaker September activity in all twelve Federal Reserve Districts, with pressure on spending, manufacturing and credit. On December 1, NBER announced that a recession had begun in December 2007. This dated an existing contraction; it did not mean the recession began on the announcement day. All three sessions belonged to the financial crisis, but the policy and economic contexts differed.October Beige Book, NBER announcement
March 12 and 16, 2020 occurred during the pandemic shock. The Fed’s May Financial Stability Report describes declines in expected income and risk appetite as concern about COVID-19 intensified, together with deteriorating equity-market liquidity and repeated circuit-breaker events in March. The subsequent rebounds describe the path after that particular shock; they are not a template for a different crisis.Federal Reserve Financial Stability Report
A large loss does not locate the bottom
The forward-return columns can tempt readers into a simple conclusion: if most eventually recovered, is a crash a buying signal? The table instead shows how differently recoveries can unfold.
March 16, 2020 was the deepest closing percentage decline in this series. From that session’s close, the index rose 23.58% over the one-month window and 95.11% over twelve months. But after falling 9.67% on April 14, 2000, it still lost another 42.51% over the following twelve-month window. A severe daily decline does not exhaust the market’s capacity to fall.
October 19, 1987 offers another path: the index was down a further 12.85% one month later, yet up 7.09% after twelve months. Short-term deterioration and eventual recovery can both be true. Looking only at the final holding-period return conceals the stress along the way; this table does not calculate the maximum intervening drawdown.
Buying at a crash-day close also differs from already owning the index before it fell. A positive forward return means the later close exceeded that day’s depressed close. It does not establish that an investor who bought at an earlier peak had recovered. After a 10% loss, an 11.11% gain is required to regain the original level because the denominator has changed.
What the ranking can establish
Percentage changes make sessions at different index levels more comparable. Recalculating both rankings and forward returns from one series also keeps the exercise internally consistent. On that basis, November 5 was an unusually severe session, while still falling short of the largest proportional closing losses in the record.
The limitation is just as clear. A ranking describes the size of a price change; it does not allocate causation among particular news or policy events. Nor is a table of ten extreme sessions a complete study of every day that lost 5%. Clustering, changes in market structure and selecting events after the fact all limit its predictive value.
Taiwanese readers face another distinction. A dollar-denominated index return is not an investor’s return in Taiwan dollars. And a Nasdaq decline is not a measure of lost orders at a Taiwanese supplier. Markets can reprice expectations quickly; changes in customer demand, capital spending and deliveries require separate evidence.
The useful record of November 5 therefore contains both findings: the largest daily point loss in this dataset, and its 36th-largest percentage decline. One reflects the index’s present scale; the other locates the session across decades. What happens next depends on whether the underlying pressures persist and how they develop. A place in the historical table cannot answer that question in advance.