Tech Stocks Just Sent a Signal They Have Sent 27 Times Since 1999. Three Times It Was a Crash. History Says One Move Covers What Comes Next

Tech stocks just sent a signal they have sent 27 times since 1999: the Nasdaq-100 closed 10% below a one-year high, the 28th time since the fund that tracks it, Invesco QQQ, began trading, and the other 27 times tell a retiree holding it most of what history has to say. The fund closed at $681.09 on Nov. 2, 10.3% under its Oct. 6 record of $759.66, fell to $647.05 by Nov. 5 in the chip selloff, climbed back to $690.38 on Wednesday, and then slipped about 1% after hours to $684, which puts it 10% under the record again.

The after-hours slide followed reports that Nvidia’s board is preparing to replace its chief executive, which sent Nvidia, one of the fund’s three largest holdings, down about 9% after the close. That is the one line this story spends on Nvidia. The pattern is bigger than one company.

A 10% drop from a high reads as a warning that the fund’s biggest holdings were priced for more than they are delivering; on the record below, it has also been a buying point far more often than the start of a crash. Both readings are true at once, and if you hold the fund, the three times the drop kept going are the reason you need a plan for the second one before you know which kind this is.

Take Tom, 68, a retired electrical engineer in Raleigh, a scenario built for this story. He and his wife, Ann, 65, have $900,000 in a rollover IRA: $300,000 in a Nasdaq-100 fund, $400,000 in an S&P 500 fund, and $200,000 in Treasury bills. Social Security pays them $2,900 a month, and they draw $36,000 a year from the IRA, a 4% rate. The tech sleeve was worth about $330,000 at the Oct. 6 record. It was $281,000 at the Nov. 5 low and is back to $300,000 at Wednesday’s close, so the paper loss from the top stands at about $30,000.

What followed the other 27 times

The count comes from every daily close of QQQ since November 1999, adjusted for its 2000 split. A drop counts on the first day the fund closes 10% or more below its highest close of the prior year, and the clock restarts only after it climbs back to a new one-year high. That produces 27 earlier drops, and here is what the fund did after each first 10% close.

Three months later it was higher 21 times out of 27, with a median gain of 10%. Six months later it was higher 22 times out of 26, with a median gain of 16%. Twelve months later it was higher 19 times out of 25, with a median gain of 18%. The two most recent drops, in March and July of this year, are too young for a full year; the March one was up 31% six months on.

On the median path, Tom’s $300,000 sleeve is about $355,000 a year from now. On the best path, the February 2020 drop, it would be $458,000, up 53%.

The six times it was lower a year later were all inside three stretches: the dot-com bust of 2000 (four of the six, counting three quick dips that winter that each bounced to a new high before the April 2000 break), the 2008 financial crisis, and the 2022 rate shock. Three times, in other words, the signal was the start of a crash, and the fund fell another 80%, 48% and 28% from the 10% line before it turned. The fund was back above the high it fell from within a year in 24 of the 27 cases. The other three are the ones to plan around.

Before you decide anything, know that in 6 of the 27 cases the fund fell at least another 10% from the first 10% close before it turned. Four of those six, in late 2018, early 2020, 2022 and the spring of 2025, were followed by a full recovery within a year or two. A second leg down is common. It is not, by itself, the start of a crash.

The three times it kept falling

April 2000. The fund closed 10% below its high on April 3, 2000, and went on to lose another 80% by October 2002. A year after the first 10% close, Tom’s $300,000 sleeve would have been $101,000, and at the bottom it would have been $58,000. His S&P 500 sleeve would have lost 27% in the same year, and the whole $900,000 portfolio would have been about $593,000 before interest on the bills. The fund did not close above its March 2000 record again until September 2016.

November 2007. The 10% close came on Nov. 12, 2007. The fund lost another 48% by Nov. 20, 2008. A year on, the sleeve would have been $185,000, the S&P sleeve $250,000, and the portfolio about $635,000 before interest. The fund was back above the high it fell from in January 2011, 38 months later.

January 2022. The 10% close came on Jan. 20, 2022, and the fund lost another 28% by Dec. 28 of that year. A year on, the sleeve would have been $240,000, the S&P sleeve $359,000, and the portfolio about $798,000. The fund was back above its old high in December 2023, 23 months later.

Two of the three were over in two to three years. The 2000 case took 16 years for the tech fund alone, and nothing in your withdrawal plan covers 16 years. What covered it then was the two-thirds of the portfolio that was not in the tech fund, which is the quiet argument for keeping the sleeve at a third or less.

The one move

The move is a withdrawal bucket: enough of your portfolio in Treasury bills that your next several years of spending never depend on the price of a stock fund. Tom’s $200,000 in bills covers 67 months of his $36,000 draw, about five and a half years, before interest.

Run that through the three bad cases. In 2008 the recovery took 38 months, and the withdrawals over it came to about $113,000, so the bucket paid every check without a single share of the tech fund sold below the line. In 2022 the recovery took 23 months and about $68,000 of withdrawals, the same result. In 2000 the bucket bought five and a half years, the S&P 500 sleeve carried the rest, and the tech sleeve was left alone to recover on its own schedule.

The alternative is to sell the sleeve at the 10% line and wait a year, which is what the pattern tempts a holder to do. On the 25 completed cases that seller dodged a loss 6 times and missed a gain 19 times, and the median missed gain on $300,000 was about $63,000. The bucket lets Tom skip that bet. He does not need to be right about which kind of drop this is, because his spending is funded either way, and neither do you.

At the Nov. 5 low the tech sleeve was $281,000 against a $300,000 one-third target, about $19,000 under. A once-a-year rebalance on a fixed date refills the bucket in good years and buys the tech fund in bad ones without anyone calling the bottom.

The bottom line

The Nasdaq-100 has closed 10% below a one-year high 28 times since 1999. A year later it was higher in 19 of the 25 cases with a full year of data, by a median of 18%, and it was back above the high it fell from within a year in 24 of 27. The six losses came in 2000, 2008 and 2022. For a retiree with $300,000 in the fund and $200,000 in Treasury bills, the bills cover about five and a half years of withdrawals, longer than two of the three bad recoveries, and the rest of the portfolio carried the third. If you own the fund, keep your bucket full, keep the tech sleeve at a third or less of what you have, and let the pattern finish on its own.

Contact [email protected] for any questions or corrections.