Yesterday we did a study:
What happens to the S&P 500 when the tech sector gets crushed and Dow rallies on an individual day? What happens when XLK (tech sector ETF) closes down more than -2.4% while the Dow Jones is positive?
Since XLK’s history is so limited (began in late-1998), the results were meaningless.
Fortunately, Sentimentrader did a similar study but used a different calculation. Here’s how they did it:
After the NASDAQ makes a new high, look at the NASDAQ’s Z-score on any given day. Note the cases in which the Z-score is lower than -1.5. (A z-score is basically a standard deviation. A large negative z-score means that NASDAQ crashed).
The results weren’t interesting on their own. Sometimes the NASDAQ (and S&P) would rally after a big negative z-score, and sometimes the NASDAQ would fall.
However, the results were interesting when there was a cluster of negative z-scores. When there was a cluster of negative z-scores, the S&P almost always made at least a 6%+ small correction very soon. There was only 1 exception in early-1999.
*Sentimentrader considered “a cluster” as 2 or more large negative z-scores within less than 3 months.
We have recently experienced a cluster of NASDAQ z-scores below -1.5.
- May 17, 2017
- June 9, 2017
Here are the historic cases.
Perhaps the 6%+ small correction that we’ve been predicting since mid-May has begun right now.
Some of the historical cases were followed by small corrections while others were followed by significant corrections.
Fortunately, our model states that a significant correction for the S&P isn’t on the horizon. We will shift from 100% cash to 100% long UPRO when the S&P makes a 6% small correction.
2017 is shaping up to be a great year for our portfolio. We’re up 17% year-to-date. If we get this one right, our fund will most likely be up 50%+ by year-end.
Here are the historical cases in detail.
July 7, 1986
NASDAQ’s z-score signal came out on June 9 and again on July 7. This was the first leg of a 8.4% small correction for the S&P 500.
June 6, 1996
This signal came out on May 2 and again on June 6. This was also the first leg of an 11% small correction. (This correction would have been smaller. But the S&P made a mini-crash on July 17 when TWI Flight 800 went down. People thought it was a terrorist attack. See our history of 1996).
July 21, 1998
NASDAQ’s signal came out on April 23 and again on July 21. This was the beginning of a significant correction. Our model was able to predict this significant correction, but we were early. (The model’s SELL signal came out in late-March).
January 21, 1999
The signal came out on January 12 and again on January 21. This was the false signal. The S&P continued to rally and only made a 7.1% small correction in May 1999.
March 13, 2000
This was a massive cluster in less than 3 months. The signal came out on January 4, January 24, February 18, and March 13. The S&P made a 10.3% small correction from January 3 to February 28, 2000.
Then the S&P rallied briefly. That was the top of the bull market.
We don’t need to worry about this case right now. Our model says that the next bear market is still at least 1-2 years away.