The U.S. stock market’s momentum is INSANELY overbought on every single time frame: daily, weekly, and monthly.
Here’s the S&P 500 and its 14 daily RSI
Here’s the S&P 500 and its 14 weekly RSI
Here’s the S&P 500 and its 14 monthly RSI
Notice how RSI is above 80 on every single time frame. This is an extremely rare event that has only happened twice in history before January 11, 2018.
- July 5, 1955
- March 14, 1986
Let’s look at these 2 historical cases. Here’s the data in Excel.
July 5, 1955
The S&P began a 10.5% “small correction” in late-September 1955 (2.5 months later).
Its next “significant correction” (as defined by our Medium-Long Term Model) began 1 year later in August 1956.
March 14, 1986
The S&P began a 9.6% “small correction” less than 4 months later in August 1986.
The next “significant correction” was almost 1.5 years away (the crash of 1987).
The S&P 500 followed a similar, logical pattern in both of these historical cases.
- The S&P began a “small correction” that’s close to 10% within the next few months.
- The next “significant correction” is still far away. When momentum is insanely strong, the next correction will not be a “significant correction” because people will buy the dip.
*If you expand the data (e.g. using >RSI 75 as a criterion instead of >RSI 80), you get a similar conclusion.
These 2 historical cases were followed by very quick corrections (TIME), one of which was a V-shaped correction. This is logical.
- When momentum is very strong, everyone rushes to buy stocks.
- When the selling begins, all the short term traders get their stops triggered at the same time. This cases a mini-crash (e.g. 8-10% in a few days).
- But because the long term outlook is so bright, investors rush to buy the dip, thereby causing a V-shaped bottom.
I expect the next correction to be V-shaped. I also expect it to be closer to 10% than 6% (both are within the definition of a “small correction”).
I think the correction will most likely begin in February. But I could be wrong. It’s insanely difficult to predict the start of a “small correction”, and I don’t have a quantitative model for that.