RSI (Relative Strength Index): Timing The Next Correction

In the world of technical analysis, there is one reliable indicator for measuring market risk. The relative strength index (RSI) measures overextension (in either direction). Developed by J. Welles Wilder in 1978, the RSI is a momentum oscillator. As such, it measures the velocity and magnitude of price changes, plotting those on a scale from 0 to 100.

At its core, the RSI compares the magnitude of recent gains to recent losses over a chosen lookback period. Investors commonly use a lookback period of 14 days to estimate whether a market (or stock) is overbought or oversold. Readings above 70 are often flagged as “overbought,” while readings below 30 are considered “oversold.” The chart below shows a stock market sample versus the 14-day RSI.

S&P 500 Large Cap

Why is it so helpful? Because markets don’t just move in a straight line forever. While it may seem that way at extremes, momentum tends to exhaust, and reversals or corrections become more probable. The RSI gives us a real-time gauge of when a trend may be vulnerable to a pullback or turn. It’s not perfect, and as is often the case, things can remain “overbought” much longer than seems logical. However, as a risk management metric, it’s among the cleaner, more actionable oscillators in an investor’s toolkit.

Importantly, the RSI also gives us a way to spot divergences. For example, notice above that in December 2023, RSI peaked and began to decline even as the market advanced. This is known as a negative divergence and tends to precede corrections. Other signals include failure swings (RSI pushes above a threshold and fails) or confirmation of trend strength (e.g., crossing 50). Notably, RSI is not a standalone “buy/sell” signal. However, it is one of the more disciplined guardrails to judge when exuberance or fear is becoming excessive.

A Thought Experiment

While I find RSI very useful in managing short-term portfolio risks, I was looking at a long-term market chart and wondered if it could give us some clues about the next more significant corrective cycle. I am not talking about minor 5-10% corrections in any given year; those happen constantly. As we noted in “Bullish Years Have Corrections,”

During bullish years, corrections happen more often than you think. However, when corrections occur, it is not uncommon to see concerns about a “bear market” rise. However, historically speaking, the stock market increases about 73% of the time. The other 27% of the time, market corrections reverse the excesses of the previous advance. The table below shows the dispersion of returns over time. Critically, note that drawdowns of greater than 10% only occur 13% of the time.

average returns

However, 10% or less market corrections are more common and occur in every bullish year, as shown.

annual vs intra year

For this thought experiment, I was curious about the lead time between more extreme weekly RSI readings and the time lag to a more notable correction or bear market. While most investors use a daily measure of RSI, such as 14 days, over extended periods, a daily reading becomes more noise than signal. Therefore, to generate a better “signal”, I used 14 weeks to smooth the volatility. That chart of the 14-week RSI versus the S&P 500 is shown below.

relative strength index