A reader asked about the usefulness of the S&P Short-range Oscillator as sometimes used by Jim Cramer to forecast U.S. stock market returns. The self-reported “Performance” of the oscillator, relying on in-sample visual inspection with snooped thresholds, is of small use. Since continuous historical values of the indicator are not publicly available, we conduct an out-of-sample test by:
- Searching CNBC.com for “Oscillator” “Mad Money” and just “Oscillator” on October 3, 2019 and identifying articles with U.S. stock market forecasts from Jim Cramer based on the S&P Short-range Oscillator.
- Extracting the date for each forecast and determining whether it is call to be “In” or “Out” of the market.
- Calculating for each call a cumulative S&P 500 Index return starting at the next open after the article date (generally timestamped after the market close) for 21 trading days.
- Computing average cumulative performances of “In” and “Out” calls.
- Comparing these averages to that for all days spanning the search results.
Using the 15 qualifying articles and daily opening levels of the S&P 500 Index during June 16, 2008 through October 31, 2019, we find that:
Of the 15 qualifying articles, six are “In” and nine are “Out” calls. Five calls have cumulative return intervals that overlap with one or two other calls. It is arguable that clustered calls indicate high conviction, and we do not therefore winnow them.
The following chart tracks average cumulative S&P 500 Index returns for the 21 trading days (open to open) after “In” calls, “Out” calls and all days from the first call through 10/31/2019. On average, “In” (“Out”) calls underperform (outperform) average S&P 500 Index behavior. In other words, doing the opposite of the calls is advantageous.
In summary, very limited evidence indicates that the S&P Short-range Oscillator as applied by Jim Cramer does not work as described.
Cautions regarding findings include:
- The “In” and “Out” call subsamples are extremely small for reliable inference (but clearly do not support belief in usefulness of the indicator).
- Conversely, results do not strongly support belief that the calls are reliable contrarian signals.
For other research related to Jim Cramer, followed this search link.
The following table provides dates, quotes and associated “In” or “Out” judgments from qualifying articles for readers who may wish to check findings.
Frequently Asked Questions
What is the S&P Short-range Oscillator and how is it used to forecast the stock market?
The S&P Short-range Oscillator is a technical indicator that Jim Cramer sometimes uses to call for being "In" or "Out" of the U.S. stock market. Its self-reported performance relies on in-sample visual inspection with snooped thresholds, which makes it of small use as evidence. Continuous historical values of the indicator are not publicly available.
Does Jim Cramer's use of the S&P Short-range Oscillator actually predict stock market returns?
Jim Cramer's use of the S&P Short-range Oscillator does not appear to work as described, based on very limited evidence. Over June 16, 2008 through October 31, 2019, "In" calls underperformed average S&P 500 Index returns over the next 21 trading days, while "Out" calls outperformed. Doing the opposite of the calls would have been advantageous.
How can you test an indicator like the S&P Oscillator when historical values aren't publicly available?
The S&P Short-range Oscillator was tested out-of-sample by searching CNBC.com on October 3, 2019 for Cramer articles with oscillator-based "In" or "Out" market calls. For each of the 15 qualifying articles, the cumulative S&P 500 Index return was computed from the next open after the article date over 21 trading days. Average returns for "In" calls and "Out" calls were then compared to the average for all days spanning the search results.
Are Cramer's oscillator calls a reliable contrarian signal, or are they just useless?
Cramer's S&P Oscillator calls are not clearly either one. Since "In" calls underperformed and "Out" calls outperformed the average S&P 500 Index return, doing the opposite would have helped. However, the results do not strongly support belief that the calls are reliable contrarian signals, and they clearly do not support the indicator's usefulness as intended.
What are the limitations of testing Cramer's S&P Oscillator calls?
The main limitation is that the "In" (6 calls) and "Out" (9 calls) subsamples are extremely small for reliable inference. Five of the 15 calls also have 21-day return intervals that overlap with one or two other calls; these clustered calls were not removed because they may indicate high conviction. The sample depends on calls found through a CNBC.com search, so the evidence is very limited.