About a year ago I went looking for the historical data on the Fear-Greed index (FGI) which was created and is being updated daily by CNN. The company ignores requests for their historical data. Looking for it online I saw forums of university researchers (not necessarily traders) who were trying to discover how CNN’s algorithm works just to recreate the index, but with no luck.
FGI is calculated daily through a proprietary algorithm, putting together seven different parameters dealing with the options market, stock and bonds return, advancing-declining NYSE volumes, volatility, NYSE NH-NL, junk bonds demand and market momentum. The result is a daily value between 0 (maximum fear) and 100 (maximum greed). I started collecting FGI values every day a year ago in order to create an indicator to plot with my Tradestaton charts: a task similar to what I already do with other indicators not available on professional trading platforms.
Luckily, next to today’s data CNN publishes FGI value for the same day a year ago; collecting data for a year I would then eventually have two years of historical data. Thanks to Michael G., a SpikeTrade Member (a community of serious traders founded by Alexander Elder and Kerry Lovvorn), I could add some more data to the CNN Fear-Greed Index and I have now 5 years of historical data in my database, enough to see whether there is anything interesting (click to enlarge).
In the above chart, FGI is plotted under the S&P500 index in different timeframes. On the left the weekly FGI is plotted under the weekly S&P500, whereas the daily, monthly and quarterly FGI timeframes are plotted on the right chart under the daily S&P500. The different timeframes have been calulated as an average of the previous days, so for example, the weekly is the average of the previous 5 days, the monthly is the average of the last 20 days and the quarterly includes the last 60 days. All the timeframes ranges from 0 to 100 as the original CNN FGI.
When charting FGI I found that the most informative results come from plotting it on the weekly chart of the S&P500.
The bars on the chart are colored red when FGI is -25 or lower (Extreme Fear) and green when FGIC is +25 or higher (Extreme Greed).
You can see on this chart that FGI tops out and starts to decline between 2 to 6 weeks in advance of S&P tops. For example, the latest top of FGI prior to the pandemic-driven downward reversal occurred in the very first week of 2020, when only some well-informed Chinese knew what was going on. That FGI top was followed by a remarkably steep descent of FGI while the S&P continued to rise. We know how it all ended up.
The weekly FGI appears more sensitive to catching market tops than bottoms.
This could be due to the very nature of the index: many of its components (options, volatility, junk bonds) are more actively traded by the big money than retail traders. The dominant emotion of market tops is euphoria, when big money decides it is time to go south, while crowds of traders continue to buy stocks looking for easy gains. Such opposite behaviors create a divergence between FGI and the S&P on our charts.
The dominant emotions at market bottoms are fear and capitulation. The mass of traders has already been thrown out of the market (some forever). When the big money decides it is time to go north, they find no opposition: no divergence.
A follow up to this post and the presentation of the Composite Fear&Greed Index, can be found here
(This article has been posted on SpikeTrade. Follow FGIC updates on www.spiketrade.com)
5 Comments
Paul Zhu
Thanks for the research. Two issues: 1. To profit from the spikes, you must complete one sell-buy cycle, or you may miss the big again because you already sold/ For example in your last chart there were 8 green signals, if you acted accordingly and sold your stock, you would indeed spare losses in 1 case, but you would be missing our all the big gains in all other 7 cases, so it is a very mis-leading signal. 2. On the other hand, if you buy when there is extreme fear, the reward is generally very good, so it is much more useful indicator for market bottom.
Gianluca Leo
Paul, thanks for comment. I would not take FGI or FGIC spikes or bottoms as trading signals. I agree that FGIC is much more useful in defining market bottoms than tops: markets have to panic before recovering.
Tonio Volcano
Very useful info. Thank you! Certainly we can use some of this to help us navigate the current market environment. I just took a quick look at VIX and McClelland indicator in confluence with this, and it seems that using the FGI as a filter could help avoid getting at the wrong time in a long trade… It seems that what could be read as extreme readings in the VIX or McClelland is not so, for as long as FIG is not on the extreme fear zone. Cheers!
Andrzej
Gianluca thanks for the research. Very interesting. Do you analyze FGI right now and actualize graphs? Today is extreme fear – 6 points. If you collect the data from the previous year you get the exactly day ? F.e. 9.05.2022 was monday but 09.05.2021 was sunday.
Gianluca Leo
Hello Andrzej, thanks. Yes I do this on a daily basis. The day is not necessarily the same, for example, 9.5.2021 could have been on a Sunday, but in analyzing historical data it is absolutely negligible.