Can web traffic predict the behavior of trading markets? Yes, says Investopedia CEO David Siegel, who says his team's "Investopedia Anxiety Index" can actually predict impending spikes in market volatility.
The index, based on data from nearly 1 billion page views, gauges investor anxiety based on the number of Investopedia readers searching for 12 specific fear-based terms, such as "default," "bankruptcy," "short selling" and "inflation." Siegel and his team have found that when the number of searches on Investopedia for these terms grows, the VIX index tends to increase, several days later.
With more than 30 million monthly readers, Investopedia's readership encompasses a significant proportion of people participating in the global markets—so it's not too much of a stretch to assume what Investopedia readers care about might also be what the market cares about.
Recently, ETF.com sat down with Siegel at the Collision conference in New Orleans to learn more about how the Investopedia Anxiety Index works. A lightly edited transcript of the conversation follows.
ETF.com: When did you first notice that there was a connection between market volatility and your web traffic?
David Siegel: It was on Aug. 24, 2015, when the market went down about 1,000 points, the “flash crash.” I'm sitting in my office, thinking, "Oooooh, ****; this is terrible." Then our head of marketing runs into my office. He says, "The craziest thing's happening right now: Our short-selling tutorial is up, like, 1,000%. Tons of people are reading about short selling at the exact moment the market is down." I thought, "Wow, that's cool."
So we had our data scientists look into what kind of content would be indicative of having an "anxious" market: search terms like "short selling" or "correction" or "bankruptcy." We figured, with 30 million people on our site a month, if suddenly hundreds of thousands of people start looking up "bankruptcy," then weird stuff is happening, generally.
We plotted this anxiety index and found huge spikes around the Lehman Brothers collapse, the Greek bailout and so on. Then we looked at the VIX. We found that whenever the anxiety index peaks, the VIX peaked a few days later.
ETF.com: But why would it lead the VIX, specifically?
Siegel: When there was a big peak in the anxiety index, that meant people were educating themselves, doing research—and as you know, research and education happen before taking action. And if people are looking things up on Investopedia, they're more serious about taking action than if they're just Googling.
ETF.com: To what extent does Investopedia being a crowd-sourced publication influence your capability to track this kind of market anxiety?
Siegel: Completely. So, you've heard the study about people predicting the number of jelly beans in a jar, right? They asked 1,000 people how many jelly beans were in this jar, and they repeated the test multiple times. The study found that if you took the average answer, it would consistently come to within 5% of the actual number of jelly beans.
Well, our content isn’t crowdsourced (we have a large editorial team with finance and investing expertise), but you could say our data is crowdsourced, because ultimately it’s our users seeking and reading our content that creates our data points.
I think that speaks to the power of crowdsourcing. Crowdsourcing gives you a very high "N" [number of data points]. And that high N takes out the extremes, if you use a median.