Twitter Chatter Packed In New Index

October 15, 2015 In this index, tell me about the concept of a predictive tweet. How do you go back and determine that someone’s bullish tweet was predictive? How does this methodology work?

Gonta: We don't go back. We actually record their activity in real time. We have a data feed where we get prices from Nasdaq in real time. As the person is tweeting, if they tweet, "I'm buying Apple," our sentiment engine automatically scores that as a positive sentiment, and creates that into an artificial kind of mock trade or a purchase of Apple stock at that time. We know the price of Apple because we've got the feed from Nasdaq.

We then measure whether that “tweet” or that “trade” made money hypothetically on a same-day basis. We look at whether the stock went up by the close of business that day.

We’ve put thousands of hours’ worth of research into this. We don't ultimately know whether these people are actually trading. We don't know if what they're tweeting is actually true. And, quite frankly, we don't care.

What our system is meant to do is measure what the public activity is and whether publicly what they're saying is accurate or not. We can do that because the market is visible and the market doesn't lie. If the tweeter made money, hypothetically, then he'll have a positive score. And if the tweeter lost money, he'll have a negative score. Our system allows you to easily filter who was the smart money and who's the dumb money. What are the plans for an ETF? Will it be a broad-market sentiment fund that owns some of these stocks everyone talks about, like Apple?

Gonta: We launched this aggregate super index with S&P Dow Jones. It is a long/short smart-beta index. Every three months, we look at what stocks on Twitter have the most tweet volume. We sort that list and pick the top 25 stocks with the most volume. We exclude ETFs, penny stocks and any stock that has a market cap of $1 billion or less. These are the rules of how we select the stocks.

Inside of the three months, every day we both market-cap-weight each component and sentiment-weight it. We rebalance the index on the close of business, where we decide whether we go long or short or how long or how short we go each of the stocks in the basket. That’s a function of the average sentiment from the crowd for that stock that day.

We allow the financial Twitterverse, if you will, all the pundits, to determine—based on their sentiment—whether we go long or short a specific stock on the close for tomorrow. This is a daily process, as in daily rebalancing?
Yes, it’s a daily rebalance. We also reconstitute the index every three months. It's published by S&P Dow Jones, and can be found on a Bloomberg Terminal under the ticker symbols CROWD, CROWDTR and CROWDNTR—price level, total return and net total return. That's meant for crowdsourcing. This index has outperformed the S&P by over 26 percent in the past two years or so, since July 2013.

We launched the index in May 2015, but the history goes back to July 2013. May, to date, it’s up 2.5 percent while the S&P 500 is down about 5 percent. Should we assume this index—and ultimately the ETF—is a large-cap U.S. stock strategy? Is Apple always going to be there, given people’s fascination with the company?

Gonta: You're going to get a couple of stocks that are always in there. Apple is always in there. But thematically, the point of the index is to be what the collective crowd is talking about. Yes, everybody talks about Apple, but there are certain stocks that go in and out of the index; for example, not too long ago, Exxon Mobil wasn't in the index. But when oil prices collapsed, everybody started chatting about oil, and so Exxon Mobil made it into the index. It’s what people care about at any given time.

And we believe in the wisdom-of-the-crowds concept, which is a statistical concept, but one that holds true. That's why our index outperforms the broader market. Conceptually speaking, in an ETF form, if this rebalances every day, do you worry it will be an expensive strategy to pull off because you’ll be dealing with high turnover?

Gonta: The turnover is high, and we worked at it. But stocks are so liquid that the bid/offer is not that high when you look at it on a relative basis. It will be higher than a pack of other passive ETFs, but the performance that you get—26 percent outperformance over the S&P 500—compensates you for that. You pay for performance.

Having said that, we're also talking about other hold periods and other rebalance features when the ETFs arrive. There could be a version that rebalances once a month. There could be a version that rebalances once every two weeks. We're open to different structures of rebalancing as we build up a family of strategies. Do you have an issuer lined up for the ETF?

Gonta: We're in discussions with ETF sponsors right now. And there's quite a bit of interest, partly because there’s no ETF today that specifically isolates the sentiment risk factor. If you think about Fama and French, the sentiment risk factor is its own component that can drive stock prices. Markets get emotional. Moods affect both short- and long-term prices, and there's no way right now to just hedge that. This index is a completely untapped risk factor. Do you think this index, and ETF, will appeal to a certain demographic?

Gonta: It can appeal to any manager who’s looking to get exposure to the sentiment risk factor. It's also got an appeal to millennials, for example, because they're tech-savvy and the strategy has a global-advisor kind of feel—it’s set-it-and-forget-it in the sense that it lets the crowd decide how to rebalance the index. It's highly technologically focused.

Contact Cinthia Murphy at [email protected].

Find your next ETF

Reset All