BUZZ Index: Using Social Media Sentiment to Outperform the S&P 500

ETFdb.com (ETFdb): Please tell us about yourself.

Jamie Wise (J.W.): I’ve been involved in the investment management industry for almost 20 years now. I started my career in the late 90s down at Citibank in New York. I ended up managing their Canadian Equity Derivatives Trading business. I moved back to Toronto in 2001 to run a variety of proprietary trading strategies on Bank of Montreal’s equity derivatives desk.

I launched Periscope Capital in 2008 as a hedge fund focused on Canadian derivative strategies. We wanted to market those services to non-Canadian investors who wanted a local manager that understood the Canadian investing landscape. We manage capital across some Canadian-focused arbitrage strategies, primarily for non-Canadian pension plans. We’re not fundamental investors at our core; our background is quantitative.

ETFdb: What was the motivation behind launching the BUZZ Social Media Insights Index (BUZZ Index)?

J.W.: Four years ago, we were interested in seeing how consumer product companies were taking advantage of data they could collect from online communities to understand what their customers were saying about their products and their brands. Those consumer product companies were able to use insights from that information to help them drive product, marketing and business decisions.

We started a project back then to understand what people were saying about stocks. We wanted to see if we could glean any insights out of the discussion which might be predictive for asset price movements. By 2013, Twitter adopted a methodology of ‘cashtagging’ stocks – which is instead of hashtagging, you put a little dollar sign in front of the stock symbol. That really was the tipping point in terms of allowing everyday investors to go online into a consistent forum, look up a stock and then connect with thousands of other investors that were interested in that stock and talking about that stock.

What happened immediately after that was the discussion online began to transition to a broad base of large-cap companies. The bigger the stock and the more widely held the stock, the more likely it would be that people would be online and talking about it. So, the amount of information online just exploded five-fold or more in very short order, and all of a sudden, we could see people talking about all the S&P 500 type stocks. We revisited our project and found that this concept could be predictive and that we could make use of it.

It doesn’t matter if five people or 50,000 are talking about that stock, you can run that conversation through an algorithm to try and determine some insight and some sentiment, but how do you really interpret that score, what do you do with it, and how can you develop a portfolio of best-in-class sentiment names that could have the best opportunity to outperform? All of our research pointed to the first big breakthrough in our analysis which is it really comes down to depth of conversation with respect to conversations online being predictive about asset prices. If you can identify the most talked about stocks, then you can have confidence that what people are saying about that stock, whether it’s positive or negative, will actually be represented in the market. This is, in essence, what (BUZ C) does. We licensed the index to Sprott Asset Management and the index launched on April 18, 2016. (For a full list of Sprott ETFs, click here.)

ETFdb: Can you explain how the BUZZ Index, (BUZ C), utilizes social media in selecting the companies at the security level in the ETF?

J.W.: We think of (BUZ C) as a smarter way of owning the S&P 500. One thing that most people in the investment world intuitively understand is that sentiment matters when it comes to determining stock valuations. The question is: how do you accurately measure that sentiment?

Our view is that we can measure sentiment from a longer-term perspective such as weeks and months.

What differentiates online social commentary from solicited sentiment or opinions is a diversity of opinion and a real independence of opinion. Individuals have become comfortable sharing their lives across social media and constantly look for feedback and validation. Moreover, they are comfortable relying on crowd consensus when it comes to making decisions. I think that is what is happening with respect to finance and stocks, in particular. What we do is harness those insights and by identifying the most talked about stocks, we can then analyze the sentiment to determine which stocks are being discussed in the most positive tone. The data allows us to calculate sentiment in a predictive manner with respect to stock prices; it is a huge breakthrough when it comes to sentiment analysis.

ETFdb: The BUZZ Index launched on April 18, 2016. Can you update us on the performance?

J.W.: The performance has been pretty good since we launched. The index is up around three percent and has kept up with the S&P 500. Over the past six months, (BUZ C) has outperformed the S&P by about four percent. In our tested models that go back to 2013, we see a 35% outperformance of our strategy relative to the S&P.

Even though it’s a relatively small representative sample since the April launch, the performance of the index is doing what we expect, which is to at least keep up with the S&P 500, and demonstrate periods of notable outperformance along the way.

ETFdb: BUZZ Index recently changed the number of ETF securities from 25 to 75. What was the rationale behind that change? What is the selection process for these securities?

J.W.: The original concept behind the 25 name index was to provide investors with the opportunity to hold a concentrated portfolio of high sentiment names. Stocks are weighted according to their positive sentiment, or positive insight score, that we refer to. So, the higher the degree of sentiment, the higher the weighting of the index; we had a maximum weight at the time of 15 percent.

Once we launched the ETF, we went around and had discussions with advisors and investors. Everyone was keen on the concept of exposure of social momentum, but there was some concern about the concentration exposure of the portfolio. So, we went from a 25 name index to a 75 name index with a maximum weight of three percent for any given holding. Again, the weightings are determined based on the degree of positive insight and sentiment; the higher the sentiment, the higher the weighting, but it’s now capped at three percent. The index still rebalances monthly. Typically, there are eight to ten security names that change every month.

Given that we are analyzing social momentum, a certain amount of turnover is needed to ensure we are at the right end of the distribution with respect to positive insights and sentiment all the time. The ETF wrapper is a good solution for that, as it can handle some turnover, typically 10-15%, in the index.

ETFdb: What type of investors should hold this ETF?

J.W.: If you are thinking of large-cap US equity exposure in your portfolio, then (BUZ ) should be a consideration for you. The sector focus within the ETF is really interesting. We have a wide collection of growth and value names in the portfolio. Through our back-tested models, we found that the consumer, technology and healthcare sectors are consistently the three biggest weights in the index. It is fascinating to watch other sectors, such as materials and financials, rotate in and out of the index. We believe this type of sector rotation can add value over the long term.

ETFdb: What type of trends do you see developing and evolving over the next two to five years in each of the following area: artificial intelligence, big data and social media?

J.W.: AI has really grown a lot in the last couple of years. Computers nowadays can understand the context of the conversation in a much more meaningful way than earlier forms of AI; the machines now can teach themselves. Finance is probably late to the game when it comes to AI. AI will have a huge positive effect on the financial community’s ability to understand all the data that’s being produced in the world and how it could lead to alpha-generating opportunities in investing.

The data out there is endless, whether it be on social media or online, and it’s growing at Moore’s law type rates. The challenge for everyone in finance is if you’re not thinking about how to make use of that data and deliver solutions to your clients that incorporate some sort of analysis or insight from that data, you’re probably already behind. I think the days of relying on a single analyst’s fundamental research view are behind us. Using big data and social media sentiment, we can analyze behavior and trends in a much more meaningful way.

ETFdb: What is your best piece of advice for investors and financial advisors in this environment?

J.W.: Embrace technological change. It’s coming to the world of finance. The robo-advisor community may be a part of that; what we’re doing at BUZZ Indexes is a part of that. Certainly the way that you interact with your advisor and the way you interact with investing will be changing over the coming years. The same amount of flexibility that you’ve had by embracing technological change in other aspects of your life, you should think about how you’re going to apply that mindset to your investing life. Because it’s coming and it will add value.

The BUZZ Index is one way to harness the power of social media, artificial intelligence and big data analytics. With a focus on large-cap U.S. equities, it has outperformed the S&P 500 since it was launched in April 2016 due to overweight positions in the consumer, technology and healthcare sectors.

For more information about the BUZZ Index, read The Creator of the BUZZ Index, Jamie Wise and Sprott’s John Ciampaglia Discuss the Recently Filed BUZZ ETF. Also make sure to read The Best ETFs to Capitalize on the Rise of Artificial Intelligence to find ETFs that help investors best capitalize on the growth of artificial intelligence.

For more ETF news and analysis, subscribe to our free newsletter.