Smart Beta Sentiment Enhanced ETF Performance Analysis
At SMA we continuously research our data. Below we discuss modifying weights of the SPDR SPY ETF based on sentiment values and examine the impact on return. Please contact SMA (info@SocialMarketAnalytics.com) to learn more.
The SPDR SPY ETF is a cap-weighted ETF which closely replicates the performance of the S&P 500. Our objective is to develop a “smart beta” strategy using the social media sentiment levels of individuals ETF constituents and amplify or accentuate the weights of the constituents in the ETF while keeping the Assets under Management constant. The transaction cost assumption is ignored for both the original and the enhanced ETF.
One of the strategies explored was looking at the sentiment levels an hour before the close (2:55 PM Eastern Time) and re-balancing the weights according to that. The stocks were bought or sold (to reduce position as per new weight only, NO short selling) at the close of the day and the positions were maintained until the next day when the re-balancing was performed again. To explore the weight modification methodology please contact SMA.
Our re-balance strategy keeps the AUM constant with no need for additional funds. Another strategy explored was to use a “lagged” sentiment. The lag being a day. So, for adjusting the weights today, we looked at the sentiment at 2:55 PM yesterday, and changed the positions based on that.
The results for the cumulative returns calculated over the period extending 7/31/2013-8/31/2015 are summarized below. Chart 1 shows the cumulative returns over the period for the “Original” which calculates fund returns using positions and closing price data. The “500% PM” makes the calculations using enhanced weights based on the pre-close sentiment. The “500% PM Lagged” has enhanced performance using pre-close sentiment from previous (trading) day.
Chart 2 shows the cumulative out performance, for the 2 “smart beta” strategies. As you can see both strategies track the SPDR SPY ETF while outperforming performance. You see the benefit of adding sentiment to your calculation process without increasing risk.
This is preliminary research we will be enhancing and updating over the coming weeks.