Pairs Trading Strategy with Geolocation Data: The Battle Between Under Armour and Nike

Abstract

We investigate the fundamental linkages between “geolocation” (human movement) data and financial market’s equity price behavior. The geolocation positions were recorded intraday and cover the period from January 1, 2018 to July 17, 2018. Our initial data set spans over 54 billion observations and was provided by Fysical. We focus our study on a popular hedge fund trading strategy known as “pairs trading.” First, we collect Under Armour (UA) and Nike’s stock price and volume data. Second, we investigate the relative activity of people visiting a particular physical store of UA and Nike, as proxied from anonymous cell phone foot traffic. Third, we collect the relative sentiment for tweets to UA and Nike. After combining all the data, we glean the following fascinating results: (1) geolocation information is an important variable in pairs trading strategy between UA and Nike, as evidenced by the results from our feature selection popularity methodology; (2) surprisingly, pairs trading only using geolocation data yields positive returns, and employing machine learning methods and rolling analysis enhances the returns; and (3) pairs trading strategy incorporating geolocation information yields a cumulative return of 13.7% from January, 2018 to June, 2018, with an Annualized Sharpe Ratio of 3.9. Some caveats are in order – our results are very sensitive to transaction costs assumptions and our data set has serious limitations.

Keywords: pairs trading, geolocation data, machine learning, tweet sentiments, alternative data, artificial intelligence, intraday trading

JEL Classification: G1, G12, G14