In this project, quantitative analysts Amanda Ramirez, Alexander Magnusson and Michal Nowak, use quantified and utilized alternative data in the form of significant short positions and insider transactions to create an investment model on the OMXSBGI equities. Alternative data is the data that cannot be found in the financial statements or the stock performance of a company.
The short positions were filtered based on varying the acceptable size and price level to use as short signals. The insider transactions were filtered into only acquisitions, and then by varying the removal of quarterly/incentive buyers, volume level and clustering acquisitions within a certain time frame and number of transactions to use as long signals. The returns of the signals were measured by simulating trades of 1, 5, 20 and 100 day periods on the individual equities, and benchmarking the returns to the same position on the OMXSBGI.
The short signal with the best excess return compared to the OMXSBGI was when a significant short position of at least 1.5% of the total floated stock was entered, when the stock price was at the 10-20% high or low level of the historical price in the past 100 days. The best holding period for the short for this signal was 100 days, which outperformed holding a short position in OMXSBGI by around 10% over around 100 trades over the 10 year period tested.
For the long position signals, for a holding period of 100 days, going long after a cluster of at least 2 insider acquisitions within 20 days, but only removing quarterly buys, was the best performing strategy. This resulted in a 3.2% excess return on average over 251 trades over 3 years. Alternatively, for a 5-day holding period, applying all filters had an excess return of between 3-4%, but only produced 64-75 trades.
To read the full report, please see attached PDF below: