The QMIT Leveraged Buyout (LBO) Model & Enhancements via Sentiment Based Alternative Data
31 Pages Posted: 7 Aug 2019
Date Written: July 30, 2019
This paper introduces the QMIT LBO model and describes its salient characteristics. In addition to a 41% long term hit rate the Top 100 model predictions can be traded quite profitably as an equal weighted long portfolio. A Russell 2000 Value index hedge increases the Sortino ratio to ~2.5 over the 19-year history. It then synopsizes the SESI (sentiment) signal from RavenPack and investigates its merits as an overlay to the base level LBO Top 100 trading signal. Given that SESI captures newsbased sentiment which may include rumors on such LBO names it is logical to ascertain whether benefits may accrue from trading such a combined quantamental signal. We conduct a series of experiments involving the daily overlay of SESI for the 10-year period (2007-16) to the weekly rebalanced LBO Top 100 and find substantial improvements in annualized returns as well as Sharpe and Sortino ratios, not to mention the drawdown profile of the overlaid strategy. The best overlay scenario tested results in a 46% boost to the Sharpe ratio with an absolute +8.6% improvement to annualized returns.
Keywords: Factor Investing, LBO Model, Risk Arbitrage, Machine Learning, Smart Betas, Alternative Data, News Sentiment
JEL Classification: C53, G11, G12, G14, G15, G34
Suggested Citation: Suggested Citation