6.7.5 Portfolio characteristics
Few firms have large amounts of novel, extremely polarized news. This strategy made
active bets in 229 distinct securities selected from the S&P 1500. Details by month are
seen in Figure 6.17.
During this period (2006–2009) it will be no surprise that there were many more stocks
that experienced an extreme negative sentiment day than stocks that experienced an
extreme positive sentiment day.
6.7.6 Return distribution
Figures 6.18 and 6.19 show the distributions of returns to positions based on extreme
sentiment signals.
The portfolio is significantly impacted by the choice of stop loss. Many negative
returns occur in excess of stop loss, due to trading at the close. Allowing intraday
position exits would likely improve the stop loss rule significantly.
6.7.7 Portfolio beta and market correlation
Event-driven trading is inherently opportunistic. Our strategy is as well, reacting to days
on which there is strongly positive or negative news by allocating the strategy to news-
worthy securities. We constrained the portfolio to put no more than 15% of NAV into
any position. As seen in Figure 6.20, the portfolio had a highly consistent short bias, but
was profitable in both bull and bear market regimes. Daily beta (seen in Figure 6.21)
ranged from approximately 0.5 to2.5, averaging0.8 over the period. The rolling
one-quarter correlation of simulated portfolio returns with the S&P 500 are seen in
Figure 6.22.
166 News and abnormal returns
Figure 6.17.Number of distinct names held by the RNSE Extreme Sentiment Day strategy in any
given month. Black bars are the number of names held long, and grey bars are the number of names
held short.