years except 2006, which was the only year that resulted in a loss. Interestingly, price
momentum outperformed the sentiment-based strategies only in 2006.
Overall, it seems that the construction of bottom-up sentiment indexes outperforms
price momentum in predicting future returns of the S&P 500. In addition, I find that
event novelty is an important element of a news-based strategy, since including only
the most novel news stories results in significantly improved information ratios both
pre and post the market high in October 2007. Finally, considering the impact of
different company events adds value to the construction of market-level news sentiment
indexes.
5.3 Industry-level sentiment
In the previous section, I demonstrated how to construct a market-level news sentiment
index that can be used to predict the future direction of the S&P 500. Here, I will focus
How news events impact market sentiment 135Figure 5.2.Cumulative strategy returns covering the out-of-sample period May 2005 through
December 2009 for the Market-Level Sentiment Index (solid line)withan Event Novelty Score
filter, the Market-Level Sentiment Index (dashed line)withoutan Event Novelty Score filter, and
the 1-month price momentum strategy based on the S&P 500 (dot-dashed line). The sentiment
indexes have been constructed based on a 90-day trailing window, and a trading decision was made
based on the monthly index delta. A positive delta resulted in a long position, and a negative delta
in a short position.