which the human is processing the information, often called ‘‘market color’’. Incor-
porating this information from ‘‘traditional’’ quant sources can only improve ‘‘quan-
textual’’ portfolios.
.Further news differentiation RNSE provides a very rich picture of the news as it is
created and flows across the wire. Topic and product codes, headlines, and item types
are probably all useful for differentiating news that presents an arbitrage opportunity
from news that is merely stating prior price movements. There is also a great deal of
information to be inferred from content, using either supervised (human in loop) or
machine learning.
.Higher frequency trading The results here were on daily intervals. With ‘‘informa-
tion leakage’’ a significant amount of alpha is lost within the trading day on which
intraday news occurs. Intraday position entries and exits would allow for more
opportunity for alpha capture. Furthermore, trading into a position over several
hours rather than trading at or near the close should reduce transaction costs and
increase capacity.
.Pros and cons of optimizer-based portfolio construction Creation of a forecast for
excess return from news analytics and other information and feeding these forecasts
into an optimizer would allow for fuller utilization of capital, and better position
sizing, based on liquidity. It also allows for risk management based on sectors and
factors. However, these constraints may limit exploitation of the signal. Given the
types of bets the simulation has made, sector ETFs may be useful to include as an
asset.
It is interesting to note that a research group at Deutsche Bank (Cahan et al., 2010)
reported on a similar approach to equity portfolio management after this chapter was
presented. They evaluate news strategies separately and in the context of a variety of
quant approaches. Pure news portfolio simulations included in the Deutsche Bank work
Relating news analytics to stock returns 169Figure 6.22.Rolling one-quarter correlation of extreme sentiment simulated portfolio with the
S&P 500.