News Parsing
How does one parse the news? How does the news have informative value? See Magerman: David Magerman's Website. The anecdotal evidence: news results in short-term momentum and (against peers) longer term reversion. How does one trade this? It's easiest to consider the problem when we have a sample set that distinguishes good and bad news fairly well… like pharmaceuticals where patent approvals generate news. Once we have a high correlation between good news and good events we are set. Statistical learning of the form of neural nets, decision trees, etc.
Rotation
Some of the new research revolves around style rotation in equity space (notably value-momentum but other styles as well). Do these problems translate into GTAA? If so, why? If not, why not? My sense is that styles have to be pervasive and well-defined, especially in fund space.
