Four questions, four studies. Each one is here with its sample size, its confidence intervals, and where it is weak. The honest answer to the headline question is that news coverage describes what has already happened and does not anticipate what happens next. We publish that conclusion rather than the version that would sell better.
Most sentiment products are sold on the claim that they predict. We tested that claim against our own data and it failed. Publishing the failure is more useful than hiding it, because it tells you what an indicator like this can honestly be used for. And stops you attributing a signal to it that isn't there.
Open an Axi Account →