Oh, it has plenty of predictive power, it's just not evenly distributed. When economists weigh in on contemporary politics, there are usually a lot of assumptions that doesn't quite make it into the soundbyte (and I don't want to pin this solely on the press, there are plenty of economists that are happy with this situation, as long as "their side" gets a boost). This, however, is the case for all members of the "political" sciences.