This paper empirically investigates learning among financial forecasters. I test for informational cascades or rational herd behavior among financial analysts. I conduct a set of nonparametric tests and show that analysts not only learn but also believe that their predecessors learn from each other. I also test a structural parametric model of social learning among financial analysts. This comes around the problem of common information in influencing similar forecasts among analysts. Here, too, I find no evidence that analysts make independent forecasts.