An assay of the accuracy of analysts’ reports

Abstract

Analysts seek to ascertain the fair value of securities and avail this information to the investing public through reports. The constituents of these reports are, typically, a fair value price, which the analysts’ discern to be the actual price and, consequently, a projection of the direction (Strong Buy, Buy, Hold, Sell, Strong Sell, Underweight, Overweight) in which the equity of a particular company is deemed to headed. These fair value estimates and projections are predicated on specific models utilised by the analysts. 124 analyst reports were collected in the course of this study. We empirically examined the accuracy of the fair value estimates and projections in these reports, and found that, at a 99% level of significance, the analysts were deemed to have been inaccurate, according to a previously (academically) established metric of forecast precision. We also determined, through second-order analyses, a measure of concomitance between the forecast error made by analysts and time periods in which the projections were made. This finding affirms prior time-series studies conducted by academics, all of whom are duly attributed in the paper