John Shon investigates herding by sell-side analysts in forecasting earnings
Associate Professor John Shon co-authored an article titled “Who Herds? Who Doesn’t? Estimates for Analysts’ Herding Propensity in Forecasting Earnings,” which was accepted by Contemporary Accounting Research and published In advance on the journal’s website.
Sell-side financial analysts provide an important function in capital markets by acting as sophisticated information intermediaries, providing analysis and information to present and potential shareholders about the firms that they follow, including forecasts for a firm’s future earnings. However, much of this valuable information is lost if the analyst places too much emphasis on peer analysts’ forecasts instead of calculating a unique forecast.
Though prior studies have documented this phenomenon of “herding” on an economy level—that is, in the aggregate—there has to date been no way to specifically estimate the precise amount of herding that a specific individual analyst may engage in. Shon’s paper’s novel approach of exploiting a “rational expectations” framework finds that more than 60 percent of analysts are herding, and that herding is more prevalent for analysts with less experience, in smaller analyst firms, when following a firm with small number of analysts, and/or when the forecast is made for a longer time period.
Furthermore, the paper “adjusts” analysts’ actual forecasts by backing out the herding effect to calculate the analyst’s raw, uncontaminated, unadjusted forecast. The paper finds that after these adjustments, analyst forecasts are more accurate, i.e., less biased, than their stated unadjusted forecasts. Such adjustments can help market participants better anticipate firms’ future earnings.