Investigating the Accuracy of Analyst Consensus for Earnings per Share of S&P 100 companies

Abstract: This study investigated what affects how accurately financial analysts can predict the earnings per share of companies included in the Standard & Poor’s 100 index. To achieve this goal data on earnings forecasts was gathered for the years 2000 through 2013. Further this study investigated if there are any differences in the accuracy of optimistic respectively pessimistic earnings consensus forecasts. Multiple linear regressions were used in order to answer the imposed questions. The factors found, in this study, to affect the forecast accuracy were consensus dispersion, company size, net margin, which exchange a company is listed on and, to some extent, a company’s industry classification. Further, the result of the study implies that there is no difference in the accuracy of optimistic respectively pessimistic earnings consensus forecasts. In this study the models utilised, and the factors investigated, could only explain a limited part of what affects the earnings forecast accuracy. Lastly, a concluding qualitative attempt was made to find factors which affect the accuracy but which were hard to incorporate in this quantitative study.