CMStatistics 2017: Start Registration
View Submission - CFE
Title: Nowcasting earnings Authors:  Michele Modugno - Federal Reserve Board (United States) [presenting]
Abstract: Like GDP, companies earnings are quarterly data that begin to be released one or two weeks after the last month of each quarter. In other words, the majority of the public companies releases their earnings in early to mid- January, April, July and October. We propose a model to nowcast S\&P500 companies earnings, that makes use of data at higher frequency, and that are published in a more timely manner than earnings. Nowcasting earnings is extremely important given that they are the primary metric to evaluate the performance of a company. Moreover, market participants form expectations about them, and every time earnings are released, and they differ from market expectations stock price of publicly traded company move dramatically. We find that in order to predict the average earnings of S\&P500 companies a factor that capture the real side of the U.S. economy is enough to generate nowcast that are more accurate than market expectations, while factors capturing inflation, monetary policy, and the same stocks return have limited impact on improving the accuracy of the nowcast.