Do Analysts' Exclusions Lead to Better Forecasts?: An Analysis of the Effects of Street Earnings Exclusions on Target Price Forecast Accuracy

Abstract: Sell-side analysts regularly make earnings forecasts, stock recommendations, and target price forecasts in their research reports. As part of their earnings forecasts, analysts commonly make "exclusions", which is the process of excluding expense and income items that they believe will not reoccur in future periods. This is done in an attempt to derive an earnings number that is devoid of transitory items that will likely not have earnings impacts in future periods. These earnings are regularly used as inputs into valuation models as part of analysts' target price formation. This thesis explores the effects of exclusions on analysts' target price forecast accuracy on firms listed on the S&P 500 between 2004 and 2013. Using a threefold approach to examine accuracy, we find that exclusions are statistically and economically significant factors that inhibit analysts' ability to accurately forecast target prices. Using an absolute error measure and two accuracy measures observing whether the target price is met or exceeded, we find the size of exclusions to be positively associated with target price forecast error and negatively associated with the target price being met. We also find that analysts' subjective (incremental) exclusions reduce overall forecast accuracy, and accuracy is improved by only excluding objective nonrecurring (special) items. Our findings contribute to the emerging body of research examining target price forecast accuracy and to the ongoing debate regarding non-GAAP earnings exclusions.