This paper documents that the VAR method for estimating the Discount Rate and Cash Flow News components of stock returns at the firm level produces results that are highly sensitive to research design choices. To alleviate these concerns, an alternative method of the news estimation is proposed. This method is based on Campbell and Shiller's (1988) return decomposition, but directly computes the Cash Flow News as the sum of the revisions in analysts' forecasts without relying on the VAR system as a modeling mechanism for market expectations (RAF method). Moreover, this method does not require estimation of the implied cost of capital as proposed by Chen and Zhao (2010). Empirical results provide evidence that the RAF method outperforms both VAR and Chen and Zhao's (2010) methods. Monte-Carlo simulations suggest that the RAF method more accurately captures Cash Flow and Discount Rate news.