Using MD&A to Improve Earnings Forecasts
May 1, 2013 Leave a comment
Using MD&A to Improve Earnings Forecasts
Khrystyna Bochkay Rutgers, The State University of New Jersey – Rutgers Business School at Newark & New Brunswick
Carolyn B. Levine Rutgers, The State University of New Jersey – Rutgers Business School at Newark & New Brunswick
April 17, 2013
Abstract:
We estimate and compare quantitative and text-enhanced earnings forecasting models to evaluate the extent to which MD&A disclosures improve earnings forecasts. Incorporating MD&A disclosures into forecasting models significantly improves forecasting accuracy. The gains in accuracy are much greater following regulatory reforms, providing some of the first large-sample empirical evidence on the success of recent MD&A regulatory actions. The MD&A section is less informative between 2007-2009, particularly for those firms hardest hit by the financial crisis (i.e., firms with low cash and large changes in performance). Further, we find that text improves forecast accuracy most for firms in the consumer staples sector, firms with low profit margins, large changes in performance, high political and/or legal costs and high complexity. Last, we find that models enhanced by MD&A disclosures are generally less accurate than analysts’ consensus forecasts for large and medium sized firms, but equally accurate for small firms.