An Examination of the External Validity of the Matching Between Concurrent Revenues and Expenses Estimated Using OLS Regression

I employ MM-estimation to investigate the external validity of the declining matching between concurrent revenues and expenses found in Srivastava (2014). Srivastava (2014) finds lower matching for successive listing cohorts. However, the low matching values in that study indicate the potential influence of observations with large losses that are a minority of the data. After relieving the impact of the influential observations using MM-estimation, I find no overall declining matching and no declining matching for successive listing cohorts over time. I find that the severe influential observations are mostly loss firms in the introduction stage of their life cycles, and their losses increase dramatically over time. This study updates accounting researchers and other readers on the status of matching over time for the majority of a large sample of public firms. Besides, this study directs the attention of accounting researchers, accounting standard-setters, and regulators to the declining matching of the firms in the introduction stage. The deteriorating performance of the firms in the introduction stage over time discourage startups, potentially explaining the decreasing rate of U.S. startups in recent years. Also, this study contributes to the research methodology in accounting research. I demonstrate the increasing impact of influential observations on ordinary least squares regression over time, the relative performance of winsorization and MM-estimation in treating influential observations, and how to use weights assigned in MM-estimation to identify the severe influential observations