Statistical Based Earnings Estimates: Further Reflections and Extensions

Spizman (2013) compares earnings estimates using the Census Bureau’s PINC-04 Tables to earnings estimates reported in the publication Full-Time Earnings in the United States. His focus is primarily on whether it is better to use estimates of mean or median earnings to estimate lost earnings in a personal injury or death lawsuit, when the plaintiff has no earnings history. Public-use microdata from the Annual Social and Economic Supplement of the Current Population Survey and from the American Community Survey are applied to offer some critiques of the Spizman (2013) study and some extensions of it. Consistent with Spizman (2013), median earnings estimates are significantly lower than mean earnings estimates, leading to the conclusion that median earnings calculated using the ACS are the preferred approach if the goal is to estimate lost earnings for a typical person. This study finds, however, that other concerns Spizman (2013) raises are less justified.