Estimation of Employment Survival Probabilities with Firm-Specific Data from
Abstract
It is well established in the forensic economics literature that workplace attrition or job tenure should be considered when calculating the economic damages associated with lost pay resulting from wrongful failure to hire or employment termination. (Baum, 2013; Trout, 1995; White, Tranfa-Abboud, & Holt, 2003) When available, firm specific data should be used in specific cases dealing with employment in industries where employment attrition is high because reliance on aggregate data would likely understate employment attrition, overstate the duration of employment and consequently overestimate any economic damages associated with forgone employment opportunities. This is especially the case in an industry such as long-haul trucking where employee turnover is high and employment durations are low relative to most other industries. In this paper, employment records for more than 6000 applicants who entered a training program to become truck drivers between 2009 and 2013 are used to illustrate alternative statistical techniques that can be used with firm-specific data to determine the probabilities of continued employment and expected employment durations that are required to properly estimate the expected value of forgone earnings in cases where workplace attrition adjustments are warranted. The paper contrasts the Binary Logistic (Logit) model and the Kaplan – Meier right-censored survival model that can be used with this type of firm-specific data.