Determinants of quits and dismissals on a long-lasting unionized industrial construction project

Labor-market conditions, remuneration, and personal and worksite characteristics all affected the probability of a worker quitting or being fired prior to a reduction in force (RIF) on a large, industrial, union project under construction between 1999 and 2002 in the South Central United States. Using a Cox proportional hazard survival regression model applied to weekly payroll and monthly labor-market data, hazard ratio probabilities for adverse separations (quits and dismissals) are estimated. Straight-time and overtime hours, in particular, retained workers more effectively than did periodic increases in collectively bargained wage rates. Workers coming from afar were less likely to be fired, whereas, ironically, increases in travel incentives made workers more likely to quit. Older workers were more likely to stay until a RIF, and compared to journey workers, apprentices were more likely to stay. Overmanning crews made workers more likely to quit or be fired. This case study underscores the importance of labor-market options and worksite remuneration as factors influencing adverse separations.