Dips and Floors in Workplace Training: Using Personnel Records to Estimate Gender Differences

Life-long learning is considered necessary to sustain employability at older age in light of continuous changes in the labor market. Nevertheless, it is a challenge to sustain the training of employees throughout their professional careers. In particular, this is the case for females, who typically receive less formal workplace training than males. As professional careers of females are characterized by family breaks and periods of part-time work, they are less likely to invest time in workplace training. Since disparities in career paths, in particular between males and females, might lead to differences in training, which then tend to reinforce career inequalities, this paper analyzes the determinants of training not only with a special interest in age but also with a focus on gender. A detailed analysis of the training gap between males and females requires detailed data and a decomposition capturing the dynamics over professional life. We use personnel records from a single company with a high-skilled workforce from the financial industry. Our outcome variables of interest are the probability to participate in company-provided formal training and the length of training per year. We extend the standard Blinder- Oaxaca approach and decompose our two outcomes - training probability and training duration - into three terms: an age-specific coeffcients effect, an age-specific characteristics effect, and an age composition effect. We additionally include supervisor fixed effects to analyze if supervisors are treating male and female employees differently and we analyze the gender match between supervisor and employee. Our empirical results show that the divergence in training duration between females and males can mainly be attributed to differences in characteristics (such as wage, working time or hierarchical status). Although birth of children and child care are a plausible explanation of the age pattern of the gender training gap, we cannot find evidence for prebirth training effects and for post birth catching-up effects, which could explain the training investment at higher ages. Furthermore, including supervisor-fixed effects cannot explain the gender differences in training. Supervisors assign more training to all employees if they participate more in training themselves. As companies want women to work more and employees of both genders to work longer, they need to adjust training mechanisms to individual demand. Considering the dynamics in training behavior may help to improve the employability of both male and female workers.

Fitzenberger, Bernd and Grit Mühler (2011), Dips and Floors in Workplace Training: Using Personnel Records to Estimate Gender Differences, ZEW Discussion Paper No. 11-023, Mannheim.