Alternative measures of the explanatory power of multivariate probit models with continuous or ordinal responses

In 1986 German federal parental leave and benefit policy was expanded in several ways, extending the potential duration of leave from six to ten months and paying child-rearing benefits to all new mothers regardless of their employment status before childbirth. The potential duration has increased four times since 1986 and stood at 18 months in 1991 and three years starting in 1992. This study uses log-wage difference regressions to examine the effect of leave taken by the mother on wage growth for two 5-year periods, 1984-1989 and 1989-1994. In each of the five-year periods, taking maternity leave was found to have a significant negative effect on wage growth. Point estimates imply that each month of maternity leave reduced wage growth by 1.5 percent over five years. In addition, for the second five-year period only, mothers experienced lower wage growth if they chose to stay at home rather than return to work when the allowable leave period expired: from 1989 to 1994, a half-year out of the labor force after the end of the leave period lowered wage growth by an additional 15 percent over five years.