Biomarkers and long-term labour market outcomes: the case of creatine

I evaluate the impact of the UK Working Time Regulations 1998, which introduced mandatory paid holiday entitlement. The regulation gave (nearly) all workers the right to a minimum of 4 weeks of paid holiday per a year. With constant weekly pay this change amounts effectively to an increase in the real hourly wage of about 8.5% for someone going from 0 to 4 weeks paid holiday per year, which should lead to adjustments in employment. For employees I use complementary log-log regression to account for right-censoring of employment spells. I find no increase in the hazard to exit employment within a year after treatment. Adjustments in wages cannot explain this result as they are increasing for the treated groups relative to the control. I also evaluate the long run trend in aggregate employment, using the predicted treatment probabilities in a difference-in-difference framework. Here I find a small and statistically significant decrease in employment. This effect is driven by a trend reversal in employment, coinciding with the treatment.