Impact of first birth career interruption on earnings: evidence from administrative data