Reconsidering the labeling effect for child benefits: evidence from a transition economy