Forum

Using novel data from Luxembourg and comparative evidence from the UK and Australia, we doc­ument systematic under-reporting of children's socio-emotional difficulties by parents - particularly for daughters. A distinctive feature of our study is the joint measurement of first-order beliefs (how parents think their child feels) and second-order beliefs (how parents think their child reports feeling), allowing us to disentangle information frictions from evaluative bias. We find that second-order beliefs are sys­tematically biased and that their precision is negatively correlated with the level of distress reported by the child. Parents with more accurate second-order beliefs also provide closer and unbiased first-order asvmaiments, suggesting that the negative difference between parent and child reports may arise from in­formation frictions rather than deliberate minimization. Misperceptions are more common among highly educated and employed parents, and in parent-child pairs with divergent personality traits. Interest­ingly, parents with more accurate priors about general parental under-reporting tend to show greater divergence from their child's report in their rust-order beliefs. An information treatment improves be­liefs among parents who already hold priors about systematic discrepancies between parent and child reports, demonstrating the malleability of parental beliefs to light-touch interventions. We also provide exploratory evidence that receiving accurate information can influence parents intentions to invest in their children's human capital.