Endogénéité du statut du répondant dans les enquêtes sur la santé. Quelles implications pour la mesure des besoins d'aide ?

Most of population health surveys appeal to proxy respondents, when surveyed persons with poor health are not able or do not want to answer the questionnaire by themselves. Yet, it may give rise to a strong endogeneity problem since the use of a proxy respondent is likely to be linked to subjects? health status. That paper aims at giving a convenient and tractable way to cope with the twofold potential endogeneity of the respondent status in the estimation of two probit equations concerning the need for care with daily activities and unmet or undermet need reported by people aged 60 years and over living in the community. Unlike previous papers dealing with that topic, evidence is found that proxy respondents do not entail any subjectivity when they assess old people?s need for care but they seem to overestimate undermet and unmet need. Classification JEL : C35 ; I12 ; J14.