Towards a Precise and Accurate Analysis of the Situation of Two-Home Children in France from New Demographic Panel Data

Laurent Toulemon, Institut National d’Études Démographiques
Benjamin Marteau, Institut national d’études démographiques (INED)
Sebastien Durier, INSEE

Among children from separated parents, most live with their mother, but the proportion of children commuting between the two parental homes is not negligible. The aim of the current paper is to construct consistent estimates from the available data sources on children from separated parents, and to provide, for the first time, a description of family situation of children commuting between both parental homes based on the complete information combining both households. Except in case of specific surveys base on a sample of separated couples (Kitterød and Lyngstad 2011), it is almost impossible to combine information for all households where the children usually live in.

In France, three sets of data may be used to study multiple residence of children: household surveys including a question on another “usual residence” for each member of the household (Toulemon and Denoyelle, 2012). Second, the 2011 Family and Housing Survey, conducted within the census (Lapinte and Buisson, 2017). Third, the Demographic panel which includes information from the census, vital statistics and, since 2011, tax data, and allows estimating double counts at census, as well as describing the household composition and the family situation of inhabitants in their both households (Toulemon et al. 2016).

Our aim is threefold. First, to reconcile estimates of multiple residence from these three data sources. Second, to accurately describe the social differentials in the probability to live in more than one home among children of separated parents. Third, to take benefit of situations where the same children are present twice in the census the same year to accurately describe their family situation in both households.


Introduction

Among children from separated parents, most live with their mother, but the proportion of children commuting between the two parental homes is not negligible. The aim of the current paper is to construct consistent estimates from the available data sources on children from separated parents, and to provide, for the first time, a description of family situation of children commuting between both parental homes based on the complete information combining both households.

Most data do not provide such information: in family and household surveys, assuming that sample size is sufficient, a set of questions can be introduced in order to grasp some information of “another usual residence”, but it is almost impossible to ask a parent of a partially coresident child questions on the family situation in the “the other home” of that child, as adults are reluctant to speak about the current family and couple situation of their previous partner, who is the other parent. Register data do not consider the possibility for a child to live in two places at the same time. Thus, except in case of specific surveys base on a sample of separated couples (Kitterød and Lyngstad 2011), it is almost impossible to combine information for all households where the children usually live in.Data and methods

In France, three sets of data may be used to study multiple residence of children. First, since 2004, most of household surveys conducted by the INSEE, the French institute of statistics, include with the list of household members a question on another “usual residence” for each member of the household (Toulemon and Denoyelle, 2012). Second, the 2011 Family and Housing Survey, conducted within the census, included similar questions on multiple residence (Lapinte and Buisson, 2017). Third, the Demographic panel conducted by the INSEE since 1968 includes many information from the census, vital statistics and, since 2011, tax data. Since 2004 the census has been transformed into census annual surveys, and the demographic panel includes all information coming from the census, including double “echoes” when the same individual is included twice the same year in the census, in two different dwellings. This allows estimating double counts at census, as well as describing the household composition and the family situation of inhabitants in their both households (Toulemon et al. 2016).

Our aim is threefold. First, to reconcile estimates of multiple residence from these three data sources, especially on the issue of bias in the prevalence estimate due to double counts (Toulemon and Pennec, 2009). Second, to accurately describe the social differentials in the probability to live in more than one home among children of separated parents. Third, to take benefit of situations where the same children are present twice in the census the same year to accurately describe their family situation in both households: is the parent living as couple, how many brothers and sisters are present, how many of them are also commuting between the two homes.Preliminary results

The issue of double count is far from being negligible: according to INSEE, some 900,000 children aged 1-17 are commuting between both parental homes, but we estimate from the Demographic panel that 400,000 are counted twice in the census, leading to a severe overestimation of the frequency of these situations from the census data, as the logical consequence is that the accurate figure is 500,000, among whom 400,000 are included twice in the census, leading to the false estimate of 900,000. This means that 400,000 children are falsely included in those living in a one-parent family (2,450,000) or in a reconstituted family with a step-parent (945,000). We have already checked that in almost all cases of multiple presences in the census, children live with their mother in one household, with their father in the other.

Preliminary results from surveys have shown that 15% of children with separated parents (2.9% of all children 0-17) commute between parental homes, while 74% of children are living with their mother and 11% with their father only. For these children, multiple residence is much more frequent among children living with their father than among children living with their mother (59% vs. 17%; in such estimates, children with multiple residence are logically counted twice, once with each parent). Children whose co-resident parent (Where the survey took place) are living as a couple (especially LAT) are more often commuting. Multiple residence is more common among children with the most educated parents as well as those with parents working as senior or intermediate manager. These results will be updated and completed with recent data from the Demographic panel, where the same proportion of children (2.9%) is estimated to be counted twice.

Presented in Session 1100: Families and Households