The Hungarian Central Statistical Office recorded data about support provided to adult children in two representative national surveys entitled “Housing Conditions 2003” and “Housing Conditions 2015”. The 2003 database contains 8781 dwellings while the 2015 survey recorded data about 9781 dwellings and their inhabitants. Surveys were answered by heads of households in each dwelling and datasets of both surveys available for research contain data about support provided by parents to their child(ren) living independently in obtaining their dwelling, socio-economic characteristics of the household and detailed information about their dwelling.

Although the survey is representative of dwellings and not households, data can be used to measure parental support with high accuracy. Based on the 2011 Census, 8,6% of households share a housing unit with at least one other household. Since more than two households occupying a single dwelling is rather rare, the dwelling-based database does not contain data about only cca. 4% of households sharing a dwelling with other households. Among households sharing a dwelling with other households there are almost 10% more households with at least one member under 29 years and 11% more households living alone (KSH 2013, pp. 113–116). This suggests that among these households, young people in co-residence with peers are over-represented, while people with adult children living independently are underrepresented. However, such differences are so insignificant that we can consider results gained from the dwelling-based database accurate without weighting for multi-household dwellings.

The other weakness of the surveys, stemming from their focus on dwellings and not people, is that people not living in dwellings remained outside the sample. People living in institutions such as homeless shelters or elderly homes are excluded from the two surveys. Since questions about parental support were answered by heads of households, answers of elderly parents living with their child’s family are also missing from the survey.

The above factors already make it hard to define the share of parents supporting their children, however, it is further complicated by the lack of information about the share of divorced parents. The survey records support by two divorced parents forming separate households as two responses while that of a couple only counts as a single response. Since differences stemming from the lack of consideration of multi-household dwellings are minor, data are largely representative of the share of parental households providing support, but not of parents providing support.

Another issue that needs to be taken into consideration is the comparability of data from the two waves of the survey. Abrupt demographic changes can induce variation in the ratio of parental households providing support. The most significant such change occurring between 2003 and 2015 is undoubtedly the increasing ratio of divorced individuals. Data recorded in the 2001 Census and the 2016 Microcensus (around the two waves of the housing survey) show that the share of divorced individuals increased by 20%, from 8,9% to 10,7% in the period (Makay and Szabó, 2019, p. 37). Although the increase may seem remarkable at first glance, even if we do not take into consideration that, in many cases, after divorce only one of the parents supports their child(ren); and that most divorced people form new households with partners who (or with whom together they) also have children; and we consider all data about divorced parents as evidence about parental support recorded twice, we still get only 1% observational error which is rather small. With the above factors considered, the impact of the increase of divorce on a prospective change in parental support is even milder. Hence, data from the two waves are comparable.

Since the surveys focused on housing conditions rather than the characteristics of individuals living in the dwellings, datasets contain only a limited amount of data about heads of households, further, data were not recorded about children living elsewhere to which support was potentially provided. For this reason, in the following only characteristics of parents can be examined, but no parent-child dyads.

As expounded in the literature review, the international literature highlights two qualities of adult children that have an impact on the occurrence of parental support: merit and need. The lack of variables denoting characteristics of children prevent the scrutiny of merit and need. However, need was partly examined by the surveys in the case of non-provision of support as the survey differentiated between parental households not support their children due to the lack of child(ren)'s need and parents lacking means to support their child(ren) (see below). In this sense, some evidence is available about the lack of the child’s need or the lack of parental resources.

The question about the provision of parental support is the following (in the 2003 survey, question III/5; in the 2015 survey, question VII/5):

 

 

Since co-res pertains to temporary accommodation provided in the past to adult children living independently, it does not perfectly overlap with intergenerational co-residence which also includes long-term and present-day co-residence. For this reason, the extent of the provision of temporary accommodation is expected to be lower than actual intergenerational co-residence. However, its determinants are expected to be similar to those of intergenerational co-residence, but should also be evaluated carefully.

The meaning of the two options available for non-supporting parent households are somewhat ambiguous. The option “did not need to support” suggests one’s (financial, physical, housing, etc.) conditions allow them to support their child(ren), but so far their support has not been needed by their children. In turn “could not support” seems to mean one’s conditions do not allow them to support their child(ren). In most cases, the latter refers to problems of affordability of support while the former the lack of opportunity or the lack of acceptance of support on the child(ren)’s side.

As mentioned earlier, the survey primarily records evidence about housing units and data about the inhabitants of dwellings are limited compared to a population census. Still, the two surveys recorded a wide range of characteristics of responding households and heads of households that could serve as independent variables, and are sufficient for the current inquiry. Datasets contain variables such as per capita income, settlement type, tenure, number of inhabitants in the dwelling, the modified Erikson-Goldthorpe-Portocarero (EGP) occupational classification of the head of the household and the year of provision of parental support that are expected to significantly impact the provision of housing-related parental support.

First, in order to test H1 the development of support types is examined through periods of housing system formation identified in Section 4.1 based on Hegedüs and Tosics (1992b), Hegedüs and Somogyi (2016), Augustyniak et al. (2019) and Pósfai (2018):

Determinants of the provision of parental support and its different types are explored through logistic regression performed separately on major types of parental support in the two surveys measured by dependent categorical variables, and independent categorical and continuous variables denoting socio-economic and housing conditions of the parental household.