VoxEU Column Institutions and economics Poverty and Income Inequality

Insurance in extended families

The significance of informal sources of insurance against income risk has important consequences for the design of social insurance programs. A particular concern is that public programs simply crowd out informal institutions. This column uses US household data to investigate whether the extended family acts as such an informal institution. Although there is a large potential for the family to insure against income shocks, no such insurance occurs. 

The ability of individuals and households to absorb income and resource shocks has substantial implications for their welfare and is critical to the design of policy interventions. For instance, a social safety net may be of crucial value for households that bear the full brunt of income shocks, while less necessary for households who can smooth out these shocks through other means. Understanding the available mechanisms through which households combat income risk, their efficacy, and the interactions between them, is central to the policy debates around social insurance programs.

The extended family is one such avenue through which households may share risk. Due to its ubiquity – almost everyone has a family – this informal mechanism has policy implications across many different contexts, from families in small agricultural villages in rural India to families in post-industrial Europe. In the US, where politicians and policymakers heatedly debate the merits of social programs, one common argument is that social insurance crowds out informal institutions such as the family.

Evidence from the US

Using intergenerational data from extended families in the US, economists have made headway in understanding the role of the family. Recent work by Blundell et al (2008) found that households are insulated from income risk above and beyond the impact from social programs and their own ability to smooth over shocks through savings (‘excess smoothness,’ as was first pointed out in a seminal paper by Campbell and Deaton 1989). At the same time, research by Altonji et al (1992, 1996) found that extended families do not completely insure their members against income risk.

In a new working paper, we complement both of these sets of studies by examining two new questions:

  • What is the potential for the family to insure its members?
  • To what extent does the family exploit this potential to share risk?

To answer these questions, we use over 30 years of longitudinal income and consumption data from households across the US in the Panel Study of Income Dynamics. The key feature of the data that we exploit to answer these questions is the full set of intergenerational links between parents and grown children.

Methodologically, we develop a theoretically motivated decomposition of income shocks. This method allows us to isolate income shocks that are aggregate to the family from shocks that are statistically idiosyncratic to individual households within the family. Since ‘family-aggregate’ shocks act as resource shocks to the entire family, they are by definition uninsurable by the family. In contrast, idiosyncratic shocks, which only affect the distribution of income between households but not the overall resources of the family, are insurable by the family.

When we apply this method to our data, we find that over 60% of income shocks are idiosyncratic within the extended family. This implies a large potential for the family to have a non-trivial impact on the transmission of income shocks into consumption and therefore well-being. Put another way, the family can potentially mitigate 60% of household income risk.

But does it? We answer this second question by applying a similar decomposition method to consumption changes and test whether idiosyncratic income shocks, which are potentially insurable by the family, are better insured than family-aggregate income shocks, which are not insurable by the family.

We first find that only 54% of a family-aggregate income shock translates into a consumption change, implying that other sources aside from the family – such as savings (‘self-insurance’) – play an important insurance role against income risk. We also find that 51% of an idiosyncratic income shock translates into a consumption change. The insignificant difference between these numbers implies that the family provides no detectable insurance against idiosyncratic risk.

There are several reasons why families may not take advantage of informal insurance opportunities, such as an inability to monitor the true economic circumstances of family members, or an inability to enforce an informal ‘insurance contract’ (see theoretical contributions by Attanasio and Pavoni 2011 and Ligon et al 2002). Our work finds suggestive evidence that the ability to monitor family members may play a role, but this is an important area for future research.

Concluding remarks

In the design of social insurance programs, policymakers often worry about how these programs could affect household behaviour, such as creating work disincentives or crowding out other sources of insurance. In the US, it is well known that public welfare programs provide incomplete insurance. Our research shows that private informal networks – in particular, the extended family – do not fill these gaps in insurance.

Thus it is not the case that extending public insurance programs necessarily would crowd out private networks. However we caution against extrapolation to the converse – our results do not speak to the crowd-in nature if social insurance programs became less generous, and we view this as an important policy question for future investigation. More broadly, our decomposition method easily generalises to any grouping of households that may share risk, such as households in small rural villages or church communities; as such, it can contribute to policy research in a variety of contexts besides families in the US.


Altonji, J G, F Hayashi and L J Kotlikoff (1992) “Is the extended family altruistically linked? Direct tests using micro data”, The American Economic Review, 1177–1198

Attanasio, O P and N Pavoni (2011) “Risk sharing in private information models with asset accumulation: Explaining the excess smoothness of consumption”, Econometrica, 79(4): 1027–1068.

Blundell, R, L Pistaferri and L Preston (2008) “Consumption inequality and partial insurance”, The American Economic Review, 98(5): 1887–1921.

Campbell, J J and A Deaton (1989) “Why is consumption so smooth?”, The Review of Economic Studies, pages 357–374.

Hayashi, F, J Altonji and L Kotlikoff (1996) “Risk-sharing between and within families”, Econometrica, 64(2): 261–294.

Ligon, E, J Thomas, and T S Worrall (2002) “Informal insurance arrangements with limited commitment: Theory and evidence from village economies”, Review of Economic Studies, 69(1): 209–44.

1,155 Reads