Illness is one of the greatest and least predictable shocks to the poor. Households often rely on costly and inefficient coping strategies, as they are unable to insure against illness.1 In Mexico a typical uninsured household with three dependents spends up to 12 hours per week caring for someone who is ill. Given the traditional role of women in Mexican families, the increased demand for caregiving during spells of illness is borne almost exclusively by women. These shocks have profound repercussions on the employment prospects of women, for whom family responsibilities account for as much as 44% of all firing or quitting events.
In a recent paper,2 I show that the provision of publicly subsidised health insurance in Mexico has led to an increase in the labour supply. This increase can be as high as one-quarter percent of GDP. This labour supply effect is important for two reasons:
- The labour market response is concentrated among women for whom insurance reduces the burden of caring for sick dependents, and
- It allows me to show that the provision of insurance need not entail an efficiency loss in the labour market, as suggested by past papers.
The nuts and bolts: Deriving causal estimates
The key challenge in deriving estimates of the causal impact of health insurance on the labour market is the endogeneity of access to health insurance. I overcome this problem by exploiting the variation created by the municipal-level rollout of Mexico’s Seguro Popular (SP). This program provides free health insurance and improved access to health services to those not covered by employer-based schemes. SP provides a unique opportunity to learn from a large-scale intervention. SP currently insures 52.6 million individuals. This means that, in terms of affiliates, SP covers more people than the Affordable Care Act in the US.
I use a difference-in-differences design that compares changes in the labour market outcomes of individuals that reside in municipalities already reached by SP with individuals in municipalities not yet reached by SP. I isolate plausibly exogenous variation, focusing on two years in the middle of the rollout period for which control municipalities are likely to provide good counterfactuals;3 that is, accurate estimates of what would have happened to labour market outcomes had the program not been implemented.
Health insurance increases labour supply
My main finding is that SP increases labour supply by reducing the exit flow from employment, that is, by helping workers stay employed.4 This labour market response is, however, not contemporaneous. I find that SP reduces the exit flow from employment roughly one to two years after its introduction. This time profile is consistent with the idea that individuals are learning about the value of the program, and that the program is operating through a health channel.5
SP enables women to stay employed
By reducing the frequency and duration of illness, particularly among children, SP enables women to stay employed.6 I find that the labour supply response triggered by SP is specific to women. In particular, women not fully specialised in caregiving, but who are likely to care for dependents during spells of sickness, are the ones who benefit the most from SP. I also find that SP leads to an increase in the hours worked by this group of women and to an almost symmetrical decrease in the hours spent caregiving.7
SP leads to an overall efficiency-enhancing response in the labour market
Given that programs like SP are modeled as transfers, it is often argued that their introduction will encourage workers to seek jobs without employer-based health insurance. In Mexico, these jobs are informal. Policymakers are understandably concerned about growing informality in view of the loss in tax revenues and the potential loss in overall efficiency associated with informal jobs.8
In keeping with past papers,9 I find that SP increases the relative size of the informal sector. However, this finding does not imply a trade-off in the labour market. The new insight is that the increase in informality is not driven by workers moving to informal jobs, but by informal workers being retained in the labour force.
The intuition behind these findings is that better-off workers who are voluntarily informal place little value on SP, while marginalised women, trapped in informal jobs, benefit from SP because the program delivers a valuable reduction in the exposure of their household to health shocks.
Two invitations for policymakers
- Concern about SP reallocating labour to the informal sector is important, but it should be weighed against the positive impact SP has on labour force participation.
- The design of social insurance should leverage the way families operate in order to achieve complementary objectives. The case of SP highlights the potential of insurance to both improve health and empower women.
Author's note: If you are interested in the details of this article, please download my paper at www.ocf.berkeley.edu/adv. If you have feedback or suggestions, please send me an email at [email protected] or follow me on twitter @alejandrodvs
Alcaraz, C, D Chiquiar, M J Orraca, and A Salcedo (2012), “The Effect of Publicly Provided Health Insurance on Academic Performance in Mexico”, Technical Report.
Bernal, P (2014), “Essays on the effect of health insurance on utilization and health”, PhD dissertation, The University of Chicago.
Bosch, M and R M Campos-Vazquez (2010), “The trade-off’s of social assistance programs in the labour market: The case of the “Seguro Popular” program in Mexico,” Serie documentos de trabajo del Centro de Estudios Económicos 2010-12, El Colegio de México, Centro de Estudios Económicos.
Bosch, M, M B Cobacho, and C Pages (2012), “Taking Stock of Eight Years of Implementation of Seguro Popular in Mexico”, Washington, DC, United States: Inter-American Development Bank. Mimeographed document.
Del Valle, A (2014), “From Caring to Work: The Labour Market Effects of noncontributory health insurance”, Job Market Paper.
Gertler, P and J Gruber (2002), “Insuring Consumption against Illness”, The American Economic Review.
Levy, S and N Schady (2013), “Latin America’s Social Policy Challenge: Education, Social Insurance, Redistribution”, The Journal of Economic Perspectives.
Pfutze, T (2014), “The Effects of Mexico’s Seguro Popular Health Insurance on Infant Mortality: An Estimation with Selection on the Outcome Variable”, World Development.
1 See, for example, Gertler and Gruber (2002).
2 See Del Valle (2014).
3 The paper makes a strong claim of internal validity because it shows that those municipalities reached in the middle of the rollout period had similar pre-program trends across all outcomes, that SP was not anticipated, and that the municipalities were similar at baseline in terms of observables. I additionally use lights-by-night satellite imagery and detailed records of local government expenditures, in order to account for two key time-varying confounders: changes in the level of economic activity, and political targeting of the program.
4 I am able to make this precise statement because I use individual level labour market transitions derived from the labour force survey rotating panel.
5 Bosch and Campos-Vazquez (2010) show that increases in health inputs occur in this same timeframe.
6 While the effect of insurance on morbidity is notoriously difficult to identify, the medical literature shows that SP gives rise to large increases in health care utilization (Bernal 2014), large reductions in infant mortality (Pfutze 2014), and improvements in school performance (Alcaraz et al. 2012).
7 Other mechanisms could potentially create this pattern of effects. See Del Valle (2014) for a discussion of the role of own health effects, delayed child bearing, reallocation of time among household members, and the freeing up of other household resources.
8 See, for example, Levy and Schady (2013).
9 See Bosch et al. (2012) for a review of this literature.