Abstract
We use data from the Familias en Acción program in Colombia to examine the spillover or indirect effects of a conditional cash transfer program. Our results show that the program has significant spillover effects: it leads to an improvement in the health of non-targeted individuals in treatment households in terms of both incidence and severity of illness. The benefits are stronger for women and the elderly in the short run and for men in the medium run. Our analysis suggests that these spillovers are driven by increased access to information in the household that creates a public good.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
See Lagarde et al. (2007) for a survey of the direct effects on health of CCT programs.
We use the terms indirect effects and spillovers synonymously. Importantly, we define spillovers as different from unintended impacts. While unintended impacts are defined in the light of outcomes that were not targeted by the program design, spillovers are defined in terms of the effects on individuals who were not targeted in the program design. Adults in the household are not targeted in the program we consider, and the program’s design does not include any incentive to improve the health of non-targeted adult members of the household.
US $20.45 at the 2002 exchange rate.
In Table 8 of this paper, we also show that the program has a positive effect on the likelihood of stunting (child’s height-for-age z-score < − 1).
Around 82% of the titulars were the mother of the child and 95% of the titulars were females.
Contrast this with the PROGRESA program. As Gertler (2004) notes, the program imposes the explicit condition that other family members visit clinics once a year for physical checkups and that all adult family members participate in regular meetings at which health, hygiene, and nutrition issues and best practices are discussed (see Gertler 2004, p. 337). Thus, one could argue that other adults in the household were also targeted beneficiaries of the program. That is not the case in the FA program.
There is a larger literature on across-household spillovers of CCT programs that take the form of gifts or other transfers (Angelucci and De Giorgi 2009), an increase in overall incomes (Angelucci and De Giorgi 2009), learning from peer interaction (Bobonis and Finan 2009; Lalive and Cattaneo 2009), changes in behavior due to changes in social norms (Avitabile 2011), or the desire to behave like the eligible population in the hope that they would become eligible, particularly when the eligibility criteria are not well defined within the community and knowledge spillovers occur (Bobba and Gignoux 2019). There is also evidence that CCTs have significant effects on other domains of life, for example, crime and conflict (Chioda et al. 2016; Crost et al. 2016).
In the Appendix A, we present a simple theoretical framework to show how the three channels could manifest. This is, of course, not to say that these three channels are exhaustive. The program can lead to improvements in the supply of basic health services either as part of the program or as a part of a complementary strategy to expand health services in areas where the program is implemented. Alternatively, the program could provide preferential or facilitated access to services to the eligible. We do not consider these supply-side factors in this paper.
The SISBEN is a proxy means test indicator of economic wellbeing that is used throughout Colombia to target welfare programs. Families were surveyed by the municipal authorities and centrally classified into one of the six categories according to their level of measured poverty. The poorest families were classified in level 1, and the richest in level 6.
In the main regression results we present the difference-in-difference estimates, controlling for baseline observables, bearing in mind that there is a potential for the estimates to be biased. To analyze the extent of this (potential) bias, we examine the robustness of the results using propensity score matching in the comparison of treatment and control households. These results are discussed in Section 4.4.4.
See http://www.dnp.gov.co for more details.
We include the number of visits to complete the interview, the number of enumerators to complete the interview, the number of supervisors of the enumerators and if the interview was incomplete as measures of the quality of the interview. We also include dummies for the supervisor code and the percentage of attrition in the municipality. See Fitzgerald et al. (1998) for more on this methodology.
For example, individuals who are more educated, wealthier and from socially advantaged groups, are typically more aware of the limitations imposed on them by their health status and are more likely to report themselves (and their family) as being of poor health. This is known as the cultural conditioning problem.
We also examine whether the results are driven by heterogeneity of average health in the household at the baseline. We divide the sample into two groups: households that were in better health at the baseline and households that were in poorer health. We define better and poorer health accordingly to the average household health relative to the sample mean (proportion of household members reporting ill or reporting hospitalization); those with average < mean for the sample (better health at baseline) and those with average ≥ mean for the sample (poorer health at baseline). The regression results for the two sub-samples are presented in Appendix Table 9: columns 1–4 for households with better health at baseline and columns 5–8 for households with poorer health at baseline. We include two different ways of classifying households: in panel A according to the average health of adults aged 18–59 and panel B according to the average health of the elderly, 60 and higher. The results in panel A of Table 9 show that the results in Table 2 are not driven by average health of adults at baseline. However, the corresponding results in panel B of Table 9 imply that the results are driven by households with elderly in poor health at the baseline.
We find the level of education is, in general, higher for men than for women; and this pattern is stronger for the elderly.
The set of observable characteristics used to compute the propensity score are identical to the ones used by Attanasio et al. (2010).
Matching households with the same p(x) assumes that assignment to treatment or control is random for individuals with the same propensity score. However, this method relies on the assumption of no significant differences between treatment and control households in terms of the unobservable characteristics.
We are unable to run the corresponding regressions by age because of sample issues.
The specifications in columns 2 and 4 are, therefore, problematic in a difference-in-difference setting as they compare changes in different time periods.
Ver Ploeg (2009) writes that it is not possible to tell whether this is due to increased food benefits that are then shared with the non participating children in the family or whether the income offset by the WIC benefits is used to improve the diets of nonparticipating members with other foods (p. 425).
In all regressions we control for labor supply including the number of hours worked. This helps us to isolate the program income effect from any labor supply income effect. The IFS-Econometria-SEI (2006) program report shows an increase in the job market participation by adults but no program effect on the number of hours worked.
Stunting (or more generally height-for-age z-score) reflects the cumulative effect of under-nutrition and infections since birth and therefore, it can be interpreted as an indicator of poor environmental conditions. We define a child to be stunted if his/her height-for-age z-score is less than − 1 The height-for-age z-score is calculated using the WHO Child Growth standards (WHO 2019).
References
Angelucci M, De Giorgi G (2009) Indirect effects of an aid program: how do cash transfers affect ineligibles’ consumption? Am Econ Rev 99(1):486–508
Attanasio O, Battistin E, Fitzsimons E, Vera-Hernandez M (2005) How effective are conditional cash transfers? Evidence from Colombia. Technical report, Institute of Fiscal Studies
Attanasio O, Fitzsimons E, Gomez A, Gutierrez MI, Meghir C, Mesnard A (2010) Children’s schooling and work in the presence of a conditional cash transfer program in rural Colombia. Econ Dev Cult Chang 58(2):181–210
Attanasio O, Fitzsimons E, Gomez AN (2005) The impact of a conditional education subsidy on school enrolment in Colombia. Technical report, Institute of Fiscal Studies
Attanasio O, Gomez LC, Gomez Rojas A, Vera-Hernandez M (2004) Child health in rural Colombia: determinants and policy interventions. Econ Hum Biol 2(3):411–438
Attanasio O, Gomez LC, Heredia P, Vera-Hernandez M (2005) The short-term impact of a conditional cash subsidy on child health and nutrition in Colombia. Technical report, Institute of Fiscal Studies
Attanasio O, Mesnard A (2006) The impact of a conditional cash transfer programme on consumption in Colombia. Fisc Stud 27(4):421–442
Avitabile C (2011) Spillover effects in healthcare programs: evidence on social norms and information sharing. CSEF Working Papers 271 Centre for Studies in Economics and Finance (CSEF), University of Naples, Italy
Baez JE, Camacho A (2011) Assessing the long-term effects of conditional cash transfers on human capital: evidence from Colombia. Technical report, World Bank
Baird S, McIntosh C, Özler B (2011) Cash or condition? Evidence from a cash transfer experiment. The Quarterly Journal of Economics 126 (4):1709–1753
Bastagli F, Hagen-Zanker J, Harman L, Barca V, Sturge G, Schmidt T (2016) Cash transfers: what does the evidence say? A rigorous review of programme impact and of the role of design and implementation features, Report, ODI
Behrman JR, Deolalikar AB (1988) Health and nutrition. In: Chenery‡H, Srinivasan T (eds) Handbook of development economics, Volume 1 of Handbook of Development Economics, Chapter 14, Elsevier, pp 631–711
Behrman JR, Hoddinott J (2005) Programme evaluation with unobserved heterogeneity and selective implementation: the Mexican PROGRESA impact on child nutrition. Oxf Bull Econ Stat 67(4):547–569
Bobba M, Gignoux J (2019) Neighborhood effects in integrated social policies. World Bank Economic Review Forthcoming
Bobonis GJ, Finan F (2009) Neighborhood peer effects in secondary school enrollment decisions. Rev Econ Stat 91(4):695–716
Bustelo M (2012) Who else is benefiting from Conditional Cash Transfer Programs? Indirect effects on siblings in Nicaragua. Ph.D. thesis, University of Illinois at Urbana-Champaign
Chaudhuri A (2009) Spillover impacts of a reproductive health program on elderly women in rural Bangladesh. J Fam Econ Issues 30(2):113–125
Chioda L, De Mello JMP, Soares RR (2016) Spillovers from conditional cash transfer programs: Bolsa Familia and crime in urban Brazil. Econ Educ Rev 54:306–320
Crost B, Felter JH, Johnston PB (2016) Conditional cash transfers, civil conflict and insurgent influence: experimental evidence from the Philippines. J Dev Econ 118:171–182
DFID (2011) Cash transfers evidence paper. Technical report DFID Evidence Paper Policy Division, London, UK
Economist T (2010) Give the poor money
Fiszbein A, Schady NR (2009) Conditional cash transfers: reducing present and future poverty. The World Bank
Fitzgerald J, Gottschalk P, Moffitt R (1998) An analysis of sample attrition in panel data: the Michigan Panel Study of income dynamics. J Hum Resour 33(2):251–299
Fitzsimons E, Mesnard A (2008) Are boys and girls affected differently when the household head leaves for good? Evidence from school and work choices in colombia. Technical report, CEPR Discussion Paper DP7040
Gertler P (2004) Do conditional cash transfers improve child health? Evidence from PROGRESA’s control randomized experiment. American Economic Review Papers and Proceedings 94(2):336–341
Gertler P, Boyce S (2001) An experiment in incentive-based welfare: the impact of PROGRESA on health in Mexico. Technical report, University of California, Berkeley
IFS (2004) Baseline report on the evaluation of Familias en Accion. Technical report, Institute of Fiscal Studies
IFS-Econometria-SEI (2006) Evaluación de impacto del programa familias en acción: Informe final - diciembre de 2006 (Impact evaluation of the familias en accion program, Final report - December 2006, Technical report, IFS-Econometria-SEI
Kazianga H, de Walque D, Alderman H (2013) School feeding programs, intrahousehold allocation and the nutrition of siblings: evidence from a randomized trial in rural Burkina Faso. Journal of Development Economics https://doi.org/10.1016/j.jdeveco.2013.08.007
Lagarde M, Haynes A, Palmer N (2007) Conditional cash transfers for improving uptake of health interventions in low- and middle-income countries. Journal of American Medical Association 298(16):1900–1910
Lalive R, Cattaneo MA (2009) Social interactions and schooling decisions. Rev Econ Stat 91(3):457–477
Maluccio J, Flores R (2005) Impact evaluation of a conditional cash transfer program: the Nicaraguan Red de Proteccion Social. Technical report, IFPRI, Research Report 141, Washington D.C
Rosenbaum P, Rubin D (1983) The central role of the propensity score in observational studies for causal effects. Biometrika 70:41–55
Velez CE, Castano E, Deutsch R (1998) An economic interpretation of Colombia’s SISBEN: a composite welfare index derived from the optimal scaling algorithm. Technical report, Inter American Development Bank
Ver Ploeg M (2009) Do benefits of U.S. food assistance programs for children spillover to older children in the same household? J Fam Econ Iss 30:412–427
WHO (2019) Nutrition Landscape Information System (NLIS) country profile indicators: interpretation guide. Guide, World Health Organization
Acknowledgment
We would like to thank the editor, Shuaizhang Feng, and two anonymous referees of this journal, Sarah Baird, Paul Christian, Juan Miguel Gallego, Bansi Malde, Manuel Ramirez and seminar and conference participants at the Centre for Health Economics at Monash University, the Universidad del Rosario, the Australasian Econometric Society Meetings, the Australian Health Economic Society Meeting and the NEUDC Meetings for their comments and suggestions. The usual caveat applies.
Author information
Authors and Affiliations
Corresponding author
Additional information
Responsible editor: Shuaizhang Feng
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Theoretical Framework
To explain the mechanisms by which a spillover effect might result, we use a stylistic unitary household model (see Behrman and Deolalikar 1988, Chaudhuri 2009). Consider a household with n members. The utility function that defines the preferences of the household is well behaved and can be written as:
where Uj is the utility of the j th household, Hij represents the vector of the health of individuals i = 1, 2, … , n in household j and Zij represents the vector of health inputs and Xij represents the vector of all other consumption goods of household members. Utility maximization is subject to the household budget constraint and the health production functions of all the individuals in the household.
The health production function of the household members can be written as:
Health production within the household depends on the use of health inputs (Zij), consumption of all other goods (Xij), household public good (Wj), health of all other members in the household excluding oneself (H−ij) and all the observed and unobserved endowments of the household (μ). We subdivide the household into two groups: the targeted or T members (for example children aged 0–6 who are the direct beneficiaries of the program) and the other or O members of the household who are not the targeted beneficiaries. Also H−ij = {H1j, … , Hi− 1j, Hi+ 1j, … , Hnj} and \(H_{ij}\in [H_{ij}^{T},H_{ij}^{O}]\). Health inputs (Zij) depends on health inputs provided by the FA program (zFA) and private health inputs (zP), so that we can write
Since zFA is only available to targeted individuals residing in the treatment municipalities, it is a function of the health program (F). Note that the health and nutrition component of the FA program involves a lump sum payment to the household, irrespective of the number of targeted individuals. Hence, \(z_{j}^{FA}(F)\) is defined at the household level. Likewise household public good (W) is also a function of F, generated when the program is present in the household. Define F = 1 when the program is available (for households with targeted individuals in the treatment municipalities) and F = 0 if otherwise. Then \(z_{j}^{FA}(F)=0~\text {if}~F=0\) and \(z_{j}^{FA}(F)>0~\text {if}~F=1\). Likewise Wj(F) = 0 if F = 0 and Wj(F) > 0 if F = 1.
The household budget constraint when Y is the pooled household income, \(p_{z^{P}}\) and px are prices of the private health inputs and consumption goods respectively can be written as:
Maximizing utility (given by Eq. (A.1)) subject to the production constraints (given by Eqs. (A.2) and (A.3)) and the budget constraint (given by Eq. (A.4)), the reduced form demand functions for health inputs, consumption and outcome variables can be written as:
Program intervention (through F) that changes any of the right-hand side variables will change the allocation of resources and outcomes within the households to conform to the optimizing allocation. The impact of the program on the targeted and non-targeted population can therefore be written as:
The focus of this paper is on spillovers, and therefore, we are interested in the effects captured through Eq. (A.7). The first two terms \([(\frac {\partial {H^{O}}}{\partial {X_{ij}}})(\frac {\partial {X_{ij}}}{\partial {F}})+(\frac {\partial {H^{O}}}{\partial {z_{ij}^{P}}})(\frac {\partial {z_{ij}^{P}}}{\partial {F}})]\) denote the income effect, the third term \([(\frac {\partial {H^{O}}}{\partial {W_{j}}})(\frac {\partial {W_{j}}}{\partial {F}})]\) denotes the household public good effect and the last term \([(\frac {\partial {H^{O}}}{\partial {H_{-ij}}})(\frac {\partial {H_{-ij}}}{\partial {F}})]\) denotes the contagion effect. Note that the total effect on the targeted individuals (Eq. (A.6)) has an additional term which is the direct effect of the program on those targeted (given specifically by the program requirement—regular attendance and check-ups in the health clinics).
Appendix Tables
Rights and permissions
About this article
Cite this article
Contreras Suarez, D., Maitra, P. Health spillover effects of a conditional cash transfer program. J Popul Econ 34, 893–928 (2021). https://doi.org/10.1007/s00148-020-00809-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00148-020-00809-y