Skip to main content

Perception of HIV risk and the quantity and quality of children: the case of rural Malawi

Abstract

The empirical literature on the impact of HIV on the quality (Q) and quantity (N) of children provides limited and somewhat mixed evidence. This study introduces individual HIV risk perceptions, as a predictor of mortality, into a Q–N investment model. In this model, higher maternal mortality predicts lower N, while higher child mortality predicts lower Q. Thus, the two effects together make likely negative associations between HIV and both Q and N. Based on longitudinal micro-data on mothers and their children in rural Malawi, our results suggest that higher mothers’ reported HIV risk reduces both child quality, as reflected in children’s schooling and health, and child quantity, when the perceived risk is already moderate or high. The effects are sizable and, in the case of Q (schooling and health), are found for children and teenagers, both boys and girls, while in the case of N, they are found for young and mature women.

This is a preview of subscription content, access via your institution.

Fig. 1

Notes

  1. Studies that report on schooling declines include Evans and Miguel (2007), Yamano et al. (2006), and Case and Ardington (2006). Mishra and et al (2007) analyze malnutrition, while Weinhardt et al. (1999) and Oster (2012) study on changes in sexual behavior.

  2. Some examples include Case and Paxson (2009) on health care, Papageorgiou and Stoytcheva (2008), and Young (2005) on income and Whiteside (2002) on poverty.

  3. Other mechanisms include supply-side variables, such as teacher and health worker availability (Kelly 2000; Bennell 2005), funding (Grassly et al. 2003; Chen et al. 2004), and services (Case and Paxson 2009). Among the demand-side factors, the expectations about future resources, for example, the expectations about medical costs might influence the way individuals make decisions about investments in children.

  4. An evolutionary perspective intuitively suggests a positive relationship between fertility and the probability of death of children. The literature reviewed in this article, however, did not find a clear relationship between HIV and fertility, nor between mortality and fertility.

  5. Fortson (2009), Kalemli-Ozcan (2012), Kalemli-Ozcan and Turan (2011), Juhn et al. (2008), Fink and Linnemayr (2008), and Young (2005) analyzed regional and local HIV rates, as well as own HIV status among sub-Saharan African countries, with mixed results.

  6. Durevall and Lindskog (2008), Noël-Miller (2003), and Ueyama and Yamauchi (2009) find that Malawian HIV rates (or self-reported risk) are negatively associated with fertility rates, with the exception of young women with no children, for whom the probability of a first birth increases.

  7. In a concurrent study, Shapira (2011) examines the impact of perceived HIV risks on fertility, but does not consider child quality.

  8. De Lannoy (2005) presents related qualitative evidence.

  9. The empirical literature on links between HIV rates and child quality focuses on schooling. However, another important investment in children’s human capital is in their health, which may have important long-run effects in poor populations (e.g., Hoddinott et al. 2008, Maluccio et al. 2009, Behrman et al. 2009).

  10. If the mother dies, there are C 2 resources available, for example, for children to consume. The model presented in this section, however, does not include children’s consumption explicitly. But the inclusion of children’s consumption would not change the major lessons learned in this exercise.

  11. However, changes in the relative preferences between Q and N could change the total amount to be invested in children if the marginal utility of investments in children, W’, changes.

  12. The first-order conditions with respect to C 1, Q, and N, respectively, are the following:

    $$ U^{\prime} -\lambda =0 $$
    (4’)
    $$f( {H_{m} } )\left[ {( {1-H_{c} } )W^{\prime} \ast({\beta N^{\alpha} Q^{\beta-1}})} \right]-\lambda ( {P_{q} +P_{qn} N} )=0 $$
    (5’)
    $$f( {H_{m} } )\left[ {( {1-H_{c} } )W^{\prime} \ast ( {\alpha N^{\alpha-1} Q^{\beta}} )} \right]-\lambda ( {P_{n} +P_{qn} Q} )=0 $$
    (6’)

    Equations 5 plus 6 imply Eq. 3. Equations 4 plus 5 imply that \(f(H_{m})[(1-H_{c})W^{\prime }(w_{q} N^{wn} Q^{wq-1})]/(P_{q}+NP_{qq})=U(C_{1})^{\prime }\). If H m rises, then the left side decreases and thus U(C 1)decreases, which implies that C 1 increases. Because of Eq. 3, adjustments in Q and N move in the same direction, so both Q and N decrease. Therefore, the total investment in children decreases. The analysis shows that a rise in H c goes along the same lines.

  13. The 2010 MLSFH data, and in particular the longitudinal linkages of children across waves, were not yet available at the time of writing this paper.

  14. Detailed descriptions of the MLSFH sample selection, data collection, and data quality are provided on the project website http://www.malawi.pop.upenn.edu/, in a Special Collection of the online journal Demographic Research that is devoted to the MLSFH (Watkins et al. 2003) and in a recent follow-up publication that incorporates the 2004 and 2006 MLSFH data (Anglewicz et al. 2009).

  15. Tabulations in De Paula et al. (2010) show that two thirds of individuals receiving HIV-positive test results in 2004 classify themselves in the next round in 2006 as zero or low likelihood of having HIV.

  16. For example, Matovu et al. (2005) and Nyblade et al. (2001) find self-selection in testing in Uganda.

  17. Other studies in this discussion have used “HIV prevalence rate” in the geographical area of residence of the individual. This alternative has advantages and disadvantages compared to ours, so we would say that they complement each other. For example, individuals may have a vague idea of the prevalence in their region, just as they may have a vague idea about the “likelihood of contracting HIV,” as is asked, in different forms, in the surveys. Another example is the role of unobservables given that HIV rates are not distributed randomly in the geographical areas, while our identification strategy is susceptible to endogeneity bias. In general, it is difficult to attribute a measure to specific biases in the methodology, so it is useful to have different methodological approaches to the same problem.

  18. The exact wording of the question depends on the language of the interview. Most of the interviews were conducted in Chitumbuka and Chichewa (District of Rumpi, north Malawi), Chichewa (District of Mchinji, central Malawi), and Chiyao (District of Balaka, south Malawi).

  19. We are not aware of evidence about the better predictive power of these extreme answers; Idler and Benyamini (1997) summarize evidence in relation to self-reports on health.

  20. Recoded as 0, “very good” or “good”; and 1, “poor” or “very poor.”

  21. Indeed, our data do not show significant results for these women.

  22. The significance of the differences in the coefficients of interest between the fixed and the random-effect models are investigated by the Durbin–Wu–Hausman test. The random-error model used to conduct the test also included other regressors: age of the mothers, number of siblings, regional fixed effects (one for each of the three regions included in the survey), and perceived HIV status.

  23. Women with HIV, whether or not they know it, change their behavior for health reasons. In this case, the dynamic does not have an effect on the main idea guiding the article. In this article, the test results are not included as a control because it is not available for all individuals, but the health of the mother and a measure of medical expenses are included as controls. Regardless of the above, of the people who received the test, approximately 5 % tested positive, and excluding them from the statistical analysis does not alter the results.

  24. Q 2 and Q 1, as dependent variables, are taken from MLSFH 2008 and MLSFH 2006. Q 1 and Q 0, as control variables, are taken from MLSFH 2006 and MLSFH 2004. But MLSFH 2004 does not contain the required measure of child health. Thus, the included control for child health ( Q1 and Q0) is the answers to “Is your child ill?” as included in MLSFH 2006 and 2004.

  25. Another possible wealth indicator is household crop production. However, given that the survey was conducted in the dry season, crop accounts have to rely on retrospective reports that probably have considerable measurement error as usually is the case for near-subsistence agriculture. Also, crop production is affected considerably by weather fluctuations that may obscure the relation with longer-run wealth. Nevertheless, crop variables were included in exploratory regression analysis (not shown), although they have little significance.

  26. In Table 3, 0.36 is just the simple average of the coefficients in child education.

  27. For the sake of simplicity, the budget constraint in this section does not contain the P n N and the P n q × Q N components.

  28. MLSFH 2006 respondents report, by selecting a number of beans from zero to ten, their likelihood of dying within the next 1, 5, or 10 years. On average, individuals who reported a high perceived risk of HIV reported a 10 % higher likelihood of dying, compared to individuals who reported no perceived risk of HIV. We also assume that this represents the children’s mortality risk.

  29. Q t / H t − 1 = − C t − 1 × β s × f(H t − 1; b s ), where f(H t − 1; b s ) = (1 − H t − 1)(H t − 1 + 3) / (b s (1 − H t − 1)2 + (2 − H t − 1)).f is a positive function, and decreases with H.

  30. See a brief summary in Datar et al. (2007) and Jayachandran and Lleras-Muney (2009) respectively.

  31. The association between mortality and investments in children quantity (fertility) and quality (education and health) has been discussed for long. A substantial fraction of this body of research is focused on developed countries, for which empirical analysis found that (a) mortality is linked with fewer investments in child education and health (see a brief summary in Datar et al. 2007 and Jayachandran and Lleras-Muney 2009, respectively) and (b) mortality is not clearly linked with fertility. Preston (1978) identified three conceptual mechanisms for the mortality–fertility link, insurance, replacement, and lactation interruption effects, and Montgomery (2000) included the role of perceptions and social definitions, but the empirical counterpart of such associations remains rather weak and poorly understood (Montgomery 2000; Cleland 2001); Cigno (1998) and Strulik (2003) show that there is a negative relationship between mortality and total fertility when mortality is high and a positive relationship when mortality is low. Albanesi and Olivetti (2010) found mixed results for the adult mortality–fertility association in the USA.

References

  • Akbulut-Yuksel M, Turan B (2013) Left behind: intergenerational transmission of human capital in the midst of HIV/AIDS. J Popul Econ 26(4):1523–1547

    Article  Google Scholar 

  • Albanesi S, Olivetti C (2010) Maternal health and the baby boom, NBER Working Papers 16146, National Bureau of Economic Research, Inc.

  • Anglewicz P, Kohler H-P (2009) Overestimating HIV infection: the construction and accuracy of subjective probabilities of HIV infection in rural Malawi. Demogr Res 20(6):65–96

    Article  Google Scholar 

  • Anglewicz P, Adams J, Obare-Onyango F, Kohler H-P, Watkins S (2009) The Malawi Diffusion and Ideational Change Project 2004–06: data collection, data quality and analyses of attrition. Demogr Res 20(21):503–540

    Article  Google Scholar 

  • Arellano M, Bond S (1991) Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Rev Econ Stud 58:277–297

    Article  Google Scholar 

  • Ashenfelter O, Krueger A (1994) Estimates of the economic return to schooling from a new sample of twins. Am Econ Rev 84(5):1157–1174

    Google Scholar 

  • Becker G, Lewis H (1973) On the interaction between the quantity and quality of children. J Polit Econ Part 2 New Econ Approach Fertil 81(2):S279–S288

    Google Scholar 

  • Behrman JR, Rosenzweig MR, Taubman P (1994) Endowments and the allocation of schooling in the family and in the marriage market: the twins experiment. J Polit Econ 102(6):1131–1174

    Article  Google Scholar 

  • Behrman JR, Calderon CM, Preston S, Hoddinott J, Martorell R, Stein AD (2009) Nutritional supplementation of girls influences the growth of their children: prospective study in Guatemala. Am J Clin Nutr 90:1372–1379

    Article  Google Scholar 

  • Behrman JR, Kohler HP, Jensen VM, Pedersen D, Petersen I, Bingley P, Christensen K (2011) Does more schooling reduce hospitalization and delay mortality? New evidence based on Danish twins. Demography 48(4):1347–1375

    Article  Google Scholar 

  • Bennell P (2005) The impact of the AIDS epidemic on teachers in sub-Saharan Africa. J Dev Stud 41(3):440–466

    Article  Google Scholar 

  • Bíró A (2013) Subjective mortality hazard shocks and the adjustment of consumption expenditures. J Popul Econ 26(4):1379–1408

    Article  Google Scholar 

  • Case A, Ardington C (2006) The impact of parental death on school enrollment and achievement: longitudinal evidence from South Africa. Demography 43(3):401–420

    Article  Google Scholar 

  • Case A, Paxson C (2009) The impact of the AIDS pandemic on health services in Africa: evidence from demographic health surveys research program in development studies and center for health and wellbeing. Princeton University, Princeton

    Book  Google Scholar 

  • Chen L, Evans T, Anand S, et al (2004) Human resources for health: overcoming the crisis. Lancet 364:1984–1990

    Article  Google Scholar 

  • Cigno A (1998) Fertility decisions when infant survival is endogenous. J Popul Econ 11(1):21–28

    Article  Google Scholar 

  • Cleland J (2001) The effects of improved survival on fertility: a reassessment. Popul Dev Rev 27:60–92

    Google Scholar 

  • Datar A, Ghosh A, Sood N (2007) Mortality risks, health endowments, and parental investments in infancy: evidence from rural India. NBER Working paper no. 13649

  • De Lannoy A (2005) There is no other way out: educational decision-making in an era of AIDS: how do HIV-positive mothers value education? Centre for social science research working paper no. 137. University of Cape Town

  • Delavande A, Kohler HP (2012) The impact of HIV testing on subjective expectations and risky behavior in Malawi. Demography 49(3):1011–1036

    Article  Google Scholar 

  • De Paula A, Shapira G, Todd P (2010) How beliefs about hiv status affect risky behaviors: evidence from Malawi. PIER working paper no. 11-005

  • Durevall D, Lindskog A (2008) Adult mortality and fertility: evidence from Malawi. Country Econ Report 2008:1

    Google Scholar 

  • Evans D, Miguel E (2007) Orphans and schooling in Africa: a longitudinal analysis. Demography 44(1):35–57

    Article  Google Scholar 

  • Fink G, Linnemayr S (2008) HIV, education and fertility: long term evidence from sub-Saharan Africa. Working paper, Harvard School of Public Health

  • Fortson J (2008) Mortality risk and human capital investment: the impact of HIV/AIDS in sub-Saharan Africa. Rev Econ Stat 93(1):1–15

    Article  Google Scholar 

  • Fortson J (2009) HIV/AIDS and fertility. Am Econ J Appl Econ 1(3):170–194

    Article  Google Scholar 

  • Grassly N, Desai K, Pegurri E, Sikazwe A, Malambo I, Siamatowe C, Bundy D (2003) The economic impact of HIV/AIDS on the education sector in Zambia. AIDS 17(7):1039–1044

    Article  Google Scholar 

  • Hoddinott J, Maluccio JA, Behrman JR, Flores R, Martorell R (2008) The impact of nutrition during early childhood on income, hours worked, and wages of Guatemalan adults. Lancet 371:411–416

    Article  Google Scholar 

  • Idler E, Benyamini Y (1997) Self-rated health and mortality: a review of twenty seven community studies. J Health Soc Behav 38(1):21–37

    Article  Google Scholar 

  • Jayachandran S, Lleras-Muney A (2009) Life expectancy and human capital investments: evidence from maternal mortality declines. Q J Econ 124(1):349–397

    Article  Google Scholar 

  • Juhn C, Kalemli-Ozcan S, Turan B (2008) HIV and fertility in Africa: first evidence from population-based surveys. J Popul Econ 1–19

  • Kalemli-Ozcan S (2012) AIDS, reversal of the demographic transition and economic development: evidence from Africa. J Popul Econ 25(3):871–897

    Article  Google Scholar 

  • Kalemli-Ozcan S, Turan B (2011) HIV and fertility revisited. J Dev Econ 96(1):61–65

    Article  Google Scholar 

  • Kelly M (2000) Planning for education in the context of HIV/AIDS IIEP fundamental of educational planning, No. 66. UNESCO, Paris

    Google Scholar 

  • Maluccio JA, Hoddinott J, Behrman JR, Quisumbing A, Martorell R, Stein AD (2009) The impact of nutrition during early childhood on education among Guatemalan adults. Econ J 119:734–763

    Article  Google Scholar 

  • Matovu J, Gray R, Makumbi F, Wawer M, Serwadda D, Kigozi G, Sewankambo N, Nalugoda F (2005) Voluntary HIV counseling and testing acceptance, sexual risk behavior and HIV incidence in Rakai, Uganda. AIDS 19(5):503–511

    Article  Google Scholar 

  • Mishra V, et al (2007) Education and nutritional status of orphans and children of HIV-infected parents in Kenya. AIDS Educ Prev 19(5):383–395

    Article  Google Scholar 

  • Montgomery M (2000) Perceiving mortality decline. Popul Dev Rev 26(4):795–819

    Article  Google Scholar 

  • Noël-Miller CM (2003) Concern regarding the HIV/AIDS epidemic and individual childbearing, evidence from rural Malawi. Demogr Res 1(10):318–349

    Google Scholar 

  • Nyblade L, Menken J, Wawer M, Sewankambo N, Serwadda D, Makumbi F, Lutalo T, Gray R (2001) Population-based HIV testing and counseling in rural Uganda: participation and risk characteristics. J Acquir Immune Defic Syndr 28(5):463–470

    Article  Google Scholar 

  • Oster E (2012) HIV and sexual behavior change: why not Africa? J Health Econ 31(1):35–49

    Article  Google Scholar 

  • Papageorgiou C, Stoytcheva P (2008) What is the impact of AIDS on cross-country income mimeo. International Monetary Fund, Washington

    Google Scholar 

  • Preston SH (1978) The effects of infant and child mortality on fertility. Academic, New York

    Google Scholar 

  • Shapira G (2011) The effect of beliefs about own HIV status on fertility Philadelphia, University of Pennsylvania, unpublished Ph.D. Dissertation

  • Singh-Manoux A, Guéguen A, Martikainen P, Ferrie J, Marmot M, Shipley M (2007) Self-rated health and mortality: short-and long-term associations in the Whitehall II study. Psychosom Med 69(2):138–143

    Article  Google Scholar 

  • Strulik H (2003) Mortality, the trade-o? between child quality and quantity, and demoeconomic development. Metroeconomica 54(4):499–520

    Article  Google Scholar 

  • Ueyama M, Yamauchi F (2009) Marriage behavior response to prime-age adult mortality: evidence from Malawi. Demography 46(1):43–63

    Article  Google Scholar 

  • UNAIDS (2010) Global report. http://www.unaids.org/globalreport/

  • Waters E, Doyle J, Wolfe R, Wright M, Wake M, Salmon L (2000) Influence of parental gender and self-reported health and illness on parent-reported child health. Pediatrics 106(6):1422–1428

    Article  Google Scholar 

  • Watkins S, Behrman JR, Kohler HP, Zulu EM (2003) Introduction to research on demographic aspects of HIV/AIDS in rural Africa. Demogr Res S1(1):1–30

    Article  Google Scholar 

  • Weinhardt L, Carey M, Johnson B, Bickham N (1999) Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health 89(9):1397–1405

    Article  Google Scholar 

  • Whiteside A (2002) Poverty and HIV/AIDS in Africa. Third World Q 23:313–332

    Article  Google Scholar 

  • Willis R (1973) A new approach to the economic theory of fertility. J Polit Econ Part 2 New Econ Approach Fertil 81(2):S14–S64

    Google Scholar 

  • Yamano T, Shimamura Y, Sserunkuuma D (2006) Living arrangements and schooling of orphaned children and adolescents in Uganda. Econ Dev Cult Chang 54(4):833–856

    Article  Google Scholar 

  • Young A (2005) The gift of the dying: the tragedy of AIDS and the welfare of future African generations. Q J Econ 120(2):423–466

    Google Scholar 

Download references

Acknowledgments

This research was supported by the UK Economic and Social Research Council/Hewlett Foundation Grant (PI Marcos Vera-Hernandez, Institute of Fiscal Studies, University College London; PI on a subcontract to the University of Pennsylvania, Hans-Peter Kohler) on “Effects of Reproductive Health on Poverty in Malawi” and R01-HD-053781-01A1 (PI Hans-Peter Kohler) on “Consequences of High Morbidity and Mortality in a Low-Income Country”. The authors are grateful for insightful comments and suggestions of the three anonymous referees.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Ruben Castro.

Additional information

Responsible editor: Junsen Zhang

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Castro, R., Behrman, J.R. & Kohler, HP. Perception of HIV risk and the quantity and quality of children: the case of rural Malawi. J Popul Econ 28, 113–132 (2015). https://doi.org/10.1007/s00148-013-0498-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00148-013-0498-0

Keywords

  • HIV
  • Child quantity
  • Child quality
  • Malawi
  • Fertility
  • Schooling
  • Child health

JEL Classification

  • D13
  • I12
  • I20