Reproductive health, female empowerment and economic prosperity. Elizabeth Frankenberg Duncan Thomas

Transcription

1 Reproductive health, female empowerment and economic prosperity Elizabeth Frankenberg Duncan Thomas

2 Studies suggest females with more resources under own control more likely to use prenatal care have healthier babies have healthier children

3 Studies suggest females with more resources under own control more likely to use prenatal care have healthier babies have healthier children Why? Several potential reasons Relative to males, females care more about own health care more about health of children (more altruistic) rely more on children in old age (higher expected return to invest in children) care more about longer run (more patient) + other reasons not mutually exclusive

4 Do males and females have different preferences?

5 Do males and females have different preferences? Mexican Family Life Survey (MxFLS) On-going nationally representative longitudinal survey of individuals, households and communities in Mexico Baseline conducted in First follow up in 2005 Preferences Pilot (MxFLS-PP) Re-interview adults in 2 states in 2005 wave of MxFLS Drawing on methods established in behavioral economics Completed incentivized tasks in group setting ( lab ) to measure preferences E.g. Offer $15 for sure/gamble with equal chance of $50 or $0 Resp. paid for 1 randomly selected decision at session end

6 Gender differences: Decisions of married respondents Incentivized tasks Males Females Diff Dictator game: % give away [Altruism] Diff is significant at 5%

7 Gender differences: Decisions of married respondents Incentivized tasks Males Females Diff Dictator game: % give away [Altruism] Ultimatum bargaining game [1 5] [1 4] [1 5] % give away [Fairness] Minimum % accept [Reciprocity] [1 0] 34 [1 5] 32 [1 8] 1.7 Trust game [0 8] [0 7] [1 0] % give away [Trust] % return [Trustworthiness] [1 4] 28 [1 7] 25 [2 1] 3.2 Diff is significant at 5%

8 Suggests females not more altruistic than males But altruism likely relationship-specific Difficult to measure Experiments involving family members Respondent gives about ½ endowment away No difference in % given by males and females to children (and other family members) Exception: Husbands give ~55% to wives Wives give ~40% to husbands (Significantly different) Turn to other preferences Reflects social norms?

9 Gender differences: Decisions of married respondents Incentivized tasks Males Females Diff Dictator game: % give away [Altruism] Ultimatum bargaining game [1 5] [1 4] [1 5] % give away [Fairness] Minimum % accept [Reciprocity] [1 0] 34 [1 5] 32 [1 8] 1.7 Trust game [0 8] [0 7] [1 0] % give away [Trust] % return [Trustworthiness] [1 4] 28 [1 7] 25 [2 1] 3.2 Risk aversion [1 3] [1 3] [1 3] Coeff of variation [Risk loving] Time preferences [0 02] [0 02] [0 02] % always impatient Sample size [0 02] 492 [0 02] 492 [2] Diff is significant at 5%

10 Gender differences Incentivized decisions indicate Males more pro-social more altruistic, more trusting, more trustworthy possibly reflects social norms? Females more inclined to save for the future (more patient)

11 Gender differences Incentivized decisions indicate Males more pro-social more altruistic, more trusting, more trustworthy possibly reflects social norms? Females more inclined to save for the future (more patient) Supportive evidence in observational data PROGRESA pays benefit to women Marginal propensity to consume PROGRESA income about ½ spent on food, child clothing and household goods about ½ invested in small livestock (pigs, chickens) (tend to be under control of women)

12 What happens if give females more choice? Examine impact of randomized family planning services intervention (Matlab) Studies established exposure to program resulted in reduced fertility Building on work on Joshi and Schultz examine impact on savings decisions separating males and females

13 Old age security If children provide resources at older ages as fertility declines both parents should save more for old age if females live longer than males and females rely on savings in older age more than males effect should be stronger for females than males Preferences If females have lower inter-temporal discount rates should save more than males

14 Old age security If children provide resources at older ages as fertility declines both parents should save more for old age if females live longer than males and females rely on savings in older age more than males effect should be stronger for females than males Preferences If females have lower inter-temporal discount rates should save more than males Access to contraception should increase saving, esp for females

15 Effect of treatment on asset ownership and log(value assets) % Own ln(value) Gold Any Gold Any Females [2.5] [2.4] [0.12] [0.13] Males [1.6] [2.5] [0.32] [0.17] Diff (F-M) [3.0] [3.5] [0.34] [0.21] Bari FE [3.3] [3.8] [0.43] [0.21] HH FE [3.8] [3.9] [0.43] [0.23] [Standard errors]

16 Effect of treatment on asset ownership and log(value assets) % Own ln(value) Gold Any Gold Any Females [2.5] [2.4] [0.12] [0.13] Males [1.6] [2.5] [0.32] [0.17] Diff (F-M) [3.0] [3.5] [0.34] [0.21] Bari FE [3.3] [3.8] [0.43] [0.21] HH FE [3.8] [3.9] [0.43] [0.23] [Standard errors]

17 Effect of treatment on asset ownership and log(value assets) % Own ln(value) Gold Any Gold Any Females [2.5] [2.4] [0.12] [0.13] Males [1.6] [2.5] [0.32] [0.17] Diff (F-M) [3.0] [3.5] [0.34] [0.21] Bari FE [3.3] [3.8] [0.43] [0.21] HH FE [3.8] [3.9] [0.43] [0.23] [Standard errors]

18 Effect of treatment on asset ownership and log(value assets) % Own ln(value) Gold Any Gold Any Females [2.5] [2.4] [0.12] [0.13] Males [1.6] [2.5] [0.32] [0.17] Diff (F-M) [3.0] [3.5] [0.34] [0.21] Bari FE [3.3] [3.8] [0.43] [0.21] HH FE [3.8] [3.9] [0.43] [0.23] [Standard errors]

19 Conclusions Why do women invest in children?

20 Conclusions Why do women invest in children? Women not appear to be more altruistic than men But do appear to be more patient If give females money, tend to save more than males If give females access to contraceptives, tend to save more than females not have access

21 Effect of children on wealth A = α + αt + α CEB + α C + Spline in cohorts+ μ+ ε First stage: Effect of treatment on CEB CEB Treatment [0.10] [Standard error] IV % own ln(value) CEB [0.14] [0.64] [Standard error]

22

23 Give respondent money offer opportunity to donate some to randomly assigned partner partner does not know about

24 Altruism Percentage of endowment given to partner By gender of donor Male Female Percent All Spouse Child Mother Father Other HH mem Same village Different village

25 Altruism Percentage of endowment given to partner By gender of donor Male Female Percent All Spouse Child Mother Father Other HH mem Same village ~50% given to HH members, ~25% given to neighbors/strangers Males give more to neighbors/strangers than females Males give more to HH members (driven by husbands give more to wives) % given to children no different between mothers and fathers Different village

26 Old age security If children provide resources at older ages as fertility declines both parents should save more for old age if females live longer than males and females rely on savings in older age more than males effect should be stronger for females than males

27

28 Effect of improved access to family planning services Reduced form intent-to-treat effect on assets A A = β + βt + β T * C + β C + Spline in cohorts+ μ+ ε Compare impact on individual level of assets of married females and males living in treatment area and birth cohort females exposed to treatment relative to older birth cohort in treatment area and same birth cohort in control area