Schooling, Income, Marriage, and Pregnancy: Evidence from a Cash Transfer Experiment Berk Özler...
-
Upload
maude-reed -
Category
Documents
-
view
212 -
download
0
Transcript of Schooling, Income, Marriage, and Pregnancy: Evidence from a Cash Transfer Experiment Berk Özler...
Schooling, Income, Marriage, and Pregnancy: Evidence from a Cash Transfer Experiment
Berk ÖzlerDevelopment Research Group, World BankDecember 1, 2009.
Early marriage, teenage pregnancy, and future outcomes As the invitation to this event suggested,
evidence on the consequences of teen pregnancy and early marriage is scant in the developing world.
In terms health, for the mother, early childbearing is suggested to be risky due to a variety of factors: obstructed labor, unsafe abortions, anaemia/malaria, hypertensive disease, etc.
Not for citation without explicit permission from the authors. 2
Early marriage, teenage pregnancy, and future outcomes Similarly, for the child, biological immaturity
may lead to low birth weight, preterm births, and higher rate of infant mortality.
Field and Ambrus (2008) find that each additional year that marriage is delayed is associated with 0.22 additional years of schooling and higher literacy. They suggest minimum age laws for marriage.
Not for citation without explicit permission from the authors. 3
If we believe these findings, what action should we take? Consent laws Invest in life skills, knowledge of reproductive
health and HIV, empowerment of young women
Job training, improved access to markets and services
Transfers to parents conditional on keeping daughters unmarried.
Not for citation without explicit permission from the authors. 4
Cash transfers and sexual behavior Of course, there is already a tool that many
governments are utilizing that can produce these desired outcomes as a ‘direct’ or a ‘side’ effect: conditional cash transfer programs for schooling: In a country like Malawi, marriage and schooling
are practically mutually exclusive (unlike child labor and schooling in other settings)
Incentives to attend school are aligned with those to avoid pregnancy (Duflo et al., 2006; Baird et al., 2009)
Not for citation without explicit permission from the authors. 5
Cash transfers and sexual behavior Nonetheless, even if we have faith in the
ability of these programs to generate the desired outcomes in marriage, pregnancy, and sexual behavior, there are at least two questions we should ask:1. Could unconditional transfers work equally well
when it comes to improving these outcomes?
2. To whom should we be making the transfers?
Not for citation without explicit permission from the authors. 6
Not for citation without explicit permission from the authors. 7
Evidence from a randomized cash transfer experiment
Zomba Cash Transfer Program (ZCTP) is a two-year randomized intervention that provides cash transfers (and school fees) to young women to stay in or return to school.
Program has multifaceted research design with contract variation in various dimensions. Schoolgirls in randomly selected villages receive
unconditional transfers. Transfers split between parents and girls:
Parents’ transfer randomized at village level between $4-10. Girls’ transfer randomized at individual level between $1-5.
Cash transfers and sexual behavior Nonetheless, even if we believe in the ability
of these programs to generate the desired outcomes in marriage, pregnancy, and sexual behavior, there are at least two questions we should ask:1. Could unconditional transfers work equally well
when it comes to improving these outcomes?
2. To whom should we be making the transfers?
Not for citation without explicit permission from the authors. 8
Marriage and schooling status by treatment status after one year
Not for citation without explicit permission from the authors. 9
Control CCT UCT
Married 4.8% 4.4% 1.7%
In school 78.2% 85.0% 83.2%
Never married, not in school 17.0% 10.6% 15.1%
CCT had no effect on marriage rates Girls who got married between baseline and
one-year follow-up in the control group or in the CCT have the baseline characteristics.
This implies that there was no movement between the marriage and the schooling cell for CCT recipients.
Not for citation without explicit permission from the authors. 10
UCT had a significant effect on marriage rates. That entire effect is due to girls moving from the
‘marriage’ cell to the ‘home’ cell.
Had the UCT had the effect of moving girls from the ‘marriage’ to ‘school’ cell, the CCT would have had to have the same effect.
But, we know that marriage rates did not change under CCT.
Results are similar for pregnancy.
Not for citation without explicit permission from the authors. 11
Treatment effects on marriage and pregnancy
Not for citation without explicit permission from the authors. 12
Never-married
Currently Pregnant
Year 2 -0.047*** 0.036***
(0.008) (0.005)
Treatment in Year 2 0.031*** -0.031***
(0.012) (0.007)
Conditional Treatment in Year 2 -0.032** 0.022**
(0.014) (0.010)
Constant 1.000*** 0.003
(0.003) (0.002)
note: *** p<0.01, ** p<0.05, * p<0.1
Treatment effects on marriage and pregnancy
Not for citation without explicit permission from the authors. 13
Never-married
Currently Pregnant
Year 2 -0.047*** 0.036***
(0.008) (0.005)
Treatment in Year 2 0.031*** -0.031***
(0.012) (0.007)
Conditional Treatment in Year 2 -0.032** 0.022**
(0.014) (0.010)
Constant 1.000*** 0.003
(0.003) (0.002)
note: *** p<0.01, ** p<0.05, * p<0.1
Cash transfers and sexual behavior Nonetheless, even if we believe in the ability
of these programs to generate the desired outcomes in marriage, pregnancy, and sexual behavior, there are at least two questions we should ask:1. Could unconditional transfers work equally well
when it comes to improving these outcomes?
2. To whom should we be making the transfers?
Not for citation without explicit permission from the authors. 14
Not for citation without explicit permission from the authors. 15
Never-married
Currently Pregnant
Never-married
Currently Pregnant
Year 2 -0.047*** 0.036*** -0.047*** 0.036***
(0.008) (0.005) (0.008) (0.005)
Treatment in Year2 0.029 -0.023*** 0.024 -0.023***
(0.019) (0.008) (0.021) (0.008)
Conditional Treatment in Year 2 -0.040 0.026* -0.037 0.028*
(0.028) (0.014) (0.028) (0.016)
HH Transfer 0.004 -0.005** 0.004 -0.005**
(0.005) (0.003) (0.005) (0.003)
HH Transfer under Conditional Treatment
-0.003 0.005 -0.004 0.005
(0.006) (0.005) (0.006) (0.005)
Girl Transfer -0.005 0.004
(0.003) (0.002)
Girl Transfer under Conditional Treatment
0.009 -0.009**
(0.006) (0.004)
High Girl Transfer (above median) -0.011 0.018
(0.009) (0.012)
High Girl Transfer (above median) under Conditional Treatment
0.034* -0.049**
(0.018) (0.020)
_cons 1.000*** 0.003 1.000*** 0.003
(0.003) (0.002) (0.003) (0.002)
note: *** p<0.01, ** p<0.05, * p<0.1
Conclusions Cash transfer programs for schooling can
significantly affect sexual behavior, age at first marriage, and teenage pregnancy.
At least in SSA, unconditional cash transfer programs may be equally (if not more) effective in improving these particular outcomes.
If programs are conditional, it may make sense to make the transfer directly to the target beneficiary, i.e. the young women.
Not for citation without explicit permission from the authors. 16
Conclusions Given that, after one year, the effects
from conditional transfers on schooling and risk of HIV (and HSV-2) infection are similar to those from unconditional ones, we don’t have a good reason to recommend a CCT program over a UCT.
Are we just delaying the inevitable? It remains to be seen whether the longer-term
impacts of the program will be as strong as the short-term impacts described in this paper.
Not for citation without explicit permission from the authors. 17
Not for citation without explicit permission from the authors. 18
No Yes TotalNever-marriedNo 7.95 0.85 8.8Yes 24.63 66.56 91.2Total 32.58 67.42 100
Regularly attended school
Not for citation without explicit permission from the authors. 19
Schoolgirls CCT vs. UCT Schoolgirls CCT vs. UCT
Year 2 -0.018** -0.020* -0.022*** -0.025**
(0.008) (0.012) (0.009) (0.011)
Conditional Treatment in Year 2 0.003 0.005
(0.011) (0.010)
Constant 0.027 0.026 0.074** 0.072*
(0.033) (0.033) (0.038) (0.038)
Number of observations 1,284 1,284 1,280 1,280
note: *** p<0.01, ** p<0.05, * p<0.1
HIV HSV-2
Time
HIV prevalence
T=0 T=2 (end of CCT) T=4
0.05
0.08
0.10
Hypothetical HIV incidence by treatment status
Treatment
Control