TEST BANK For Essentials of Negotiation, 7th Edition by Roy Lewicki, Bruce Ba...
Impact of Zambia's Child Grant Program on Women's Savings
1. The impact of the Child Grant Program (CGP)
on women’s savings
Experimental evidence from rural Zambia
Luisa Natali, Sudhanshu Handa, Amber Peterman, David Seidenfeld, Gelson Tembo
on behalf of The Zambia Cash Transfer Evaluation Team
Luisa Natali
University of Sussex / UNICEF Office of Research
CSAE Conference, Oxford
21st of March 2017
1
2. Introduction
UCTs on the rise, especially in SSA
In Zambia, the Child Grant Program (CGP) ->
positive impacts on a broad range of outcomes
Savings:
•Medium-term outcome
•Secondary-level effect
•Crucial role: precautionary + investment reasons
•Focus on women’s cash savings
•Empirical literature on liquid savings is thin
(Angelucci et al. 2012; Blattman et al. 2014; Haushofer and Shapiro 2016; Stoeffler and Mills 2014)
Innovative programs to directly provide financial
services to the poor, what is the role of CTs?
2
3. The Zambia Child Grant Program
• Government run unconditional cash transfer
• Where: in 3 districts of Zambia
• Target: All households with a child under 3 at baseline
• Virtually all transfer recipients are women
• Transfer size: 60 ZMW a month (≈US $12), flat
• Equivalent to 27 percent of household’s monthly expenditure.
• Frequency: bi-monthly
• Goal: Reduce extreme poverty and the intergenerational
transfer
4. 4
0 hours, Lusaka
16 hours
20 hours
12 hours
Not covered
Kaputa
Kalabo
Some of the most
remote districts
5. Study design and timeline of the evaluation
•Cluster Randomized Control Trial
• Three districts selected (geographical targeting)
• In each district: random selection of 30 communities
• Identification of all eligible households with at least one child
under 3 y.o. 28 hhs randomly sampled in each community
2010, Oct-Nov:
Baseline
2012, Oct-Nov:
24 month follow up
2013, Oct-Nov:
36 month follow up
2013, June-July:
30 month follow up
2011, Feb:
First transfer
Randomization
Baseline
2010, Oct/Nov
24-month
2012, Oct/Nov
36-month
2012, Oct/Nov
30-month
2013, June/July
First transfer
2011, Feb
6. Data and outcome variables
• 2,519 hhs surveyed at baseline (Oct/Nov 2010)
• 2 follow-ups:
• 24-month (Oct/Nov 2012)
• 36-month (Oct/Nov 2013) [48-months not used]
• The survey collected: consumption, socioeconomic status,
women’s empowerment modules, etc.
• Primary outcome variables:
• Savings: dummy equal to 1 if woman is currently saving in cash, 0
otherwise
• Amount saved: amount saved in cash in the last one month
(logged)
6
8. Randomization Balance Test
Baseline characteristics of women and test for equivalence
• Overall, randomization was successful in producing balanced
treatment and control groups
8
All Control Treated P-value of diff.
Age (years) 29.46 29.27 29.64 0.55
Ever attended school 0.71 0.7 0.73 0.37
Never married 0.11 0.11 0.11 0.81
Divorced or separated 0.1 0.11 0.08 0.05
Widowed 0.06 0.06 0.05 0.65
Shangombo district 0.32 0.31 0.32 0.92
Kaputa district 0.24 0.25 0.22 0.74
Consumption expenditure per capita (ZMW) 40.66 39.57 41.75 0.44
Household size 5.67 5.62 5.72 0.56
Proportion of women holding any cash savings 0.17 0.16 0.19 0.38
Amount saved (ZMW) 13.67 14.32 13.03 0.68
N 2,069 1,035 1,034
9. Attrition
• Household attrition: 9% at 24-months; 2% at 36 months
• Individual attrition: higher but equal across treatment and
control areas.
• The analysis sample is not differentially lost to follow-up in
terms of important/prognostic baseline characteristics
either.
9
Overall Control Treatment P-value of diff.
Attrition rate 16.8 16.8 16.9 0.98
10. 10
Impact on women’s cash savings
Any cash savings
Amount saved
(last month, logged)
(1) (2) (3) (4
VARIABLES Unadjusted Adjusted Unadjusted Adjusted
24-months follow-up 0.0580* 0.0506 0.231** 0.204*
(0.0308) (0.0310) (0.114) (0.114)
36-months follow-up 0.0696** 0.0588* 0.366*** 0.326**
(0.0331) (0.0337) (0.126) (0.128)
Treated 0.028 0.021 0.0822 0.0492
(0.0313) (0.0306) (0.109) (0.106)
DD 24-months 0.230*** 0.230*** 1.102*** 1.102***
(0.0459) (0.046) (0.177) (0.178)
DD 36-months 0.0968** 0.0971** 0.528*** 0.529***
(0.0451) (0.0451) (0.179) (0.179)
Observations 6,207 6,207 6,207 6,207
Baseline control mean 0.16 0.59
Notes: Estimations use difference-in-difference modeling (DD indicates treatment effect). Robust standard
errors clustered at the community level are in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Unadjusted
specifications include district dummies as this was the stratifying indicator for the randomization. Estimations
with adjustment include woman’s age, education and marital status, household size and household
demographic composition, and districts.
11. 11
Dependent
variable
DD 24-months
DD 36-
months
Baseline,
control mean
Any milk cows 0.0319 0.0154 0.064
(0.0233) (0.0218)
Any cattle 0.0918*** 0.105*** 0.111
(0.0237) (0.0296)
Any goats 0.0406*** 0.0155 0.011
(0.0134) (0.0153)
Any chickens 0.119** 0.160*** 0.435
(0.0492) (0.0419)
Any ducks 0.0309*** 0.0299*** 0.034
(0.0115) (0.00894)
Any livestock 0.171*** 0.183*** 0.502
(0.05) (0.0416)
Note: Estimations use difference-in-difference modeling (DD indicates treatment effect). Robust standard
errors clustered at the community level are in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Impact
estimates reported are from the unadjusted model. Unadjusted specifications include district dummies as
this was the stratifying indicator for the randomization; impact estimates from the adjusted model
(controlling for woman’s age, education and marital status, household size and household demographic
composition, and districts) are consistent.
Impact also on other forms of saving
(household level)
12. How did the CGP lead to such positive
impacts on women’s savings?
Potential pathways
12
13. Shifts in intra-household preferences
13
– Does the CGP increase women’s decision
making-power within the household, and in
particular in women’s control over resources?
– Bonilla et al. (2016):
• Quantitative evidence: small but significant increase
• Qualitative evidence: the cash transfer allows choices
(C & I) which makes women feel empowered
14. Women’s productive investments
14
Impact on participation in NFEs (but also revenues and profits)
Does the CGP increase women’s income generation?
24-month 36-month
(1) (2) (3) (4)
Unadjusted Adjusted Unadjusted Adjusted
Treated 0.166*** 0.164*** 0.140*** 0.135***
(0.0408) (0.0402) (0.0345) (0.0326)
Observations 2,069 2,069 2,069 2,069
Follow-up control mean 0.30 0.31
Notes: Estimations use single difference modeling. Dependent variable: whether household operates a NFE (dummy).
Robust t-statistics clustered at the community level are in parentheses. *** p<0.01, ** p<0.05, * p<0.1. Unadjusted
specifications include district dummies as this was the stratifying indicator for the randomization. Adjusted estimations
include: woman’s age, education and marital status, household size and household demographic composition, and
districts.
15. Conclusions
• The CGP program enabled higher saving rates and
larger value of savings for poor women
• The increase in women’s cash savings did not crowd
out other forms of household savings (increase in
livestock holdings)
• The CGP likely facilitated impacts on savings through :
• an increase in women’s control over resources
• an increase in income-generating activities (NFEs)
• Thought-provoking results given MFIs evidence