Bargaining Power and Biofortification: The Role of Gender in Adoption of Orange-Fleshed Sweet Potato in Uganda Julia Behrman, Daniel O. Gilligan, Neha.

Bargaining Power and Biofortification: The Role of Gender in Adoption of Orange-Fleshed Sweet Potato in

Uganda

Julia Behrman, Daniel O. Gilligan, Neha Kumar, Scott McNiven, J.V. Meenakshi, Agnes Quisumbing

Land Conference,

The World Bank

Washington, DC, March 2014

Gender and Biofortification

HarvestPlus is promoting biofortification as a strategy to reduce malnutrition (e.g., vitamin A deficiency (VAD); iron deficiency)

– strategy: breed staples crops to be a rich source of missing micronutrients like iron, vitamin A, and zinc

– potential: sustainable in rural areas, self-targeting toward the poor, cost-effective over time

Success of biofortification depends on widespread adoption and consumption of new crop varieties. Gender may be important:

– women provide much of the on-farm labor in Africa and elsewhere and are primarily responsible for child diets

– there is often a complex dynamic of intrahousehold gender relations for crop choice (von Braun, Puetz and Webb, 1989)

• HarvestPlus Orange-Fleshed Sweet Potato (OSP) Project

• disseminate provitamin-A-rich OSP as a strategy to increase vitamin A intakes and reduce vitamin A deficiency

• In 2007, 10,000 households in Uganda were given OSP vines plus agriculture, nutrition and marketing trainings in two models: • Model 1: 2 years of trainings (intensive)• Model 2: 1 year of trainings (less intensive)

An Evaluation of Biofortification in Uganda

• The IFPRI/HarvestPlus/CIP evaluation

• randomized, controlled trial • 2007-2009: baseline &

endline surveys• n=1,472 households• outcomes: OSP adoption,

dietary intakes of vitamin A, serum retinol • 2011: qual study; follow-up

survey

Key Findings of OSP Evaluation1. OSP adoption: Project increased OSP adoption by 61

percentage points (ppt). Share of OSP in total SP area increased to 43%.

2. Dietary Intakes of vitamin A (from dietary recall modules)• dietary intakes of vitamin A increased sharply for two

cohorts of children (3-5 yrs, 6-35 mos) and for adult women

• prevalence of inadequate intakes of vitamin A fell 33 ppt for young children and 26-36 ppt for women

3. Serum retinol• significant 9.5 percentage point reduction in

prevalence of low serum retinol (retinol<1.05μmol/L) at endline among children with low serum retinol at baseline

Key Findings of OFSP Evaluation:1. Impact on OFSP Adoption in 2009

Control

Model 2

Model 1

0 10 20 30 40 50 60 70 80

Cultivated OFSP

%

Impact: Model - ControlM1: 64 % ***M2: 57 % ***

• Project resulted in a 57-64 % point increase in OFSP adoption

• Project increased the share of OFSP in total sweet potato (SP) area by 41 to 46 % points

2. Prevalence of Inadequate Vitamin A Intakes, Uganda

•Prevalence of inadequate vitamin A intakes

•Fell 33% for young children (age 6-35 months)

•Fell 26-36% for adult women

•No impact on reference children (3-5 yrs) due to improvement in control group

Model 1 Model 2 Control Model 1 Model 2 Control Model 1 Model 2 ControlYoung children Reference children Women

0102030405060708090

100

Baseline Follow up

%

M1-C: -1%M2-C: -5%

M1-C: -36%**M2-C: -26%**

M1-C: -34%**M2-C: -31%**

3. Impact on Vitamin A Deficiency

• Serum retinol from blood samples was used to estimate impact on prevalence of mild vitamin A deficiency (VAD) (retinol<1.05μmol/L) for children age 3-5 at baseline or for adult women

• For children mildly VAD at baseline

• weakly significant effect reducing prevalence of mild VAD at endline by 7.6 percentage points

• significant reduction in prevalence of mild VAD of 9.5 percentage points in model with more control variables (e.g., age, deworming), but a smaller sample

• Women: project had no impact on mild VAD

• Summary: broad adoption of OFSP substantially increases vitamin A intakes and can reduce child mild VAD

1. How do men’s/women’s control over land and decisionmaking affect OSP adoption and vitamin A intakes?

2. What is the role of gender in the effect of social networks in OSP diffusion and sustainability of OSP adoption?

Topics of GAAP Research on Gender and Biofortification

1. Control over assets and OSP adoption• The share of assets exclusively owned by women or by

men does not affect the household decision to grow OSP in a given season

• In female-headed households, the share of exclusively owned...• ...land assets: weakly increases OSP adoption• ...nonland assets: decreases OSP adoption

1. What have we learned about the role of gender in OSP adoption?

Table 2: HH OSP adoption, controlling for women’s baseline asset ownership

Dep. Var.: Pr(Adopt OFSP)All project households

Female headed households

Male headed households

Share of land exclusively 0.038 0.365* -0.011 owned by women, 2007 (0.070) (0.217) (0.076)

Share of nonland assets exclusively -0.029 -0.540** 0.032 owned by women, 2007 (0.069) (0.232) (0.074)Observations 1305 138 1167Notes: RE model with controls. * significant at the 10% level; **significant at the 5% level.

2. Gender, control over land and crop choice

Household crop choice decisions are complex and are usually joint decisions by men and women, but with men taking the lead. OSP adoption is most likely on plots where decisions are joint, but women play a leading role.

• Plots exclusively controlled by women are not more likely to have OSP

• Conditional on HH adoption, male controlled plots are least likely to have OSP

1. What have we learned about the role of gender in OSP adoption?

Table 7: OSP adoption by female ownership of nonland assets• Where female share of assets is higher, decision-making on joint plots appears more egalitarian, but OSP adoption is lower on male-controlled plots

Dep Var: Grow OSP on this

Low share of female ownership of nonland assets

High share of female ownership of nonland assets

parcel (1) (2)Parcel control: female only 0.032 -0.036

(0.049) (0.035)Parcel control: male only -0.085 -0.198

(0.065) (0.082)**Parcel control: joint, female 1st 0.097 0.021

(0.029)*** (0.032)Observations 2377 2655 Notes: Other control variables not reported.

3. Women’s assets, control over land and OSP adoption

• Households in which women have lower asset ownership are more likely to grow OSP on joint plots with women in primary control

1. What have we learned about the role of gender in OSP adoption?

4. Role of gender in diffusion of OSP to other households• Women do not play a unique role in OSP diffusion, but

their participation in nutrition trainings (women only) increased diffusion

1. What have we learned about the role of gender in OSP adoption?

Dep Var: Shared OSP vines with other households All Kamuli Bukedea Mukono

Panel A Household has a female FG member 0.012 0.145** -0.082 0.067 (0.052) (0.066) (0.099) (0.146) Panel B Household has a female FG member 0.010 0.092 -0.092 0.062 (0.051) (0.059) (0.091) (0.096) Attended at least one nutrition training 0.228*** 0.347* 0.165 0.302** (0.072) (0.180) (0.109) (0.112) Notes: Models control for land area under the household’s control in 2007 and an indicator for Model 2. Models of the full sample include district dummy variables. * significant at the 10% level, ** significant at the 5% level, *** significant at the 1% level.

Table 8: Gender-based differences in diffusion of OSP, 2007-2009

5. Gender, asset control and intakes of vitamin A• Share of nonland assets exclusively controlled by

women is associated with higher increase in vitamin A intakes

• But treatment effects are not larger for women with more asset control

1. What have we learned about the role of gender in OSP adoption?

Table 9: Gender differences in control over assets vitamin A intakes

Dep Var: Change in dietary intake of vitamin A, 2007-2009

Change in dietary intake of vitamin A, 2007-2009 (μmol/L)

Panel A Average impact of OSP project 445.5*** (146.0) Share of nonland assets exclusively controlled by women 509.3** (237.8) Panel B Average impact of OSP project 269.0* (140.1) Interaction of treatment effect with share of nonland assets 356.2 exclusively controlled by women great than 5 percent (279.1) Notes: Sample includes children age 3-5 years in each round. * significant at the

2. What have we learned about the role of gender in social networks and sustainability of OSP adoption

This research is still ongoing• We have found large impacts of social and information

networks on diffusion of OSP to neighboring households

• Preliminary work on the role of gender in this diffusion does not show large effects, but this is ongoing

Implications for Programming

• In the Uganda study• Women do play an important role in decisions about

adopting OSP and about vitamin A consumption by children, but intrahousehold dynamics on the adoption decisions are complex

• suggests continuing to target women for nutrition trainings, but there may be some benefits to bringing men in as well

• In new diffusion experiments being conducted with HarvestPlus, we are giving a unique role to opinion leaders in health, who are almost always women. This may shed new light on the role of female leaders in promoting adoption

• In other biofortification studies, context really matters