mens womens assets_social protection_oct 2011_v2
TRANSCRIPT
Wednesday, October 19, 2011
Men’s and Women’s Asset Accumulation and
Implications for Social Protection:
Evidence from Bangladesh and Uganda
Agnes Quisumbing, Neha Kumar, and Julia Behrman
International Food Policy Research Institute
October 2011
Research supported by AMA-CRSP, the Swiss Development
Corporation, and the World Bank
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Assets: the wealth of families
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Different assets can be held by men, women, and
jointly—this varies across cultures
Page 3
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE Page 4
Presentation overview
1. Motivation: why look at differential impact of shocks on
men’s, women’s, and jointly held assets?
2. Survey design and data
• Bangladesh: CPRC-DATA-IFPRI long-term impact study + food
price crisis survey (2006/7 and 2010)
• Uganda: HarvestPlus OFSP Reaching End User Impact
Evaluation survey (2007 and 2009)
3. Descriptives
4. Impacts of shocks on men’s, women’s, and joint assets
5. Implications for social protection
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Motivation: why look at gender-differentiated asset
accumulation?
• Accumulating evidence rejecting unitary model of the household
in many countries—resources are not pooled within the
household
• Growing evidence that risk is not pooled within households and
that risk perceptions may also differ between men and women
• Anthropological evidence (Thailand, Indonesia, Bangladesh)
suggests that men and women have different asset
accumulation strategies, and use their assets in different ways
to cope with shocks
• If this is true, how do we help women, men, and their families
protect their livelihoods in the face of adverse events?
Page 5
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Research questions
• Is the impact of negative events and processes (flood
shocks, dowries, illness, death) different on husband-,
wife- and jointly-owned assets? Are these mitigated by
positive events?
• Do these impacts differ depending on country and
context?
• …And a policy-related question
• What are the implications for the design of social
protection systems?
Page 6
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Survey design and data
Page 7
Page 8
Bangladesh: Map of study sites of longitudinal study 2006/7
and 2010
Page 9
Uganda: Map of study sites of OFSP REU impact
evaluation 2007 and 2009
BUKEDEA
L. Victoria
L. A
lber
t
L. Kyoga
L. Edward
L. George
SO
RO
TI
KUMI
KAMULILUWEERO
MU
KO
NO
KAYUNGAÊÚ
ÊÚ
ÊÚKampala
Jinja
Mbale
Implementation Areas
Non-OFSP Areas
ÊÚ Major Towns
LEGEND
ÊÚÊÚ
ÊÚ
N
50 0 50 Kilometers
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
In both countries
• Surveys conducted with baseline in 2007 and
follow-up after the food price crisis
• Detailed gender disaggregated data (in Uganda,
baseline gender disaggregation collected
retrospectively)
• Ownership categories: joint, husband, wife
assets
• Analysis limited to intact, monogamous couples
(couples that stayed together between 2007 and
2009/10, excluding polygamous households)
Page 10
Assets have grown over time, but there are clear
gender differences in asset ownership
Bangladesh Uganda
0
10
20
30
40
50
60
70
Total value of nonland
assets owned by the
household
Jointly owned nonland assets
Nonland assets
exclusively owned by the
husband
Nonland assets
exclusively owned by the
wife
Total value of assets by ownership status, in '000 taka (2007 values)
2007
2010
0
1000
2000
3000
4000
5000
6000
7000
Total value of
household assets
Total value of assets
jointly owned by husband and wife
Total value of assets owned by husband
Total value of assets owned by
wife
Total Value of Assets by ownership status in '000 UGS (2007 values)
2007
2009
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Different types of assets are held individually and jointly in
Bangladesh—joint assets dominate, except for land
Page 12
1 2 1 2
33
9
83
36 41
59
10
42
16
6258
39
5749
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Land Consumption durables
Agricultural durables
Nonag durables
Jewelry Livestock
Ownership shares by type of asset, 2010
Wife Husband Joint
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
In Uganda, joint assets are less important than husband’s
assets across most categories
Page 13
0.110.07
0.25
0.14
0.41
0.53 0.65
0.44
0.43
0.43
0.360.28 0.31
0.43
0.16
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Consumer durables
Land Livestock Productive equipment
Cash savings
Ownership Shares by type of Asset, 2009
2009 Joint
2009 Husband
2009 Wife
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Floods and droughts
Page 14
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Bangladesh: Negative shocks and positive events
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
45.00
50.00
Proportion of households reporting shocks and positive events
Type of shock
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Uganda: Negative shocks and positive
events
Page 16
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Percentage of households affected by shocks, 2007-09
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Equation to be estimated
Ait-AiB= ß0 + ß1 AiB + ß2AiB2 + ß3AiB
3 + ß4AiB4+
Zi(Shocks, Positive events)i + Ci(HH demographics,
baseline wealth, village dummies)i + εit
Dependent variable: Asset growth
Regressors: Lagged assets (linear, squared, cubed, fourth)
Covariate shocks (floods, drought)
Idiosyncratic shocks (illness, death, dowry/wedding expenses)
Positive events (remittances, inheritance, received dowry)
HH demographic characteristics: age of head, age squared, hh size,
proportion in age-sex categories
Value of (assets) land at baseline [assets in land equation; land in
assets equation)
Location dummies
Page 17
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
What is the impact of shocks?
• Look at the coefficients on shocks from
Zi(Shocks, Positive events)I
• Depending on specification, this is the change
(or percentage change) in assets as a result of
the shock
• We obtain this separately for men’s, women’s,
and joint assets, and for different types of shocks
Page 18
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Change in asset as a proportion of baseline holdings,
Bangladesh
Page 19
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
Flood Drought Price shock Illness Death
Hus Land
Hus Assets
Wife Assets
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Change in asset as a proportion of baseline holdings,
Uganda
Page 20
-140
-120
-100
-80
-60
-40
-20
0
20
40
Flood Drought Price Shock Illness Death
Joint Assets
Hus Land
Wife Assets
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Consumer goods, livestock and jewelry: where the action
is in Bangladesh
Page 21
-60
-50
-40
-30
-20
-10
0
Flood Drought Price shock Illness Death
Hus consumer
Jt jewel
Hus jewelry
Wif jewelry
Hus livestock
Wif livestock
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Conclusions--1
• Shocks appear to have differential impacts depending on
ownership, and depending on context
• In Bangladesh:
• women’s assets are negatively affected by illness—a really big
hit (40% of her baseline assets)
• Joint assets are protected, men’s are affected but only to a
limited extent
• Illness is the shock most frequently reported by
households=>implications for asset disposal and social
protection?
• Analysis for disaggregated assets indicates that there is a lot of
movement in consumer durables, livestock, and jewelry
• Floods have negligible impacts on assets in Bangladesh,
possibly because emergency assistance system works
Page 22
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Conclusions--2
• In Uganda:
• Husband’s land negatively affected by death, impact of flood
needs to be looked into
• Wife’s assets are negatively affected by drought, price shocks,
and death, and to a greater degree than husband’s assets
• Joint assets are also negatively affected by price shocks
• Indications are that men’s assets (aside from land) are protected,
but women’s assets and joint assets are sacrificed
• In both countries, women’s assets are a small proportion
of household assets, but they take a big hit
Page 23
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Implications--1
• Need to devise social protection strategy to provide
insurance against shocks
• Health insurance may help protect asset stocks (as well
as individual health)
• Social safety nets (public works, income transfer
programs) may help prevent asset depletion, which
would also help protect future livelihoods
• We are now developing experiments that will look at
men’s and women’s willingness to pay for flood
insurance (as part of gender and climate change project)
Page 24
INTERNATIONAL FOOD POLICY RESEARCH INSTITUTE
Implications--2
• Need to provide mechanisms for poor to save and build
up asset stocks, and to rebuild them after shocks
• Need to provide mechanisms to prepare adequately for
(anticipated) life-cycle events
• Women, in particular, need to be able to build up assets
(savings?) that they can control, not only while the
couple is together, but perhaps more importantly if the
marriage dissolves through death or divorce
Page 25