1. Cash or Vouchers?
The Relative Impacts of Cash and Vouchers in
the Democratic Republic of Congo
Jenny C. Aker, Tufts University
IFPRI
November 7, 2013
Jenny C. Aker Cash and Vouchers in DRC
2. Motivation
• How should society (donors, governments and
NGOs) redistribute to the poor?
• Cash transfer programs are an increasingly important
part of social protection programs in both developed
and developing countries
• Majority of welfare transfers are still “in-kind”
Food aid, food stamps, medicines, inputs, vouchers
Especially in developing (conflict) countries
Jenny C. Aker Cash and Vouchers in DRC
3. International Cash and In-Kind Transfers
% of USAID Emergency Food Security Program
Budget
90%
80%
2011
Total: $232 million
79%
2012
Total: $374 million
% of EFSP Budget
70%
60%
50%
44%
42%
40%
30%
20%
14%
9.0%
10%
12%
0%
In-Kind
(Local/Regional Purchase)
Vouchers
Cash Transfers
Type of Assistance
Source: Congressional Research Service, 2013
Jenny C. Aker Cash and Vouchers in DRC
4. Motivation
• If program recipients would weakly prefer an equalvalued cash transfer to an in-kind transfer, why not
use cash?
• Cash might not be the preferred modality for
redistributing wealth
Targeting
Local supply
Political feasibility
“Paternalism”
• Little evidence of the relative effects of each
modality in developing countries
Jenny C. Aker Cash and Vouchers in DRC
7. Research Questions
Goals
Research
• What are the relative effects of cash versus in-kind
transfers (vouchers) on household
purchases, consumption and other measures of wellbeing?
• Households randomly assigned to equal-valued cash
or voucher transfer
No pure comparison group
• Assess impact on purchasing
decisions, consumption, other indicators of wellbeing
• Investigate mechanisms
Jenny C. Aker Cash and of each modality
• Calculate the cost-effectiveness Vouchers in DRC
8. Preview of Findings
• Starkly different purchasing patterns between the
two modalities
Voucher transfer extra-marginal for some food items
• Few differential effects on other outcomes (food
security, food expenditures, assets, coping
strategies)
Cash households were able to save more of their transfer
• Different purchases primarily due to program
design, but voucher could be resold
• Costs of voucher program higher for implementing
agency (and potentially program recipients)
Jenny C. Aker Cash and Vouchers in DRC
9. Related Literature
• In-kind transfers
Food stamps in the US (Moffitt 1989, Fraker et al
1995, Whitmore 2002, Hoynes and Schazenbach 2009)
• Unconditional cash transfers
Increased food consumption (Hoddindott and Skoufias
2004, Attanasio and Mesnard 2005, Maluccio 2007), improved
child health (Gertler 2004)
• Conditional versus unconditional cash transfers (Baird et al
2011)
• Cash versus food transfers (Del Ninno and Dorosh
2003, Ahmed et al 2009, Cunha 2012, Hoddinott et al
2013, Gilligan et al 2013)
• Cash, food and vouchers (Hidrobo et al 2012)
Jenny C. Aker Cash and Vouchers in DRC
10. Democratic Republic of Congo
• Civil war since the 1990s
“Africa’s world war”
Over three million people killed
Peace deal in 2003, but continued fighting and millions internally
displaced (1.7 million as of July 2011)
• Internally displaced populations live in host communities
and/or camps
Few income-generating opportunities, little access to basic needs
International organizations provide in-kind transfers (seeds and
tools, non-food items, vouchers) and cash (primarily cash-for-work)
• Limited investment in basic infrastructure
(roads, power, landlines, mobile phones)
Markets still function (three markets within a 20-km radius of the
camp)
Jenny C. Aker Cash and Vouchers in DRC
11.
12. Transfer Program
• Implemented by Concern Worldwide in one
informal camp (Bushani)
All households in the camp eligible for the transfer
Transfer provided to the woman within the household
• US$130 transfer provided in three installments
between September 2011 and March 2012
US$90 provided in September 2011, US$20 provided in
October 2011 and US$20 in March 2012
Provided to increase asset accumulation (NFIs) and
meet food needs
All households had to travel to Masisi (3 hours’ walk) to
obtain their transfer
Jenny C. Aker Cash and Vouchers in DRC
14. Research Design
• 474 households stratified by neighborhood and randomly
assigned to one of two transfer modalities
• T1: Unconditional cash transfer. Cash transfers provided in
three installments
• T2. Voucher. Vouchers provided in three installments
First voucher could be spent on food and non-food items at a
“multisectoral fair”
Second and third vouchers could only be spent on food items
• Equivalently-valued transfers provided at same time and same
amounts
• Both groups had to travel to main urban center (Masisi, 15 km
from camp) to pick up their transfer
• Voucher households informed of the restriction in advance
• No pure comparison group
Jenny C. Aker Cash and Vouchers in DRC
16. Voucher Fair
• First voucher fair: Included over 122 vendors
and four primary schools in the area, and
provided access to NFIs, household
items, clothes, school fees, agricultural inputs
(seeds and tools), and small animals.
• Second and third voucher fairs: Vouchers
could be spent on food at “open
markets”, whereby program recipients could
circulate freely among pre-arranged
boutiques and kiosks.
o Vouchers could not be used to purchase meat and
condiments.
19. External Validity
• Design similar to other voucher programs in
humanitarian contexts
• Differs from “traditional” voucher transfers by:
• Timing – had to be spent in one day
• Location – could only be spent at voucher fair
(rather than vendors, kiosks or marketes in
different locations)
22. Voucher with no resale
A’’: Voucher is extra-marginal and binding
qNFI
A’
A’’
B’: Voucher is infra-marginal
A
B’
B
qF=T/pFood
qFood
Jenny C. Aker Cash and Vouchers in DRC
23. Voucher with resale
A’’: Voucher is extra-marginal and binding
qNFI
A’
A’’’
A’’’: Voucher is extra-marginal and non-binding
A’’
B’: Voucher is infra-marginal
A
Can also affect
preferences and make
the goods more
substitutable
B’
B
qF=T/pFood
qFood
Jenny C. Aker Cash and Vouchers in DRC
24. Theoretical Predictions
• If transfers are extra-marginal, then there should
be differential effects of cash and vouchers on
household purchases
• If vouchers are binding, then there should be
differential effects of cash and vouchers on
consumption
• But would these differences lead to improved wellbeing?
Jenny C. Aker Cash and Vouchers in DRC
25. Data
Data
• Panel survey of 250 households over three rounds
(September 2011, November 2011 and March 2012)
Data on uses of transfer, assets, income sources, food
security
No full expenditure or consumption module but income and
food expenditures from the previous week
Extensive margin of purchases for each transfer, and
quantities purchased for subset of goods
• Price data for 40 products on the primary market and
voucher fairs
• Exit surveys from voucher fairs and cash transfer recipients
on purchases
• Cost data for each transfer modality
Jenny C. Aker Cash and Vouchers in DRC
25
28. Variables
Table 1. Comparison of Pre-Program Characteristics
Voucher
Cash
Mean (s.d.)
Panel A: Socio-Demographic Characteristics
Household size
Number of children (less than 15 years of age)
Program recipient is married
Program recipient is widowed
Age of program recipient
Program recipient has some education
Program recipient born in Masisi Territory
Number of years living in the camp
Mean (s.d.)
5.40
(2.00)
3.23
(1.87)
0.69
(0.42)
0.21
(0.35)
34.15
(14.19)
0.50
(0.50)
0.96
(0.19)
1.53
(0.76)
5.55
(1.85)
3.21
(1.58)
0.78
(0.42)
0.15
(0.41)
34.87
(13.12)
0.48
(0.50)
0.87
(0.33)
1.42
(0.82)
p-value
Jenny C. Aker Cash and Vouchers in DRC
0.314
0.888
0.285
0.272
0.503
0.767
0.04**
0.867
29. Table 1. Comparison of Pre-Program Characteristics
Variables
Voucher
Cash
Panel B: Income and Income Sources
Number of income sources
1.73
1.82
(0.72)
(0.93)
Total income earned during the past week (Congolese Franc)
2387
2491
(4610)
(4863)
Value of food purchases in the past week (Congolese Franc)
1883
1729
(4531)
(1536)
Panel C: Agricultural Production and Livestock
Had access to land
0.02
0.02
(0.15)
(0.13)
Owned poultry
0.02
0.04
(0.12)
(0.20)
Panel D: Asset Ownership
Total value (USD) of assets
62.30
60.60
(24.52)
(25.50)
Number of durable goods categories owned
0.01
0.01
(0.09)
(0.09)
Number of non-durable goods categories owned
10.87
11.00
(3.58)
(3.55)
Panel E: Food Security
Household diet diversity score (out of 12)
2.77
3.04
(1.82)
(1.67)
Number of meals in last day (household
1.27
1.29
(0.58)
(0.47)
Number of meals eaten in last day (children)
1.29
1.29
(0.59)
(0.49)
Suffered from food insecurity since last harvest
0.99
0.99
(0.09)
(0.09)
Months of adequate food provisioning
1.57
1.86
(1.23)
(1.16)
Number of observations
133
120
Well-balanced
Jenny C. Aker Cash and Vouchers in DRC
p-value
0.345
0.885
0.678
0.863
0.284
0.821
0.834
0.453
0.204
0.908
0.885
0.838
0.042**
253
30. Empirical Strategy
'
yi = g + acash + Xi0 l + q N + ei
• yi is outcome of household i (uses of the transfer, quantities
purchased, food security and asset ownership)
• cashi is the treatment status indicator of household i
• Xi0 is vector of baseline covariates
• θN is neighborhood fixed effects (level of stratification)
• εi individual shocks or ability
• Use pooled data (November and March) and control for round
fixed effects
• Also control for baseline outcome variable as a robustness
check (MacKenzie 2012)
• Main threats to identification: Attrition, spillovers, multiple
hypothesis-testing
Jenny C. Aker Cash and Vouchers in DRC
30
31. Are Transfers Infra- or Extra-Marginal?
• Quantity of goods purchased by cash transfer
households as compared with voucher households:
Identifies whether transfer is infra- or extra-marginal
• Quantity of goods purchased versus consumed for
voucher households: Identifies whether transfer is
binding or non-binding
Jenny C. Aker Cash and Vouchers in DRC
31
32. Total Food Expenditures
Voucher inframarginal for 50% of
households
2346 FC =
Weekly value of of
Weekly value
voucher: 2400 FC
Voucher
Cdf of weekly household food expenditures:
Cash Transfer Group
Jenny C. Aker Cash and Vouchers in DRC
33. Table 2: Uses of the Transfer
Panel A: Food Items
Number of different purchases made (last transfer)
Staple grains (maize, maize flour)
Other grains (cassava flour, rice)
Extramarginality for
some food
items
Stronger
differences for
second and
third transfers
Beans
Condiments
Oil
Meat
Vegetables
Salt
Fish
Panel B: Agricultural Items
Livestock
Seeds
Panel C: Other Non-Food Items
Clothing
Housing Materials
Panel D: Education and Health Expenditures
School fees
Medicines
Reimburse debts
Observations
All Transfers
(1)
(2)
Voucher
Cash
Mean (s.d.)
Coeff(s.e.)
4.94
3.07***
(2.75)
(0.33)
Third Transfer
(3)
Cash
Coeff(s.e.)
4.03***
(0.45)
0.49
(0.50)
0.73
(0.45)
0.27
(0.45)
0.15
(0.36)
0.56
(0.50)
0.03
(0.18)
0.08
(0.26)
0.93
(0.26)
0.45
(0.50)
0.25***
(0.06)
-0.12**
(0.06)
0.22***
(0.06)
0.27***
(0.05)
0.21***
(0.06)
0.65***
(0.04)
0.36***
(0.05)
-0.12**
(0.05)
-0.02
(0.08)
0.25***
(0.07)
-0.10*
(0.08)
0.38***
(0.07)
0.26***
(0.05)
0.27***
(0.08)
0.55***
(0.06)
0.35***
(0.06)
-0.12**
(0.05)
-0.02
(0.08)
0.09
(0.30)
0.37
(0.48)
0.05
(0.04)
-0.06
(0.06)
0.08**
(0.03)
0.03
(0.05)
0.38
(0.49)
0.232
(0.42)
0.26***
(0.06)
0.15***
-0.06
0.42***
(0.06)
0.11***
(0.03)
0.28
(0.45)
0.01
(0.12)
0.301
(0.46)
0.42***
(0.06)
0.08***
(0.03)
0.31***
(0.06)
308
0.64***
(0.06)
0.05*
(0.03)
0.43***
(0.06)
178
34. Uses of the Transfer (Second)
100.0%
90.0%
80.0%
Cash
Voucher
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
35. Are Specific Items Extra-Marginal?
• Among food items, salt and rice were more likely to
be purchased by voucher households (for all three
transfers, but stronger for second and third
transfers)
o Fish not statistically significant across all rounds, just
second round
• This gives a sense of the extensive margin of overprovision
• What about intensive margin of over-provision?
Jenny C. Aker Cash and Vouchers in DRC
35
36. Intensive Margin of Overprovision
Table 3: Quantities Demanded of Specific Food Items
(1)
(2)
Voucher
Cash
Mean (s.d.)
Salt (kg)
Coeff(s.e.)
1.77
-0.57***
(0.18)
1.75
-1.41***
(2.52)
Number of observations
(2.64)
(1.40)
Fish (number)
-9.97***
(18.57)
Rice (kg)
12.34
(0.32)
130
130
Jenny C. Aker Cash and Vouchers in DRC
37. Are the Transfers Binding?
• Individual welfare losses associated with overprovision
can be mitigated if the transfer is non-binding
• To measure whether the transfer is binding, we want to
measure the difference between purchases and
consumption for extra-marginal goods
• No data on actual quantities consumed, just whether
consumed
• (Given quantities of fish and salt, we could estimate daily
consumption of these goods between last transfer and
food recall period, assuming transfers were binding)
Jenny C. Aker Cash and Vouchers in DRC
38. Table 5: Food Security
All Transfers
(1)
Voucher
Mean (s.d.)
Panel A: Household Diet Diversity
Household diet diversity (out of 12)
Data were
collected 3
weeks after
each
transfer, so
difficult to
separate
between
resale, storage
or lumpy
consumption
Grains
Tubers
Beans
Vegetables
Fruits
Fats
Eggs
Milk
Meat
Fish
Condiments
Sugar
(2)
Cash
Coeff(s.e.)
Third Transfer
(3)
Cash
Coeff(s.e.)
2.62
(1.91)
0.708
(0.46)
0.736
(0.44)
0.222
(0.42)
0.618
(0.49)
0.041
(0.20)
0.416
(0.49)
0
(0.00)
0
(0.00)
0.06
(0.23)
0.153
(0.36)
0.013
(0.12)
0.298
(0.46)
0.09
(0.19)
-0.02
(0.06)
-0.00
(0.06)
0.02
(0.05)
0.03
(0.06)
-0.01
(0.03)
0.08
(0.06)
0.02
(0.01)
0.00
(0.00)
0.02
(0.03)
0.03
(0.05)
-0.01
(0.02)
-0.04
(0.05)
-0.07
(0.23)
0.02
(0.08)
-0.01
(0.08)
-0.01
(0.07)
-0.01
(0.08)
-0.03
(0.04)
0.02
(0.08)
0.01
(0.01)
0.00
(0.00)
-0.03
(0.03)
0.04
(0.06)
-0.04
(0.03)
-0.02
(0.07)
39. Did the Vouchers lead to Differential
Effects on Well-Being?
• Food security (number of meals per day, number of
months of adequate household food provisioning,
food insecurity since previous harvest)
• Asset ownership and savings (durable and nondurable goods categories)
• Agricultural assets (land, livestock)
• Coping strategies
Jenny C. Aker Cash and Vouchers in DRC
40. Food Security
Table 5: Food Security
All Transfers
Third Transfer
(1)
(2)
(3)
Voucher
Cash
Cash
Mean (s.d.)
Coeff(s.e.)
Coeff(s.e.)
1.41
0.01
-0.03
(0.28)
(0.07)
(0.10)
2.12
-0.26
-0.51
(2.79)
(0.31)
(0.51)
2.06
0.13
0.08
(0.84)
(0.09)
(0.12)
308
178
Panel B: Other Measures of Food Security
Number of meals per day (household)
Number of meals per day (children)
Months of adequate food provisioning
Number of observations
Jenny C. Aker Cash and Vouchers in DRC
41. Income and Asset Ownership
Table 5: Proxy Indicators of Well-Being
All Transfers
(1)
(2)
Voucher
Cash
Mean
(s.d.)
Coeff(s.e.)
Panel A: Income, Expenditures and Savings
Income in the previous week (Congolese Francs)
Amount spent on food in previous week
(Congolese Francs)
Money left from transfer (Congolese Francs)
Amount of money remaining (savings)
Panel B: Assets
Total value of household assets (USD)
Number of durable assets owned
Number of non-durable assets owned
Own poultry
Own rabbits
Third
Transfer
(3)
Cash
Coeff(s.e.)
3,357
(2938.83)
739
(639.74)
834.22
(1,039.19)
2,601
(2371.64)
0.00
(0.00)
0.00
(0.00)
-119
(274.58)
0.09***
(0.03)
1,787**
(704.44)
-336.90
(373.81)
0.07**
(0.03)
1,130**
(513.64)
89.07
(35.20)
0.00
(0.00)
10.62
(7.03)
0.1294
(0.34)
0.0117
(0.11)
-1.15
(4.53)
-0.01
(0.02)
0.33
(0.67)
-0.04
(0.04)
0.02
(0.02)
0.36
(5.09)
0.01
(0.01)
0.44
(0.91)
-0.08
(0.05)
0.04
(0.03)
Jenny C. Aker Cash and Vouchers in DRC
42. Coping Strategies
Table 5: Proxy Indicators of Well-Being
All Transfers
Third Transfer
(1)
(2)
(3)
Voucher
Cash
Cash
Mean (s.d.)
Coeff(s.e.)
Coeff(s.e.)
0.0161
-0.02
-0.01
(0.13)
(0.02)
(0.01)
0.27419
0.03
0.07
(0.45)
(0.06)
(0.08)
0.0161
-0.02
-0.01
(0.13)
(0.02)
(0.02)
308
178
Panel C: Coping Strategies
Sold household goods
Reduced number of meals per day
Took children out of school
Number of observations
Jenny C. Aker Cash and Vouchers in DRC
43. Summary
• Different purchases in both groups – voucher
households purchased more rice, salt and perhaps
fish
No differential effects on individual food consumption
• Proxy indicators of well-being the same between
the two groups
However, cash households saved some of their transfer
and reported suffering less from food insecurity
Jenny C. Aker Cash and Vouchers in DRC
44. Mechanisms
• Why did purchases differ between the two groups?
• Vouchers restricted to food fairs for last two transfers
Voucher households had to purchase items in food fairs on
that day (location and timing of purchases)
Voucher households would have had to pay transport costs
for many (or heavy) items
• Voucher households had more (less ) bargaining
power vis-à-vis traders – had to pay higher prices
• Women in voucher households had more (less)
bargaining power within the household
Jenny C. Aker Cash and Vouchers in DRC
44
45. Table 7: Mechanisms on Purchases
(1)
Voucher
Mean (s.d.)
Panel A: Location of Purchases
Boutique in camp
Market outside camp
Masisi Center market (20 km from camp)
Nyabiondo market (2 km from camp)
Hospital
Voucher fair
School
Panel B: Timing of Purchases
Spent money in more than one purchase
(2)
Cash
Coeff(s.e.)
0.01
(0.11)
0.02
(0.12)
0.00
(0.00)
0.00
(0.00)
0.00
(0.00)
0.99
(0.12)
0.00
(0.00)
0.00
(0.01)
0.97***
(0.02)
0.45***
(0.08)
0.46***
(0.08)
0.01
(0.01)
-0.99***
(0.01)
0.01
(0.01)
0.013
(0.12)
0.77***
(0.04)
0.94
(0.24)
0.46
(0.50)
0.06
(0.24)
-0.02
(0.03)
-0.03
(0.06)
-0.02
(0.03)
0.79
(0.40)
-0.03
(0.05)
308
Panel C: Intra-Household Decision-Making with Respect to Transfers
Beneficiary responsible for spending all or part of transfer
Husband responsible for spending transfer
No one else responsible for spending transfer
Discussed how to use transfer in advance with other person
Number of observations
Jenny C. Aker Cash and Vouchers in DRC
46. Mechanisms
If something was too heavy, I didn’t buy it…I wanted to
buy two boxes of salt but could only carry one, so I
bought one plus other things.
Voucher program recipient, Masisi camp
Jenny C. Aker Cash and Vouchers in DRC
47. Mechanisms
• Why did consumption not differ between the two
groups?
• Time delay between purchases and consumption
Storage or lumpy consumption
Wouldn’t explain lack of difference in salt consumption
• Resale of goods purchased
Almost all households stated that they purchased the
overprovided goods for resale
• Sharing of goods (and cash) between the two groups
Jenny C. Aker Cash and Vouchers in DRC
47
48. Table 8: Mechanisms on Outcomes
Sharing of
money and
goods
(1)
Voucher
Mean (s.d.)
Panel A: Transfers are Non-Binding
Program recipient shared part of money received
Program recipient shared part of goods purchased
No
differential
effects on
intrahousehold
decisionmaking
0.25
(0.44)
0.46
(0.50)
Panel C: Intra-Household Decision-Making with Respect to Transfers
Husband makes education decisions alone
0.3
(0.46)
Husband makes agriculture decisions alone
0.26
(0.44)
Husband decides whether to share with other households alone
0.16
(0.04)
Husband decides whether/how to save alone
0.31
(0.47)
Ratio of women to men's clothing
2.32
(1.58)
Number of observations
Jenny C. Aker Cash and Vouchers in DRC
(2)
Cash
Coeff(s.e.)
0.11*
(0.06)
-0.15**
(0.06)
0.05
(0.06)
0.05
(0.06)
0.03
(0.06)
0.02
(0.06)
-0.08
(0.40)
308
50. Threats to Identification
•
•
•
•
•
Differential take-up or leakage
Differential shocks
Differential price effects
Spillovers
Multiple hypothesis-testing
Jenny C. Aker Cash and Vouchers in DRC
50
51. Table 9: Alternative Explanations
(1)
(2)
Voucher
Cash
Mean (s.d.)
Coeff(s.e.)
Panel A: Take Up and Leakage
Received transfer
2.98
-0.01
(0.01)
80,825.00
1,142.94
-6042
(717.43)
18,352.54
159.25
(167.97)
(229.52)
0.54
0.07
(0.05)
(0.07)
0.27
0.04
(0.05)
(0.06)
0.08
-0.00
(0.03)
Amount received (second transfer)
(0.00)
(0.03)
Amount received (first two transfers)
-0.00
(0.00)
Number of transfers received
1.00
(0.04)
Panel B: Illness and Death
Affected by conflict
Household member affected by illness
Household member died
Number of observations
308
52. Differential Price Effects
• Impact of cash on vouchers:
Cash transfers relatively small (total of 474 households and
USD$60,000 over six months, amidst IDP population of
60,000 in the area)
Purchases occurred in three different markets 15-20 km
apart
Maximum prices at voucher fair were based upon recent
market prices for the same goods and same market
• Impact of vouchers on cash:
If vouchers increased prices, consumers could have chosen
other markets closer to the camp
Jenny C. Aker Cash and Vouchers in DRC
52
53. Threats to Identification
• Spillovers
Spillovers less of an issue for purchasing decisions (as
compared with other outcomes)
Would be less likely to observe a difference in purchasing
decisions
• Multiple hypothesis-testing
Bonferroni correction on families of outcomes (Sankoh et al
1997)
Most differences remain statistically significant for extensive
margin (for cash households) and intensive margin (for
voucher households)
Jenny C. Aker Cash and Vouchers in DRC
53
54. Cost Effectiveness Analysis
• Costs to implementing agency
• Costs to program recipients
Jenny C. Aker Cash and Vouchers in DRC
54
55. Costs per Program Recipient
$16.00
Transfer Fees
Account Opening Costs
$14.00
Voucher printing
Materials (plastic sheeting, sticks)
$12.00
Transport (fuel, lodging)
Staff time
$10.00
$8.00
$6.00
Cost per recipient is US$3 more for the
voucher program
Even if double the number of recipients
per voucher fair, the voucher fair still
US$1 more
$4.00
$2.00
$0.00
Cash
Voucher
Jenny C. Aker Cash and Vouchers in DRC
56. Cost Effectiveness Analysis
• Welfare cost (loss) to voucher recipients due to changes
in purchasing decisions
Mitigated by the fact that vouchers and goods could be sold
• Similar travel and opportunity costs for each modality
Both groups had to travel to Masisi Center (3-4 hour walk), but
only voucher recipients had to spend there
• Potentially higher security risks for voucher recipients
Had to travel and spend on particular days
Goods purchased were more observable to looters and militia
than cash
Specifically an issue with gender-based violence in eastern DRC
Jenny C. Aker Cash and Vouchers in DRC
57. Conclusions
• Voucher changed household purchases for some
goods
Voucher households purchased more salt, fish and rice than
they would have if given cash
Cash households purchased more diverse food items
(overall), school fees, medicines and debt reimbursement –
not “temptation” goods
• No differential effects on welfare
Non-binding transfers – could sell goods
• Cash transfers more cost effective
Jenny C. Aker Cash and Vouchers in DRC
58. Conclusions
• If markets can provide necessary goods, cash is a
preferred option from a welfare perspective
Might be other reasons to provide vouchers
• Environment with high marginal utility of income
• But how much cash and how often?
“If we receive (a large sum) of cash all at once, I will
spend it on silly things…like beer, or beers for my
friends, or cookies for the children….”
Cash and voucher recipients and spouses, Masisi camp
Jenny C. Aker Cash and Vouchers in DRC
59. External Validity
• Variety of voucher and cash-based programs in
conflict and emergency situations
DRC, Sudan, Niger
• Availability of supplies on local markets and selftargeting important
Poor and vulnerable populations are easily identified with
IDPs – more difficult among host communities
Markets were able to provide goods
Jenny C. Aker Cash and Vouchers in DRC
Notas do Editor
Good afternoon, and thank you for the opportunity to present here today. I am very excited to be presenting on this topic, not only because this is very much a new paper, but also because I like this area of research. I am looking forward to your comments, suggestions and criticisms.
Governments, economists and development practitioners alike have long recognized the importance of redistributing wealth to the poor. The key question, of course, is what is the best way to do that.While conditional and unconditional cash transfer programs have become an increasingly important part of social protection programs in both developed and developing countries since the 1990s, a majority of these programs are still in-kind – such as food aid, food stamps, medicines, and inputs – especially in humanitarian situations.Even within the US, more than $100 billion is designated for “food and nutrition” in the president’s 2013 budget,including food stamps, Medicaid and housing vouchers, relative to cash- based welfare programs. http://www.bloomberg.com/news/2012-02-28/cash-better-than-food-stamps-in-helping-poor-commentary-by-edward-glaeser.html
And in international aid programs, over 58% of USAID’s emergency budget is allocated to in-kind transfers and vouchers, and over 25% of that is vouchers.Hanrahan, Charles E. "International Food Aid Programs: Backgrond and Issues." Congressional Research Service, May 20, 2013, p. 10http://www.fas.org/sgp/crs/misc/R41072.pdf
In light of the fact that program recipients would weakly prefer an equal-valued cash transfer to an in-kind transfer – since it offers the same, or more, budgetary choices – why would governments not use cash? There are multiple reasons put forth for this, from targeting to paternalism to local supply.Yet despite the importance of this question for policymakers and donors alike, very little credible evidence exists for any in-kind program, and particularly comparing those with cash transfers. This is especially the case in developing countries.
This paper looks at the case of one developing country, DRC, one of the largest countries in Africa and, despite its vast natural resources, one of the poorest countries in the world. DRC, and especially the eastern part of the country, has been in the midst of a civil war since the 1990s, with millions killed and internally displaced. Despite a peace deal in 2003, internal fighting remains, with 1.7 million IDPs in eastern Congo as of 2011.
These IDPs are living in both formal and informal camps throughout eastern DRC. In response to the conflict, a variety of international organizations have used in-kind transfers – primarily food aid, seeds and tools distribution and, more recently, vouchers – with the goal of enabling households to meet their basic food and non-food needs.
And that is really the goal of this research. We are interested in understanding the relative effects of cash versus in-kind transfers.
Theory: Only extra-marginal and binding transfers will differentially influence purchases Household asset ownership was worth 60$, and increased by 31% (to 80$) after the program.
Baird et al show that the CCT reduced dropout as compared with the UCT, but no improved outcomes.
Bushani is a camp located about 20 km from Masisi center – 2 hours by car, 4 hours by foot.Population of approximately 2,500 households.Living in camp since 2009. The total value of the transfer was approximately 2/3 of the total annual GDP per capita for DRC, similar to the value of other income transfer programs in DRC and other humanitarian contexts.
There were 8 blocs (neighborhoods) in total.Since we have no pure comparison group we cannot estimate the impact of each program or draw Engel curves.
The multisectoral fair included over 122 vendors and four primary schools in the area, and provided access to NFIs, household items, clothes, school fees, agricultural inputs (seeds and tools, and small animals. A full list of items available at the multisectoral fair is available upon request. As no vendors were willing to sell plastic sheeting or mosquito nets, Concern purchased and sold these items at the fair. Program recipients could purchase school fees for either the entire year or on a semester basis.The second and third vouchers could be spent on food at “open markets”, whereby program recipients could circulate freely among pre-arranged boutiques and kiosks selling foodstuffs in the Masisi Centre market. For all of the voucher fairs, program recipients had to travel to Masisi Centre on a pre-arranged day to receive and spend their voucher. Eleven food vendors were eligible to participate at the second food voucher fair, and vendors were unwilling to provide palm oil. For this reason, Concern purchased and sold palm oil at local prices at the second food voucher fair. Additional items on the second food fair included sugar, cassava flour, beans, rice, vegetable oil, dried fish and salt. The third food voucher fair included 18 food vendors and the same food items, in addition to potatoes and peanuts.
First, in-kind transfers may induce the non-poor to self-select out of welfare programs, thereby assisting in targeting. Second, in certain contexts, food and non-food items might not be available on local markets, necessitating in-kind transfers to overcome these market failures. Or, in-kind transfers may be more politically feasible than in-kind transfers. Perhaps the most often cited rational for in-kind over cash transfers, however, is one of paternalsm. Finally, vouchers might be less likely to steal.A paternalistic government issues transfers in kind to encourage the consumption of specific transferred goods.
First, in-kind transfers may induce the non-poor to self-select out of welfare programs, thereby assisting in targeting. Second, in certain contexts, food and non-food items might not be available on local markets, necessitating in-kind transfers to overcome these market failures. Or, in-kind transfers may be more politically feasible than in-kind transfers. Perhaps the most often cited rational for in-kind over cash transfers, however, is one of paternalsm. Finally, vouchers might be less likely to steal.A paternalistic government issues transfers in kind to encourage the consumption of specific transferred goods.
Assume households have preferences for two goods, say, food and non-food items, and they maximize a utility function U (qF, qNF), strictly increasing and concave in both arguments. Pre-transfer, households have a budget constraint, and pF and pNF are the market prices of the two goods – this shows the indifference curves of two types of households here.
A cash transfer of T shifts the budget constraint up to CE, with resulting increases in consumption at A’ and B’.
While an equal-valued in-kind transfer of food – either food aid, or in this case, a voucher – leads to a kinked budget constraint that depends on the resale price of food items. Now, if frictionless resale is unavailable, then cash is weakly preferred to the transfer in-kind, if we compare the two types of households. The first household (B) is indifferent between the transfer type, moving from indifference curve B to B’ under either transfer modality. The second household, though, is weakly worse off under the in-kind transfer, consuming at A’’ (the kink), while it would have chosen A’ under the cash transfer.
Now, if resale is allowed – even with some costs, -- this will shift up the kink and allow households to reach a higher level of utility. This means that the voucher for household A will be non-binding – in other words, the household consumes less of the good than it was provided (or purchased) under the voucher, and binding otherwise.There can be other potential reasons that the voucher is non-binding besides resale – such as, perhaps, intra-household dynamics or inter-household sharing – but this gives you the idea.
I am going to skip most of the information on timeline, etc. Basically, there is a baseline survey and one post-treatment round of the survey. For each of the two intervention there are approximately 100 villages with 11 respondents each per village. The same survey instrument was used.We collected information on migration and migration outcomes, as well as data on household demographics, and mobile phone usage. Note: We did not collect data on income and expenditures, or the amount spent on each product, as pre-tests suggested that this was subject to a high degree of measurement error. Even if this measurement error was classical, which lead to unbiased and efficient results, the standard errors would be more precise, which would make it be more difficult to detect an effect.
Significant. Some differences between attriters and non-attriters, but overall the non-attriters were balanced between the two.Had a sample of about 200 people by March 2012, only about 130 people for November 2011.
First, it is important to note that in terms of total food consumption, the vouchers are in general infra-marginal for about 50% of households. That is, with a cash transfer equal in value to about 2400 FC per week for the last two transfers (which could only be spent on food items), about 50% of households consume more than that.
So if we look at that more formally again, we see that a simple OLS yields similar results. This is robust to the use of a Poisson.
Resale of the goods would detract from a pure paternalistic motive if it’s to increase food consumption, whereas lumpy consumption supports the paternalistic motive – again if it is oriented towards consumption.
So, as a proxy, we simply look at whether households consumed those goods in the previous 24 hours. It doesn’t appear as if any of these purchases remained binding, although this could be due to resale,
“Traders knew that we received the vouchers for free…so when we came to buy from them, they were unwilling to negotiate (the price) with us.”Voucher program recipient, Bushani Camp