Agriculture Reference
In-Depth Information
TABLE 5.11
Percentage leakage in targeted program delivery
a
World
Food
BIDS/
World
Targeted program
IFPRI
Programme
CARE
IFPRI
Bank
Vulnerable Group
14 (1994)
8 (1994)
41 (2002)
Development
8 (2003)
b
13 (2003)
c
VGF
35 (2002)
Rural Maintenance
Program
0 (1994)
Food-for-Work (CARE)
36 (1994)
Food-for-Work (World
Food Programme)
28 (1994)
Food for Education
7 (1994)
75 (2002)
Integrated Food
6 (2003)
d
Security
19 (2003)
e
SOURCES
: Ahmed (2000); World Bank (2002); Ahmed et al. (2003).
NOTES
: Blank cells indicate that there were no studies and hence no statistics are available. BIDS, Bangladesh
Institute of Development Studies; CARE, the international relief and development organization; IFPRI, Inter-
national Food Policy Research Institute.
a
Figures in parentheses are years of study.
b
Based on physical weighing of wheat received (excluding shared rations).
c
Based on beneficiary reported weight of wheat received (excluding shared rations).
d
Based on physical weighing of wheat received.
e
Based on beneficiary reported weight of wheat received.
The results suggest that, of all VGD beneficiary households, only 38.4 per-
cent belong to the first quartile group, that is, the poorest 25 percent of house-
holds. An immediate interpretation of this result is that the program has prob-
lems of errors of inclusion. However, when compared to the available estimates
of poverty, the picture does not look so grim. For instance, the most recent
poverty estimates in Bangladesh suggest that, in 2000, 53.1 percent of the ru-
ral population was below the poverty line, based on the cost-of-basic-needs
method of measuring the incidence of poverty (BBS 2003). Therefore, if the
bottom 50 percent of all households is considered to be the poor, the results sug-
gest that 62.1 percent of the beneficiaries were poor and had been correctly se-
lected as program beneficiaries. Furthermore, the average monthly per capita
expenditure of VGD beneficiaries is about 24 percent lower than the average
monthly income of nonbeneficiary households. Using these two sets of num-
bers, one can argue that the VGD targeting is pro-poor.
Furthermore, an analysis of the expenditure data of nonbeneficiary house-
holds indicates that the program also has high errors of exclusion. For example,
