Using the Household Income and Expenditure Survey, this document provides some evidence-based solutions regarding SSNPs in Bangladesh

1. `
Food Based Social Safety Net Programs
From problems to solution

2. Social Safety Nets
Programs that are mainly undertaken by the government to assist the poor and the disadvantaged
of the country are called social safety net programs (SSNPs).
According to Household Income and Expenditure Survey, the beneficiaries of SSNPs in
Bangladesh are very poor, widow/separated, landless, non-earner, crippled, disabled/ill, old age,
freedom fighter etc.
Chart 1:
Source: Household Income and Expenditure Survey
24.57 percent of the households reported to have received benefit from at least one type of program
in 2010. Only 13.03 percent households reported to have received benefit from SSNP in 2005.
There was similar increase in urban areas (HIES, 2010).
Classifications of SSNPs:
Chart 2:
Source: HIES
There are 30 types of SSNPs are in operation in Bangladesh. Chart 2 shows that maximum
people are benefited with conditional and unconditional cash benefits.
13.03 15.64
24.57
30.12
0
10
20
30
40
Total Rural
Beneficiary Households (%)
2005 2010
2.2
27.4
5.5
24.3
22.4
1.0
16.5
0.0 5.0 10.0 15.0 20.0 25.0 30.0
Conditional food
Conditional Cash
Unconditional Food
Unconditional Cash
Fund
Credit
Conditional Subsidy
% of Beneficiaries in each kind of SSNP

3. Source: Ahmed (2007)
SSNPs
ConditionalUnconditionalFundCreditSchemes
Food
Conditional
Subsidy
CashFoodCash
VGF
GR
VGD
Community
Nutrition
Widow
Freedom
Fighter
OldAge
Disabled
FFW/
CFW
TR
CtgHill
PESP
FSSAP
REOPA
Disabled
Child
Orphan
Maternity
Children
atRisk
Natural
Disaster
Emergency
Relief
Housingfor
Distressed
AcidBurnt&
Physically
Handicapped
Fertilizer&
Electricity
Marginal
Farmers
FoodSubsidy
RuralSocial
Service
Trainingfor
freedom
fighters
Rural
Mothers
Center
Disadvantaged
Women&
Children
Urban
Community

4. Causes of not Participating in SSNPs:
There are many people who are fit for the program but not included for the following reasons:
1. Did not know about the program
2. Fit for the program but not apply
3. Shortness of budget
4. Selection was not proper
5. Not any program in that area
Chart 3:
S Source: HIES
60.44 percent of people are not included in SSNP due to the selection bias. 12.39 percent did not
even know about the program, 11.23 percent did not apply, 10.72 percent had shortness of budget
and 5.22 percent were not included because there was no program in their area.
Cash vs Kind:
Chart 4:
From the beneficiaries, 64 percent are
receiving in cash, 30 percent are receiving
in kind and almost 3 percent are receiving
in both. Other 3 percent are entitled but not
receiving.
Source: HIES
Cause Frequency
Did not know about the program 2045
Fit for the program but not apply 1853
Shortness of budget 1769
Selection was not proper 9975
Not any program in that area 861
12.39
11.23
10.72
60.44
5.22
Causes of not including
1 2 3 4 5
cash
64%
kind
30%
both
3%
none
3%
SSN Receiving Type
cash kind both none

5. Spending for Inclusion:
Among those who are participating in SSNP, 14.8 percent have spent for inclusion. The average
spending for inclusion is 37 taka approximately (HIES, 2010). This gives the credibility of 10.72
percent not being included due to the shortness of budget.
Selection Bias:
The beneficiaries are selected through some selection criteria. In programs targeting poor
households, beneficiaries are usually selected by local official or community leaders. Therefore, 2
types of selection bias may occur:
Source: Food Planning and Monitoring Unit, Ministry of Food, GoB (2014)
The criteria are based on several indicators under Socio-economic, Nutritional and Vulnerability
profiles (Rahman & Choudhury, 2012).
Poverty and Food Security
Food Poverty and Food Security:
A person whose daily calorie intake is lower than 2122 Kcal is considered in the category of
absolute poverty (HIES). Thus, a household whose per capita daily calorie intake is lower than
2122 Kcal is food insecure.
Ultra-poor:
Ultra-poor are those who eat below 80% of their energy requirements (=80% of 2122 Kcal= 1698
Kcal) or those who are living at less than half the $1.25-a-day poverty line. (BRAC, 2013).
Dietary Diversity:
Household Dietary Diversity Score (HDDS) is the number of distinct food groups consumed by a
household over a given period. DDS ranges between 0 to 11 if we calculate according to the
definition of FAO (2007).
Safety Net & Food Security
From HIES data, we can show the impact of safety net on food security, calorie intake, poverty
and household dietary diversity (HDDS).
Inclusion Error
Non-poor Beneficiary
Exclusion Error
Poor Non-beneficiary

6. Table 1:
(1) (2) (3) (4)
VARIABLES Food Security
(Probit)
0=insecure
1=secure
Per capita
daily Calorie
Intake
(OLS)
Poverty
(Ordered
Probit)
1=ultra-poor
2=poor
3=non-poor
HDDS
(OLS)
Per capita
expenditure on food
0.000827*** 0.280*** 0.000897*** 0.000317***
(1.49e-05) (0.00218) (1.36e-05) (4.69e-06)
HH size -0.0914*** -40.19*** -0.0628*** 0.204***
(0.00807) (2.418) (0.00690) (0.00521)
HHH in agriculture 0.407*** 177.6*** 0.401*** -0.0216
(0.0326) (9.957) (0.0283) (0.0214)
HHH age 0.0121*** 4.456*** 0.0104*** -0.00233***
(0.00105) (0.315) (0.000915) (0.000679)
HHH Female 0.0187 3.562 0.00978 -0.0182
(0.0463) (13.72) (0.0399) (0.0295)
Cultivable land 0.000201* 0.152*** 0.000135 0.000401***
(0.000112) (0.0301) (0.000103) (6.48e-05)
Microcredit 0.127*** 31.99*** 0.105*** 0.0956***
(0.0288) (8.747) (0.0250) (0.0188)
Safety net 0.0665** 19.64* 0.105*** 0.00267
(0.0333) (10.19) (0.0287) (0.0219)
Litracy rate -0.00365*** -1.511*** -0.00380*** 0.00643***
(0.000421) (0.126) (0.000362) (0.000271)
Sanitation 0.138*** 53.21*** 0.142*** -0.102***
(0.0104) (3.133) (0.00891) (0.00675)
Migration 0.135*** 38.61*** 0.152*** -0.146***
(0.0475) (13.80) (0.0418) (0.0297)
sUrban -0.531*** -216.0*** -0.491*** 0.243***
(0.0337) (9.933) (0.0285) (0.0214)
Constant cut1 2.654***
(0.121)
Constant cut2 3.906***
(0.123)
Constant -3.552*** 987.1*** 3.838***
(0.141) (38.85) (0.0836)
Observations 12,240 12,240 12,240 12,240
R-squared 0.635 0.481
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Source: Authors’ Calculation from HIES data (12,240 households).

7. Problem 1: Gap
Khaddo Bandhob Program Beneficiary= 5 million
rural households
Average household member= 4.5
Ultra-poor people in Rural Bangladesh= 10.6
million
Cost-effective Solution:
From table 1, we can consider some determinants of being poor/ultra-poor. To select beneficiaries
without any error, there are some selection criteria (suppose, like the determinants in the table)
which have to be met properly. US $ 2.86 billion were allocated for wrongly selected beneficiaries
during 2007-08 to 2011-12 time period (Haider & Mahamud, 2017). To remove selection bias
cost-effectively in Khaddo Bandhob Program, following measures can be taken:
Plan A Union Parishad is responsible for selecting the beneficiaries. Union chairperson and
members are elected by vote. Hence the program has some political dimension.
Moreover, the outreach of SSNPs are far below the requirement (Rahman &
Choudhury, 2012). So, corruption is inevitable. Socially responsible personnel
(such as Primary school teachers, Imam of the mosque) should select the
beneficiaries instead of politically elected ones to increase the trustworthiness.
Plan B If Govt decides to keep union parishad as the selection authority, they have to
increase their accountability. An effective monitoring and supervision system can
minimize the incidence of corruption. Implementation authority is often not well-
informed about selection criteria and sometimes updated list is not available (Haider
& Mahamud, 2017). To solve this, we have to review the selection criteria.
Targeting should be based on measurable indicators that can be operationalized
easily. Govt needs to promote digitation and give proper access of database to local
authority.
Plan C Govt needs to target the poorest region and individuals. In order to implement this,
local community verification should be mixed with region-based allocation. For
example, Khulna has experienced increased poverty rate, therefore, deserves highest
allocation but Sylhet receives the highest share of benefit and Khulna receives the
lowest (Rahman & Choudhury, 2012).
NB: Above plans can reduce inclusion error which does not need any new cost. The opportunity
cost of inclusion error will offset the cost of implementing plans mentioned above!
On the other hand, there must be some exclusion errors too. Actually, more people are wrongly
excluded than wrongly included (Food Planning and Monitoring Unit, 2014). Better targeting
mechanism can increase both coverage and level of benefit.
Ultra-poor
10.6 million
Beneficiary
(5x4.5=22.5)
million
=
Inclusion Error
11.9 million (53%)

8. Problem 2: Dealers’ Tendency
Option A- Improving present status:
Reducing the number of intermediaries can solve this problem to some extent. Through effective
monitoring system Govt can ensure the dealers provide proper amount of rice to the beneficiaries.
The loss will be reduced if the coordination among ministries in planning, targeting and
implementation improves. It improves delivery mechanism. 13 ministries are working together for
SSNPs which can cause overlapping in programs. Coordination improves the capacity of the
programs through solving the overlaps.
Option B- Introducing Cash Transfer with Proper Technology:
Cost of delivering cash is cheaper than the cost of delivering food to the beneficiaries (Ahmed,
2007). Beneficiaries also mostly prefer cash to kind. Keeping these in mind, Govt can introduce
cash transfer program instead of “Food Friendly Program” though the impact of cash transfer in
reducing food insecurity (poverty) is a matter of controversy (See the solution of problem 3).
In case of implementing cash transfer program, the following structure can be followed:
Here, the number of intermediaries is
very few. Cost is reduced. Beneficiary
receives benefit directly from the Govt.
The monitoring is also very easy as it is
an electronic survey. The reply
beneficiary sends will be included in
the national survey automatically.
[Currently, bKash is running a network
of more than 180,000 agents
throughout urban and rural areas of
Bangladesh with over 30 million
registered accounts. 68% Bangladeshis
have mobile phones and bKash can be
accessed via all the mobile networks
operating in Bangladesh. Source:
https://www.bkash.com]
Budget:
Source: Budget 2018-19.
Food Friendly Program contains 6% of the
total budget in Food Security Programs. For
option B, budget need not to be extended
because it requires less cost than food-based
program. It only has a fixed cost of bringing beneficiaries into the bKash coverage (opening bKash
account if needed). Moreover, with the existing budget, Govt can increase the efficient coverage
by correcting selection error. For option A, it needs a specific budget (let, 2% 638=12 crore taka)
for monitoring and coordination. The budget amount will return through successful loss reduction.
Budget (2018-19)
Taka in Crore
Food Friendly Program 638.00
Food Security Programs 10457.16
Bkash
h
Cash
Reply
Govt agent
Beneficiary
(gets SMS)
Bkash agent
Cash
Received?
Select option in
SMS (Yes/No)

9. Problem 3: Cash or Food?
Table 2
(1)
(1)
VARIABLES Ordered Probit
pc_expend food 0.000977***
(2.46e-05)
HH size -0.0548***
(0.0117)
HHH in agri 0.324***
(0.0457)
HHH age 0.0104***
(0.00156)
Female HHH 0.0484
(0.0649)
Cultivable land 0.000607**
(0.000253)
microcredit 0.121***
(0.0414)
Litracy rate -0.00309***
(0.000607)
sanitation 0.149***
(0.0151)
migration 0.0951
(0.0709)
Region -0.602***
(0.0501)
SSN (cash) 0.104**
(0.0515)
SSN (both) 0.189
(0.145)
SSN (kind) 0.149***
(0.0537)
Constant cut1 2.753***
(0.207)
Constant cut2 4.050***
(0.212)
Observations 4,307
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Source: Authors’ Calculation from HIES data
Table 2 shows the impact of participating in
cash transfer program or food based program
or both type of program on poverty alleviation
using ordered probit model where the
dependent variable has a natural ordering (1 if
ultra-poor, 2 if poor and 3 if non-poor). The
poverty is calculated based on the calorie
intake, hence it is food poverty. The
observations excluded target selection errors
(cause of not included: 2) and the households
unfit for the program (cause 5). The categories
of SSNPs are “not receiving”, “receiving
cash”, “receiving both cash & kind”,
“receiving in kind”. The base category is “not
receiving”.
All types of SSNPs are effective. But we can
see that, receiving in kind is better that
receiving cash or receiving both for reducing
food insecurity. Because the significance and
the coefficient is higher in “kind” that in
“cash”. However, receiving “both” is
insignificant in determining food poverty.
Food-based Safety Nets are
better for increasing Food
Security!
On the other hand, cash is prefered to food by
beneficiaries as cash is the most liquid. Govt
also prefers cash now as the cost of delivering
cash is cheaper than the cost of delivering food
to the beneficiaries (Ahmed, 2007). Keeping
this is mind, we can say that Cash transfer
programs are better in terms of implementation
Cash Transfer Safety Nets
are better in terms of
Implementation!

10. References
Ahmed, S. S. (2007). Social Safety Nets in Bangladesh.
BRAC. (2013). An end in sight for ultra-poverty: Scaling up BRAC’s graduation model for the
poorest.
Food Planning and Monitoring Unit. (2014). Improving the Targeting Effectiveness of Social
Safety-nets in Bangladesh. Ministry of Food, GoB.
Haider, Z. M., & Mahamud, A. (2017). Beneficiary Selection and Allowance Utilization of
Social Safety Net Programme in Bangladesh.
HIES. (2010). BBS.
Rahman, H. Z., & Choudhury, L. A. (2012). Social Safety Nets in Bangladesh. PPRC and
UNDP.