Transcript Program Implementation Matters for Targeting Performance

Social Safety Nets and
Macroeconomic Crises:
Are ECA Countries Ready?
Emil D. Tesliuc
HDNSP
1
Workshop on Macro Risks
and Micro Responses
February 15, 2008
Outline
1
2
2
SSNs as Shock Absorbers: What does it take?
Are Existing Social Safety Net Programs
Ready for the Next Crisis?
1
SSNs as Shock Absorbers
What does it take?

SSN objectives during crises:
–
–

Characteristics of well-performing SSN programs
–
–
–
–
–

3
Protect (old and new) poor against irreversible welfare losses (physical,
human capital)
Help gain accepted to do needed reforms
Automatic stabilizers
… or at least programs that can be scaled-up timely, and scaled-down once
the crisis is over
Good at identifying the intended target groups (cost effective)
That injects money into the pockets of those with high marginal propensity to
consume (fiscal stimulus)
Self-targeted (public works)
Complementing other protective measures:
–
health, education, pensions, unemployment benefits, basic infrastructure
1
Challenges of SSN programs during
crises



When good SSN programs exist, need $$$ and
staffing for scaling-up
When well-performing programs not available,
response will be inefficient: costlier, late
Targeting difficult
–
–
–

4
Household circumstances change often
Informal economy goes up
Eligibility - recertification, more often
Which one to help: Old poor or new poor?
Chronic vs. transitory poverty?
2
Overview of Social Safety Net Programs
in the ECA region

We look at:
–
–
–
–

The SSNs in ECA fall into four groups
–
5
Spending levels
Program mix
Coverage of the poorest
Targeting accuracy
based on their preparedness to avert the effect of
the crises, and the type of SSNs
2
ECA is second only to OECD in terms
of SP spending
Social Protection Spending, Region Averages
Social Safety Nets
18
Social Insurance
% of GDP
16
14
12
10
8
6
4
2
0
13.2
8.3
3.8
2.5
OECD (N=23)
6
3.0
1.7
3.6
Eastern Europe M iddle East and
and Central Asia Northern Africa
(N=25)
(N=10)
1.3
Latin America
and Carribean
(N=25)
2.9
1.0
1.4
0.9
East Asia Pacific South Asia (N=5)
(N=4)
Notes: Data on 69 countries taken from WB Public expenditure reviews or other similar work. For OECD, data from the OECD
Social Expenditure database (OECD, 2004). Different years, most data from 2000 to 2003.
ECA Region: Social Safety Nets' Expenditures
Bosnia and Herzegovina
2
Slovak Republic
Ukraine
Belarus
Kosovo
Czech Republic
Kazakhstan
… and
SSN
spending
levels
seem
adequate,
except for
few
outliers
7
Turkey
Armenia
Montenegro
Uzbekistan
Croatia
Russian Federation
Moldova
Azerbaijan
Georgia
Serbia
Macedonia, FYR
Latvia
Bulgaria
Albania
Romania
Poland
Kyrgyz Republic
Tajikistan
0.0
0.5
1.0
1.5
2.0
2.5
% of GDP
3.0
3.5
4.0
4.5
2
Program mix

Similar with OECD in terms of programs mix:
–

SSN programs more narrowly targeted than in EU
–

less categorical targeting, in favor or income and asset
testing, hybrid means-testing, proxy-means testing
Regional specificity:
–
–
8
family allowances; social pensions; heating/housing
allowances; last resort programs
–
Legacy programs: privileges still plague some countries;
subsidies
CCTs are rare. Found only in Turkey and Macedonia
Workfare, public works programs, also rare
2
Few of the poorest are left out, except for countries
in Central Asia, Caucasus and former Yugoslavia
Hungary
Estonia
Poland
Lithuania
Russia
Coverage of
the poorest
quintile by
Social
Protection
Programs
Bulgaria
Romania
Serbia-Montenegro
Macedonia
Kazakhstan
Bosnia Hertzegovina
Belarus
Albania
Countries
ranked in
decreasing
order by per
capita GDP
in 2000 PPP
Armenia
Azerbaijan
Georgia
Moldova
Kyrgyzstan
Uzbekistan
Tajikistan
0%
9
20%
40%
60%
80%
100%
9
Only Social Assistance
Both SI and SA
Only Social Insurance
Not covered
2
Share of Individuals in the Poorest Quintile
Not Receiving Any SP Transfer
Tajikistan
50
40
Uzbekistan
Bosnia Hertzegovina
Moldova
Kyrgyzstan
Georgia
Armenia
%
Albania
Kazakhstan
Macedonia
30
Lithuania
Serbia-Montenegro
20
Belarus
Russia
Bulgaria
Azerbaijan
Poland
Estonia
10
Hungary
Romania
0
3
6
9
GNP per capita, in '000s $, 2000 PPP
Data source: World Bank, 2008
10
12
15
2
Errors of
Inclusion

Approaching OECD standards in best of cases;

But remains mediocre in many others
Share of Beneficiaries from the Poorest Quintile
Non-contributory cash transfers in US, LAC and ECA
100
Programs
using:
60
%
Income
testing
80
40
11
US
LAC
ECA
Tajikistan
Azerbaijan
Macedonia
Uzbekistan
Bosnia
Georgia
Russia
Armenia
Serbia
Belarus
Albania
Kyrgyzstan
Moldova
Poland
Estonia
Hungary
Lithuania
Bulgaria
Romania
Argentina
Mexico
Jamaica
Chile
TANF
0
Brazil
20
Food
Proxymeans
testing
2
Two cost-effective SSNs models
have emerged

Low(er) income countries (Armenia, Albania, Kirghiz Republic)
–
–
–
–

Medium Income Countries (Bulgaria, Lithuania, Romania)
–
–
–
–
12
Targeted programs = the SSN, typically one large program
Greater coverage (10-20% of the population), budget of 1-2% of GDP
Informal labor market; passive labor market policies small or no role
Passive systems (little emphasis on exit policies)
–
Targeted programs = residual programs, complementary to an extensive
social protection system
Children, elderly, heating …
… plus a last-resort program with small coverage (5-10% of the
population). With budgets between ¼ and 1% of GDP; cost-effective.
Emphasis on exit policies
Strong links with employment/labor offices
2
Regional solutions adapted to
regional circumstances


Entitlement programs of moderate fiscal costs
Accurate targeting systems, similar with OECD , but
well-adapted to country circumstances:
–
–
–


13
Income / asset testing combined with formal economy test (all
6 countries) and work-test (Albania, Bulgaria, Romania)
Hybrid means-test: Imputed incomes based on occupational
or asset profile (Bulgaria, Kirghizstan, Lithuania, Romania)
Proxy-means test (Armenia, Georgia, Turkey)
Ready for the next crisis: Open eligibility, frequent
recertification
Programs that operate with modest but sufficient
administrative costs
2
Cost-effective solutions to identify
poor beneficiaries
Share of beneficiaries from the poorest quintile
Targeted vs un-targeted programs
10%
20%
30%
40%
16%
50%
60%
14%
12%
Low income countries (LICs)
10%
6%
30%
4%
37%
7.8%
10.5%
9.9%
6.5%
58%
60%
Guaranteed Minimum Income, Romania
64%
27%
6.0%
2.2%
Mexico
Colombia
Brazil
Armenia
Romania
Lithuania
Bulgaria
Means Tested Programs
Social Benefit, Lithuania
7.1%
0%
27%
Guaranteed Minimum Income, Bulgaria
14
10.0%
2%
Medium income countries (MICs)
Average non-pension program, MIC
13.4%
11.0%
Albania
Average non-pension program, LIC
8%
United
States
Unified Monthly Benefit, Kirghizstan
%
41%
Ndihme Ekonomika, Albania
Family Poverty Benefit, Armenia
70%
United
Kingdom
0%
Share of administrative costs in total cost of the program
Proxy Means Tested programs
Substantial know-how on how
to target narrowly, adapted to
the circumstances of ECA
region, transferable to other
countries and regions
2
Many ECA countries have the
necessary SSN in place

Automatic stabilizers in place:
–
–



15
Unemployment programs, means-tested / last resort
programs
Challenge: given $$$ and increased in staff
Competent /strong administrative systems
Have developed effective, efficient and flexible/
agile mechanisms to identify the poor
Proven: SSNs already helped cushion the
effect of previous crises:
–
Example from Armenia
16
Country
Albania
Armenia
Azerbaijan
Belarus
Bosnia-Herzegovina
Bulgaria
Czech Republic
FYR Macedonia
Estonia
Georgia
Hungary
Kazakhastan
Kyrgyzstan
Latvia
Lithuania
Moldova
Poland
Romania
Russia
Serbia-Montenegro
Slovakia
Slovenia
Tajikistan
Uzbekistan
Unemployment
Last resort /
benefits/
Public works / Guaranteed
Has at least one
assistance
Workfare
Min. Income
program?
1
0
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
1
1
1
1
?
1
1
1
0
?
1
1
0
1
1
1
0
0
1
1
0
?
1
0
0
1
1
1
0
1
1
1
?
1
1
1
0
1
1
1
0
0
1
1
1
0
1
1
0
1
1
1
0
1
1
1
0
1
1
1
?
1
1
1
1
1
1
1
?
0
1
1
0
0
1
2
Armenia:
Poverty Family Benefit (PFB) Program



$873 (2003), low income, high poverty, food insecurity.
–
Triple shocks: independence from Soviet Union disrupted trade patterns; severe
earthquake wiped out significant industrial capacity and housing in 1998; conflict with
Azerbaijan and trade blockade
•
Pre-reform:
–
Inherited a generous and regressive cash benefit system (26 small cash and in-kind
programs), and extensive quasi-fiscal social programs (electricity subsidy)
Reform (1998/1999):
–
Consolidation of all social assistance program into one, the PFB, targeted to the poor using
a proxy-means test. Electricity subsidy discontinued
Fraction of Social Assistance Budget Captured by Each Quintile
Institutional consolidation of the PFB (1999 - present)
Note: Consumption ranks in 2003 and 1998/9 are not-comparable
Results to date:
50
–
Improvements in targeting accuracy , coverage
45
40
–
(27% of the households covered vs 15%)
35
30
–
and generosity (higher benefits)
25
–
Further improvements in targeting and generosity
43
20
32
–
Low share of admin costs (2% of program budget)
15
10
–
No labor disincentives
16
5
17
% benefits received by quintile

Setting: Landlocked country in the Caucasus, population 3 million people, per capita GDP US
17
1998
Poorest Quintile
1999
Q2
Q3
2003
Q4
Richest Quintile
Year
… but not all ECA countries have
found a cost-effective SSN model

Some countries are marred by:
–
–
–
–
–
–

18
Impropriate programs (legacy …)
Unclear objectives
Poor design
Understaffing
Underfunding
Inexistent or poor monitoring, oversight, evaluation
The remaining challenge is to get these countries
prepared
2
Add’l Slides: Features that
contribute to accurate targeting



19
Lowering barriers to
participation
Lowering inclusion errors
Improving program
administration
Key phases influencing targeting
accuracy for transfer programs
OUTREACH
ELIGIBILITY
DETERMINATION
APPEALS
BENEFICIARIES
20
EXIT
RECERTIFICATION
• M&E
• Audit
•Quality Control
PAYMENTS
What implementation aspects are shared by
programs / countries with well-run SSNs?
Lowering barriers to
participation
Lowering errors of
inclusion
Improving program
administration
Effective
Cross-check
Train
dissemination of
information about the program
Location of point-of-service
close to the beneficiaries,
minimize visits and waiting
Minimize documentation
required, free-of-charge
provision of documents
attesting eligibility
Introduction of one-stop or
one-window system
Single application for multiple
benefits
Use of multiple targeting
methods
21
the information
provided by applicants against
other public databases
Perform home-visits to
assess the means of the
households
Frequent re-certification
Introduction of a unique
identification number for
beneficiaries
Reduce data-entry mistakes
in computerized databases
Effective audit and control
processes
Use of multiple targeting
methods
staff
Provide adequate
documentation for staff
Strong monitoring, especially
performance monitoring
systems
Collaboration across
agencies
Strong monitoring, especially
performance monitoring
systems
Use the same targeting
instrument across many
poverty-focused programs