Cash Transfers Emil Tesliuc Social Safety Nets Core Course December 8, 2010 drawn from Kathy Lindert, Xiaoqing Yu, Cornelia Tesliuc & John Blomquist 1
Outline 1. What, Why, and When to Use Cash? 2. Types of cash transfer programs 3. Design considerations 4. Summary: Pros and Cons 2
1.1 What are Cash Transfers? Cash transfer programs provide cash or nearcash assistance to the poor and certain vulnerable groups. Objectives: Increase the incomes of the poor; Help individuals and families cope with the consequences of economic and other shocks; Facilitate government reforms 3
1.3 Why Cash? 1. Cost Effective. Can be cheaper vehicle to deliver benefits than in-kind benefits (e.g., food) 2. Consumer choice. Because cash doesn t distort consumer preferences or presume to know what the individual families need. 4
1.4 When is Cash Appropriate? Situations of Chronic poverty: To reduce current (income or consumption) poverty When the demand for health, nutrition and education services is insufficient (or the returns from child labor too high) for parents to improve children s human resource development Situations of Shocks: During emergencies when there is an adequate food supply When transitory shocks (or reforms) trigger large welfare losses among the vulnerable (e.g., global crisis, energy tariff reforms) Automatic stabilizers? (recent evidence from ECA ) Circumstances for delivery of benefits: When poor or vulnerable groups can (potentially) access financial facilities (permanent or mobile) When food is too costly to transport, and is locally available. 5
1.5 When are cash transfers an inappropriate part of a safety net? Shallow financial markets (hard to move cash) When administrative targeting is not possible, hence self-targeting is the only option (no inferior cash) When supply of essential goods and services has been disrupted (wars, natural disasters) When programs aim explicitly to modify recipient consumption behavior (e.g. cash rarely given to substance abuse victims) When safety net is funded with in-kind contributions (e.g. food aid recipient countries) 6
Outline 1. What, Why, and When to Use Cash? 2. Types of cash transfer programs 3. Design considerations 4. Summary: Pros and Cons
8
2.1 Types of Cash Transfer Programs 1. Poverty-targeted / Last-Resort Programs (conditional or unconditional) 2. Categorical: a. Social pensions: noncontributory transfers to the elderly b. Disability assistance c. Family & Child Allowances 3. Near-Cash Benefits: Food stamps and other voucher programs 9
2.1 Poverty-Targeted, Last-Resort Social Assistance ( Needs-Based ) Targeted to the poor Usually through means-testing or proxy means-testing Income threshold; Household score Benefits: Financed by general revenues Benefits can either be flat, or vary depending on household situation Benefit levels generally low Common in OECD, Middle-Income countries in LAC, ECA, and increasing use in some Lower-Income Countries (Asia, Africa, LAC, ECA). Some examples: Guaranteed Minimum Income (GMI) & Minimum Subsistence Programs Conditional Cash Transfers (CCTs) 10
Guaranteed Minimum Income (GMIs) Objectives: Guarantee a minimum income for poor households below an income threshold Benefit levels: Generally equal to the difference between monthly household income and the threshold, but vary according to household size Coverage: A safety net for the poorest, aims to cover the bottom 5-10% of the population. In practice, most cover less than 5% Targeting: Usually based on income and asset testing by social workers through social welfare offices Complementary to other social protection (pensions, unemployment benefits, family allowances) EXAMPLES: Most OECD countries Most ECA countries (EU new member states, Caucuses, Balkans) Some LICs: (e.g., Kyrgyz Republic, Moldova) China (Dibao program) 11
GMI Programs: Simplified Illustration of Benefits Calculations Actual pre-transfer income Minimum Subsistence Level of Income Benefits vary by distance to minimum subsistence level 12
Proxy Means Testing Eligibility criteria Welfare aggregate Welfare aggregate based on consumption/expenditures Adjusted for household composition Proxy Means Testing approach Need to define a metric to rank households by their welfare Welfare aggregate (score) is based on a set of measurable characteristics reported by households Score is a linear combination of these characteristics complemented by subjective assessment of the household s living conditions Households are ranked based on total score 13
Proxy-means Testing: Simple example ID CAR TV AC Radio Rooms Score 1001 0 1 1 0 4 1002 1 0 1 1 3 1003 1 1 0 1 1 1004 0 0 1 1 2 1.6 2.11 1.51 0.71 Score Formula S i = 1*Car+0.3*TV+0.5*AC+0.01*Radio+0.2*Rooms Higher score corresponds to better standards of living S 1002 1003 1004 1001 = 1*1+0.3*0+0.5*1+0.01*1+0.2*3 1*1+0.3*1+0.5*0+0.01*1+0.2*1 1*0+0.3*0+0.5*1+0.01*1+0.2*2 1*0+0.3*1+0.5*1+0.01*0+0.2*4 Cutoff score 1.55
Scoring methodology Weights for the score calculation are derived based on the regression of the total household expenditure on the set of household characteristics. The expenditure regression is estimated on a sample of households from HBS conducted by Statistics Georgia. The sample includes a detailed information on consumption expenditure for about 3400 households. PE = i β jdij + β kdik + βldil + β mdim + demo assets dwell locat head β d n in +... S i = ( PE i min( PE ))*100 ( max( PE ) min( PE )) i i i An alternative score is derived based on the subjective assessment of the interviewer 15
Set of characteristic to measure welfare Information on household members: Age Gender Education level Health condition, etc. Agricultural activity: Agricultural land Domestic animals etc Ownership of durables: Car Refrigerator TV Washer etc. Geographic dummies: Region Distance to the district center Incomes: Wages (???) Pension Social assistance and transfers Annual expenditures on durables: Car Furniture Repairs of the residence Living conditions: Number of rooms Type of dwelling Sources of water, electricity, etc. Assessment by the interviewer: Condition of furniture Condition of the residence General economic wellbeing 16
An example of log consumption regression Coef sd Coef Sd lhhsize -0.562*** 0.036 Cooking Oven 0.054*** 0.020 lhhsize2-0.014 0.011 Cooking Stove -0.001 0.019 Share of children 0-6 -0.534*** 0.037 Microwave Oven 0.067*** 0.013 Share of children 7-14 -0.520*** 0.035 Dishwasher -0.007 0.035 Share of children 15-18 -0.121*** 0.046 Vacuum Cleaner 0.060*** 0.009 Share of adult males (19-60) (dropped) TV Set 0.112*** 0.029 Share of adult females (19-60) -0.166*** 0.035 Satellite Receiver 0.042*** 0.010 Share of elderly -0.144*** 0.038 Video 0.045*** 0.012 Share of job holders -0.010 0.030 Video Camera 0.067*** 0.020 Share of employed in public sectror 0.038*** 0.012 Computer 0.063*** 0.010 Recieves any subsidy -0.106*** 0.017 Telephone 0.075*** 0.009 Characteristics of the head Mobile Phone 0.128*** 0.011 Female headed hh 0.037** 0.015 Air Conditioner 0.099*** 0.018 Age of the head 0.001* 0.001 Solar Heater 0.008 0.012 Education of the head Private Car 0.210*** 0.009 Illiterate (dropped) Radio 0.005 0.008 Read/Write 0.048*** 0.017 Sewing Machine 0.004 0.011 Primary 0.056*** 0.016 Characteristics of dwelling Elementary 0.043*** 0.016 Number of rooms 0.050*** 0.005 Basic 0.021 0.024 Villa (dropped) Secondary 0.083*** 0.018 House 0.008 0.041 Intermidiate 0.088*** 0.020 Flat 0.029 0.041 University 0.156*** 0.021 Dwelling area m2 0.001*** 0.000 Occupation of the head Type of ownership hocc_1-0.031 0.059 Own 0.040*** 0.015 hocc_2 0.101* 0.059 Rented 0.018 0.016 hocc_3 0.007 0.057 For work -0.096* 0.049 hocc_4 0.026 0.058 Free (dropped) hocc_5-0.013 0.058 Source of heating hocc_6 0.038 0.056 Kerosine 0.056 0.054 hocc_7 (dropped) Gas 0.102* 0.055 hocc_8-0.005 0.056 Central 0.249*** 0.057 hocc_9 0.006 0.056 Electricity 0.169** 0.070 hocc_10-0.033 0.055 Wood or coal 0.039 0.058 hind_1 0.030 0.040 Other (dropped) hind_2-0.010 0.060 No heating 0.005 0.065 hind_3 0.033 0.058 Source of drinking water hind_4 0.068 0.067 Piped water -0.237*** 0.068 hind_5 0.024 0.058 Well -0.217*** 0.070 hind_6-0.010 0.056 Water tank -0.180** 0.071 hind_7 0.051 0.057 Mineral/filtered water -0.185*** 0.069 hind_8 0.027 0.057 Spring water (dropped) hind_9-0.011 0.057 Public sewage 0.017 0.011 hind_10-0.000 0.060 Construction material Household assets Clean Stone 0.209*** 0.065 WASHING MACHINE 0.095*** 0.024 Stone & concr 0.136** 0.064 REFRIGERATOR 0.070*** 0.024 Concreate 0.113* 0.064 FREEZER 0.044*** 0.014 Cement block 0.096 0.063 17
Measured and Predicted Welfare Ranking 3 IV I Predicted expenditure 2 1 Poor 0 III II Beneficiaries -1-1 0 1 2 3 Actual expenditure 18
Georgia Targeted Social Assistance Program Procedures of the registration in the data base Central Office Data-processing (persons Identification) VPN Regional Offices (10) Information is entered into database Computer District Offices (70) Social Agent a person equipped with the special authority by the Agency. SA identifies families that express desire to be registered in the database, interviews family members, estimates their socio-economic condition and collects all other necessary information. refusal Motivated Households Households declarations Local Governments (1200) Applications 19
Georgia Targeted Social Assistance Program Procedures of the registration in the data base Central Office Data-processing (Score Calculation) VPN Regional Offices (10) Information is entered into data base Computer District Offices (70) Households Declarations Households Ranking score that corresponds to the level of wellbeing of a family 20
Passport number Score ID s of household members
ECA Examples: Targeting Accuracy of LRSA (GMI) Programs 70 ECA: Targeting Accuracy of Social Assistance Programs 70 ECA: Targeting Accuracy of Social Assistance Programs 60 50 40 30 AVERAGE LRSA (Nobs=19) AVERAGE CA MT (Nobs=8) 60 50 40 30 AVERAGE SCHOLARS HIP (Nobs=15) AVERAGE UTILITY (Nobs=10) 20 10 0 Q1 Q2 Q3 Q4 Q5 AVERAGE CA NON-MT (Nobs=14) 20 10 0 Q1 Q2 Q3 Q4 Q5 AVERAGE WAR VETERANS (Nobs=4) Y axis: share of program funds going to each (per capita consumption) quintile; average value across each program type Source: World Bank estimates using latest household survey data: Sundaram et. al. (forthcoming 2010) LRSA = last resort social assistance (GMI and minimum subsistence programs) CA = child allowances (Means Tested and Non-Means-Tested) 22
Conditional Cash Transfers Dual objectives: Pay a cash benefits to alleviate poverty and inequality in the short run That is conditional on household investments in education (e.g., school attendance) and health (e.g., growth monitoring) to break the inter-generational cycle of poverty in the long-run Targeting: usually means-tested or proxy meanstested, often in combination with geographic targeting Benefit levels and coverage vary EXAMPLES: Most countries in LAC Several countries in East Asia & South Asia Several countries in Africa Several countries in ECA Several OECD countries Covered in a separate presentation 23
Jamaica PATH CCT program 24
CCTs in LAC: Targeting Accuracy Conditional Cash Transfers Compared with Other Social Assistance Programs: Absolute Incidence % of Benefits received by each quintile 50% 40% 30% 20% 10% 0% Q1 Q2 Q3 Q4 Q5 CCTs Other Cash Scholarships School Feeding Other Feeding Source: Lindert, Skoufias and Shapiro (August 2006) 25
2a. Categorical Benefits: Social Pensions Objectives: To ensure basic old-age security for those not covered by the contributory pension system Eligibility: Varies. Universal for all elderly or targeted to poor elderly Financing: mostly financed by general tax revenues Benefit level & Incentive Compatibility: Setting appropriate benefit level is important: If high relative to minimum contributory pensions undermines incentives to contribute (Uruguay case) If too low, won t contribute to poverty alleviation, admin costs become large share of total (Argentina, Turkey) EXAMPLES: OECD Countries (Australia, New Zealand, Canada) Africa (South Africa,Namibia, Mauritius, Botswana) South Asia (India, Bangladesh, Nepal) LAC: (Argentina, Bolivia, Brazil Chile, Costa Rica, Uruguay) ECA: (almost all) 26
2b: Categorical Benefits Disability Assistance Objectives: To provide cash assistance to the disabled as a vulnerable group for those not covered by disability insurance Eligibility: Varies. Universal for all disabled or targeted to poor disabled. This raises two levels for screening: Disability certification. Classification of disability, institutional set-up / roles, time limits for recertification, etc. Means-testing / Screening based on (poverty) need (Also sometimes focused on disabled children categorically) Financing: mostly financed by general tax revenues Benefit level & Incentive Compatibility: Setting appropriate benefit level is important If high relative to disability insurance undermines incentives to contribute If high relative to other social assistance benefits, incentives to get certified for disability benefits EXAMPLES: Most OECD Countries ECA: (almost all countries) LAC: (e.g., Brazil Chile,Uruguay, Barbados, Bermuda, Trinidad and Tobago) Hong Kong (China) Africa: Liberia, South Africa 27
Operational Definitions of Disability in Medical Assessments Definition Advantages Disadvantages Inability to work Conceptually appropriate Consider full set of medical and other circumstances Sensitive to context (accessibility of transportation, buildings, types of jobs, etc.) Discretion implies variability among assessors and/or expensive systems to minimize discretion through use of multiple assessors or review panels Moral hazard Based on official list of impairments or diagnoses Simpler to guarantee equal treatment of people with same conditions Easier to verify Does not recognize differences in severity Does not recognize interactions among multiple conditions Lists can be politically difficult to agree on 28
Interplay between disability benefits and targeted social assistance Albania Example 2.5% Coverage: increasing for DB, decreasing for SA (NE) Benefit levels: DB 3 times > SA (NE) Expenditures on DB crowding out SA Moral hazard? Albania Social Assistance Spending (%GDP) 2.0% 1.5% 1.0% Disability assistance benefits 0.5% 0.0% Poverty-targeted social assistance (Ndimhe Ekonomike) 2000 2001 2002 2003 2004 2005 2006 2007 2008 3.1 Ndimhe Ekonomike 3.2 Disability Allowance 3.3 Care Allowances 3.4 Other 29
2c. Categorical Benefits Family & Child Allowances Multiple Objectives: Support families to reduce child-raising costs Child protection objectives (reduce abandonment potential) Support to orphans Pro-fertility policies (e.g., Birth allowances in ECA) Eligibility: Varies significantly. Means-tested child benefit to the poor Categorical benefit to all children under certain age (wide age range) Benefit Levels: often flat rate EXAMPLES: Most OECD countries ECA: virtually all Africa: Malawi, South Africa LAC: Argentina, Chile Common in OECD and East European and the Former Soviet Union 30
3. Near Cash: Food Stamps and Vouchers What are they? Food stamps or coupons are cash-like instruments that can be used to purchase food at authorized retail locations Vouchers are cash-like instruments that can be used to purchase specific services Alternative Currency. The value of the stamp or voucher is backed by government commitment to pay (reimburse retailers or service providers) Benefit Levels: In theory, some link to minimum consumption basket or minimum cost of service. Some food stamps programs restrict households to only by specific foods. In practice, food stamps benefits often only represent a small share of the cost of the food basket Eligibility and Administrative Requirements: Same as for cash but With added requirement of printing and distributing alternative currency Examples: Food Stamps: United States Honduras Sri Lanka Mexico Colombia Jamaica (until 2002) Vouchers: Numerous examples, see Georgia example next slide 31
Georgia Example: Using Vouchers to Complement Targeted Cash Assistance Cash benefits (TSA, pensions, disability, IDP benefits, etc.) Outside SSA: Education vouchers using PMT registry Social Services Vouchers + Child Protectio n PMT registry, MIS, One- Stop Shop Centers Electrici ty Voucher s Health Insuran ce Voucher s and other social insurance Outside SSA: Ministry of justice using PMT registry for provision of legal services
Outline 1. What, Why, and When to Use Cash? 2. Types of cash transfer programs 3. Design considerations 4. Summary: Pros and Cons
3. Design Considerations 1. Set objectives and keep expectations reasonable 2. Establish criteria for beneficiary selection 3. Ensure proper financing and fiscal sustainability 4. Set benefits level for adequacy and efficiency 5. Develop governance structure to ensure transparency and inclusiveness 6. Build administrative capacity and institutions to better manage information, disburse funds, handle grievance, communicate, and measure results 7. Sustain political support 34
3.1 Program objectives and expectations Cash transfers don t solve poverty or eliminate risk Some benefit leakage will occur Not all needy can be covered by any single program Coverage determined by objectives, fiscal allocation, administrative capacity Number of programs typically expands when social protection systems are developed Active vs. Passive Assistance: Going Beyond Cash Links to human capital (e.g., CCTs) Links to employment (e.g., activation services, work requirements) Links to other social services 35
3.2 Targeting: Beneficiary Selection Targeting mechanisms: categorical and geographically targeting; means-test (income, asset); proxy means; community targeting; self-targeting; [To be covered in detail in another session] There is no perfect targeting: Target groups typically receive 30-75% of direct benefits in cash transfer programs around the world; Cash transfers are better targeted than in-kind 36
3.3 Fiscal sustainability and Financing arrangements Long-term sustainability Adequate budget available for annual total benefit outlays and administrative costs Threat to sustainability: insufficient budget to meet program objectives Arrears Discretionary, instead of rule-based allocation of benefits Partial payments Understaffing, which leads to greater leakage Ad hoc adjustment to inflation, erosion of purchasing power Financing arrangement in a decentralized setting 37
3.4 Setting Benefit levels Type of program Guaranteed minimum income Last resort programs Food stamps Family allowances Heating allowances Social pension Benefit level depends on: Eligibility threshold income of beneficiary household Poverty gap Food poverty gap The cost of raising a child Seasonal increase in the heating cost during cold season Poverty line Minimum contributory pension 38 Low income countries Cost of an adequate food basket The food poverty line. 38
3.4 Benefit levels in practice Limited to 4 types of programs, 2 regions Family allowances (n = 15) Last-resort programs (n = 20) Conditional cash programs (n = 6) Social pensions (n = 14) 39 Generosity of Social Safety Net Programs from ECA and LAC Regions 0 20 40 60 Benefit in % of the consumption of recipient household Source: Tesliuc (2008, forthcoming) and Shady (2007) 39
3.5 Work Disincentives? Generosity Work disincentives depend on three key factors Marginal tax rate on earnings Beneficiary household with supply of labor 40
3.6 Making Cash Transfers Inclusive Particular attention to women Female headed households may face greater risks Women may make better use of transfers Poverty payoffs to better women s nutrition and empowerment shown to reduce half or more of infant malnutrition. Gender reviews to identify risks and legislative inequity Old-age pensions and family allowances tend to selftarget women Legislative reform to provide equal access to welfare Voice of marginalized groups Identification and decision making process, grievance 41
3.7. Administrative Capacity to Deliver Challenges Political interference in staffing & investment Fragmented policy making and programs fragmented amongst many agencies Delays in processing claims Poor record-keeping Failure to explain the schemes to members and public Poor terms of service Excessively complex procedures and regulations Neglect of compliance, M&E and policy research functions Reform Options Establish professional criteria and staff certification boards Separate policy and operational aspects--- contract out the later to social partners Consolidate and harmonize programs Central gov to prepare guidelines and local govs to prepare implementation rules Earmark funds for M&E, audit, and policy research Automation after basic 42 systems in place 42
3.8 Sustaining Political Support New programs tend to start during a crisis but political interest quickly fades Depoliticize payments and eligibility processes (institutional responsibility, pre-election quarantines) Programs with good O&C and M&E may gain and maintain support Build consensus and support with analysis, outreach, transparency, consultations Universal or narrow targeting Minimize incentives effects (perceptions of lazy ) Proactive links to human capital, activation services can help improve political sustainability (get beyond cash handout) 43
Outline 1. What, Why, and When to Use Cash? 2. Types of cash transfer programs 3. Design considerations 4. Summary: Pros and Cons
4. Summary Cash transfers: Pros and Cons Pros Least-cost solution to delivery benefits and reduce current poverty. Consumers decide how to use the cash to meet their needs Do not distort prices Potential role as automatic stabilizers (e.g., guaranteed minimum income programs) Benefits can be differentiated by level of need, household size, composition Cons Subject to price inflation Use might not be optimal (unintended use of transfers; payments to women as family beneficiary helps) Potential work disincentives (design features can minimize these) Haven t always responded flexibly in crisis (limited budgets, specific design features evidence from ECA) Implementation can be information intensive Low administrative costs Admin capacity requirements significant Can be well targeted 45 Political economy of narrow targeting? 45 (Context-specific)
Thank You Open for discussion! 46
Example: China Minimum Living Allowance program ( Dibao ) Target group Targeting method Eligibility criteria Eligibility rules Financing Implementation Benefit level Benefit formula Households whose per capita income is below the dibao line Income and asset test; community Income, assets, general living condition Nationally determined framework, locally set standard Largely provincial + municipal; Central govt support poor regions only Decentralized units of the line ministry, village/community committee Varies depending on the applicant s earned income Dibao Threshold Total income of the family 47
Example: Mongolia Child Money Program Promised in 2003-04 during election to establish universal child benefit Program established in 2005 introduced a targeted program using proxy means-testing to identify the poor. Requirements include: (a) family income below minimum substantive level; (b) have at least one child aged 18 or younger; (c) enroll an school-age children in school; (d) ensure all children have received their mandatory immunizations; (e) ensure that their children are not engaged in illegal child labor. By December 2005, 609,000 children from 303,000 households (24% of the population) were covered. Very limited capacity to implement. Amendments to the Social Welfare Law of July 2006 universalized the CMP to all families with children and the benefit was increased in January 2007 from 36,000 MNT to 136,000 MNT per child a year. As energy prices fall and affected by the global economic crisis, the need to target is recognized and being discussed. 48
Example: Sri Lanka Food Stamp Program Established in 1979, the program subsidized the consumption of basic goods for the poorest households. Eligibility was determined by a means-test based on self-reported HH income, with a marginal adjustment made to hh size. Can use the stamp to buy basic food (rice, wheat flour, sugar, milk products, etc.) Expenditure reached 1.3% of GDP in the 1980s. Replaced by cash transfers and subsidies in 1989 and 1995 49
Examples - Social Pensions Universal social pensions Bolivia provides a universal social pensions (fixed cash transfer) to all citizens over 65. Multiple objectives: to return the equity in the privatized state enterprises to the people, to cover the large majority of elderly not covered by the pension program, and to help reduce poverty. The program costs about 1% of GDP and covers 0.7% of the population. Targeted social pensions South Africa old-age pension covers all women above 60 and men above 65, subject to a means-test. The program covered about 4.2% of the population and is funded through general taxes. The total program expenditure is about 1.4% of GDP in 2000. Impact: Incidence of poverty; Health status of children and older people; Enrollment rates of school age children 50