Economics of Conditional Cash Transfer Programs

Transcript

1. 1 Economics of Cash Transfer Programs: Helping Poor “on condition” Sener Salci Dept. of Economics- Queen’s University, Canada
2. Outline of Talk 1. Introduction 2. Cash Transfer Programs 3. Cost and Benefits of Conditional Cash Transfer (CCT) Programs 4. Methods for Economic Evaluation of CCT Programs 5. Economic Evaluation of Mexico’s Oportunidades CCT Program 6. Introduction of Spreadsheet Model for Cost-Effectiveness Analysis 2
3. Introduction Ø Disparities in education, health and nutrition (human capital) exist between poor and rich. Labor is the primary resource of the poor. Ø Current consumption of poor families is paid for from labour income. Ø Poor families often take their children out of schools at early ages so that they can generate additional income for their family from employment. Ø Although returns to capital are high for poor, the families cannot invest/spent “enough” on human capital for their children resulting in inter-generational transmission of poverty (Banerjee and Duflo, 2011; Behrman et al. 2004). 3
4. Incidence of Poverty in LAC, 2011 Estimates Source: World Bank, Poverty & Equity Data, 2011 4
5. Policy Objectives Policy objectives are to implement social assistance programs aiming to: •  provide immediate cash to poor, and alleviate poverty through the transfer of cash grants. •  Encourages long run development of human capital and promote social development. 5
6. Cash Transfer Programs 6 Q: Give money to poor or provide goods and services? Maria Nilza, age 36, and mother of four children, showing her "Bolsa Familia" social plan card in Serra Azul, located in north of the state of Minas Gerais in Brazil. Source: The Guardian, November 2010 Photograph: Vanderlei Almeida/AFP/Getty Images
7. Cash Transfer Programs,cont. Long-term microeconomic intervention recipes in reaching the poor • Conditional Cash Transfer (CCT) Programs o  give a monetary transfer conditional upon the receivers’ fulfilling behavioral - co-responsibilities o  Programs found mostly in Latin America and Caribbean (LAC) since the late-1990s, mostly implemented nationwide • Unconditional Cash Transfer (UCT) Programs o  provide cash to all eligible registered beneficiaries 7
8. Cash Transfer Programs, cont. The CCT Wave (WB, 2009, p. 4) 8
9. CCT Beneficiaries in LAC, by Country (2001-10)* Source: Stampini and Tornarolli (2012, p. 10) * Number of CCT beneficiaries as of 2010 represents / covers about 25% of the population in LAC. 9
10. Design of CCT Programs Increasing the country’s social welfare by offering incentives at the right level for the right households on the concept of obligation on the part of beneficiaries 1. Who will benefit from CCT Program? Who is considered to be poor? o targeting the population (beneficiaries) for intervention using socio-economic data (e.g. eligibility criteria based on household level tests such as proxy mean test, income test, or community assessment) 2. What will be the scale of the programme, and how frequent will be the cash transfers and how will beneficiaries receive the cash transfers? o determine size, duration, type and method of payment of cash transfers o finding effective way to distribute cash to poor (e.g. bank debit, cash collection points, via mobile phones) 10
11. Design of CCT Programs 3. What are the responsibilities of beneficiaries? o  set the minimum conditionality (requirements) to receive cash transfers o  linked to objective of the programs and targeted population; eg. beneficiaries’ attention to the education, health, and nutrition of their children o  planners might apply some modifications to improve effectiveness. 4. Provide cash to beneficiaries and monitor the beneficiaries o  provide cash to poor (payees are almost always the mother) if they comply with the conditions. o  monitor the program (monthly, bi-monthly…) and verify compliance, apply penalties for noncompliance with conditions Source: Adato and Hoddinott, 2010 11
12. Design of CCT Programs Practical challenges in the implementation of CCT programs: §  targeting, conditionality, size of grants §  availability of supply of services, and provision of services to meet the increased demand (i.e. children attendance to school and distance to school, opening hours of health centers and health center visits) Sources: Das et al. 2005; Bastagli, 2010; Gantner, 2007; EPRI 2011 12
13. Cash Transfer Programs, cont. Expansion of conditional cash transfers CCT Models have spread because: 1. They address the “intergenerational transmission of poverty” 2. They provide “some” income security for poor and improve income distribution 3. They encourage parents to invest in the health and education of their children (to enhance productivity and improve their employability) These are the main objectives of CCT programs and economic benefits rationale for the implementation of CCT programs. Source(s): See Fiszbein and Schady (2009); Baird, McIntosh, and Ozler (2011); Akresh, Walque, and Kazianga (2012), for review of literature educational impacts, see Baird et al. (2013). 13
14. Issue of wasting money on “temptation goods” Allocation of cash benefits in household consumption goods •  Poor women do control the cash provided and spend for higher consumption of proteins (milk, meat, eggs), cereals, increased number of meals, improved quality of food and children’s clothing. •  Based on empirical evidence, poor make the right decisions in their private spending from cash transfers, no evidence of misuse of resources such as additional consumption of “temptation goods” such as alcohol or cigarette. •  Might increase households’ investments on assets, ownership of animals, land, and other forms of small assets Source: Evans and Popova, 2014; Covarrubias et al. 2012 14
15. Costs and Benefits of CCT Programs 15
16. Estimation of Costs Program Costs •  Administrative costs associated with the delivery of cash transfers to households: conditionality, targeting costs at the household level and operation costs o  these costs depend on program size and the size of these costs also change over time (Coady, 2000, p. 99) o  delivering cash benefits to reach rural dwellers is more cost-effective than allocating goods. Private Costs •  costs that households incur in order to receive cash transfers: time and financial costs of travelling schools, health clinics and collect the cash benefits from collection centers (e.g. costs borne by woman, see Parker and Skoufias, 2000; Coady, 2000) •  initial costs of obtaining certifications required for the program: national identity card, proof of residency, and traveling costs to and from program offices. 16
17. Quantification and Estimation of Benefits What are the impacts? 1.  Educational impacts including increased school enrollment and attendance, decrease in drop-outs, improved exam test results 2.  Health impacts including the use of health services for preventative care, and improved nutritional status 3.  Multiple impacts (e.g. impacts of free school meal on kids) 4.  Additional costs from implementing the program such as increase in teacher salaries from increased enrollment, and increase in health costs from additional hospital visits 17
18. Quantification and Estimation of Impacts (Benefits) How big / small are these impacts? 1.  The data for impact study might be conducted in a randomly selected intervention and control communities. 2.  Use data from studies based on randomised controlled trials (RCT) or cluster-randomised controlled trials (C- RCT) (e.g. Gertler, 2000, Barham, 2005), or regression results from regression discontinuity (e.g. Levy and Ohls, 2007). 18
19. In the very long-run, economic benefits of lowering poverty from increased employment potentially might: •  Increase the levels of trust, civic participation and support for democracy (WB, 2012) •  Lowers the participation in riots, crime and conflict (creates positive externality for non-poor as well) (Becker, 1968; Ehrlich, 1973; Grossman, 1991) 19 Non-quantifiable Impacts
20. Heterogeneity in Benefits •  impacts might vary based on socioeconomic status of households (i.e. different income levels of treatment groups) •  impacts might vary based on location of poor households (i.e. rural and urban treatment groups) 20 Estimation of Impacts (Benefits)
21. Incremental Impacts based on Socioeconomic Status: Case of Nicaragua Source: Maluccio and Flores, 2005 21
22. CCT Programs are “more effective” in Rural Areas than Urban Areas Source: The Economist, July 2010 22
23. Are these CCT programs cost-effective and efficient? 23
24. Methods for Economic Evaluation of CCT Programs A. Focusing on the relationship between inputs, outputs and impacts, with monetary values or desired outcomes attached to them: • Cost-Benefit Analysis (e.g. education, Brent, 2013) • Cost-Effectiveness Analysis (e.g education, Dhaliwal, Duflo, Glennerster, and Tulloch, 2012, Belli, 2001) B. How well are inputs converted to outputs: • Cost-Efficiency (e.g. Coady et al. 2005; Caldés and Maluccio 2005; Caldés et al. 2006) 24
25. Cost-Benefit Analysis (CBA) Cost-Effectiveness Analysis (CEA) -Used when benefits are quantifiable and can necessarily be expressed in monetary terms but with assumptions about the monetary values of benefits. -Used when benefits are quantifiable but cannot necessarily be expressed in monetary terms and only one outcome measure can be used. . -Ability to handle programmes with multiple outcomes but it is difficult to create a single cost- benefit analysis that would be useful for a wide range of organizations. -Possible a comparison of multiple programmes evaluated in different contexts and in different years – useful for policy decision making that aimed at achieving the same objective. -Exclusion of non-monetized benefits that they may be socially desirable, especially in health and education intervention projects. -User what can be achieved for what cost and leaves it to the user to decide whether that benefit is worth the cost, with trade-offs across alternatives. 25 Methods for Economic Evaluation of CCT Programs Sources: Dhaliwal et al. 2012; McEwan, 2012; White et al. 2013
26. Cost-Effectiveness Analysis (CEA) McEwan, 2012 Assumptions 1.  improved health does not contribute or contribution is negligible to increased school attendance (separability of education and health impacts from national household surveys) 2.  incentives are given separately, solves the problem of multiple impacts of program (incentives to attend school, incentives for nutrition) Cost Effectiveness in Education: include only the costs of the education subsidy in the cost- effectiveness analysis (exclusion of quality of learning) Cost Effectiveness in Health (Cost-Utility): include only the costs of the health incentive in the cost-effectiveness analysis Separating out the CE analysis for education and health will also necessitate the distribution of the program costs and private costs for each intervention. 26
27. Mexico’s Oportunidades (PROGRESA) Program 27
28. Source: Paes-Sousa, Regalia and Stampini, 2013, p.8 (*) Data from WB, 2009, p.16. Total households represents 60 % of households in the bottom decile of per capita expenditures (WB, 2009, p.75) – successful targeting. 28 About 70% rural, 16% semi-urban and 14% urban coverage*
29. Oportunidades “keeps more children at school” Education Benefits •  Improvement in Primary School Enrollment: 1.5% girls and 1.1.% boys (Schultz, 2001), Secondary School: 9.3% girls and 5.8% boys (Skoufias, 2005) •  Increased the probability of entering secondary school by 33%. •  Nearly doubled enrollment for upper secondary education. •  Children who participated on average 5.5 years achieved an extra year of schooling (WB, 2009, p.112). •  Question with the quality of schooling. Is more schooling mean more learning? 29
30. Oportunidades “feeds poor families better and helps poor to live healthier” Health Benefits Increase in food consumption, reduction in morbidity and malnutrition • significant effect at increasing child growth child growth – up to 16 percent – and reducing child stunting for children aged 12 to 36 months and children age <=5 years old had a 12 percent lower incidence of illness than non-Oportunidades children (Gertler, 2000) • increased the number of prenatal visits in the first trimester of pregnancy by 8 percent, food expenditures in PROGRESA households were 13 percent higher than in non-PROGRESA households, with PROGRESA households consuming higher-quality foods and more calories, children aged 12–36 months were on average one centimeter taller than non-Oportunidades children (Gantner, 2007) • children age <1 year old vaccinated 1.6% more for tuberculosis and age > 1 year but <2 years have vaccinated 2.8%more for measles than non-Oportunidades children (Barham, 2005) 30
31. Oportunidades households “earn more” 31 Notes: (i)Table is cited from WB (2009, p. 113), original source of information presented in Spanish, see Parker, S., Behrman, J. (2008), (ii) Values are in US$, conversion: 11 Mexican pesos = 1 USD (iii) benefits are net benefits excluding opportunity costs and includes only education component, (iv) The analysis values the benefits from program in the form of increased future earnings that result from additional years in schooling, excluding other impacts such as household consumption and improved health and nutrition status on future earnings)
32. Cost- Efficiency in CCT Programs (CaldÈs, Coady, and Maluccio, 2004) 32
33. 33 Cost- Efficiency in CCT Program, cont. (CaldÈs, Coady, and Maluccio, 2004)
34. 34 Cost-Efficiency in CCT Program, cont. (CaldÈs, Coady, and Maluccio, 2004)
35. 35 Cost- Efficiency in CCT Program, cont. Results for Mexico Sources: CaldÈs, Coady, and Maluccio, 2004, p. 26, 30
36. CEA - Impacts of CCT on Education (Dhaliwal et al. 2012) •  Total Impact of Program = Impact (per unit) × Sample Size × Program Duration o  When there were differential impacts (heterogeneity) on different proportions of the population (see e.g. previous slide) and those impacts occurred at different times, then this calculation requires more work. •  Results are presented based on aggregate impacts (aggregating costs and impacts across all beneficiaries / total impacts) rather than the cost cost- effectiveness ratio per beneficiary (per beneficiary / the impact per beneficiary) 36
37. •  The Ingredient Method – Incremental Costs of Intervention o  What is every single ingredient necessary for this program to have the observed impact? o  How much of each ingredient is needed? o  How much does one unit of this ingredient cost? o  When is this ingredient used? 37 CEA - Impacts of CCT on Education (Dhaliwal et al. 2012)
38. 38 CEA - Impacts of CCT on Education (Dhaliwal et al. 2012)
39. •  The program was analyzed from the perspective of 2010 in USD. •  The discount rate used was 10 percent. •  Exchange rate was standard and inflation was calculated using GDP deflators. •  Results are presented based on aggregate impacts (aggregating costs and impacts across all beneficiaries / total impacts) rather than the cost cost- effectiveness ratio per beneficiary (per beneficiary / the impact per beneficiary). •  Integrate the impact assessment study from Coady and Schultz (1997) for PROGRESA Mexico (later names as Oportunidades) 39 CEA - Impacts of CCT on Education Assumptions
40. @ 90% confidence interval Impact on Outcome* Lower Bound 0.02 Point Estimate 0.03 Upper Bound 0.04 40 Source: Dhaliwal et al. 2012, p. 10 and p.19 (*) outcome represented in primary school attendance, based on study conducted by Coady and Schultz (1997), and time frame is 4 years CEA - Impacts of CCT on Education Results ADDITIONAL YEARS OF SCHOOLING PER $100 SPENT RANGES BASED ON 90% CONFIDENCE INTERVAL OF PROGRAM IMPACT
41. Case Impact on Outcome* Including costs to beneficiaries (base) 0.03 Excluding costs to beneficiaries 0.03 41 Source: Dhaliwal et al. 2012, p.29 (*) outcome represented in primary school attendance and time frame is 4 years, and values are discounted at the end of years. CEA - Impacts of CCT on Education Sensitivity Results SENSITIVITY OF COST-EFFECTIVINESS TO BENEFICIAR’S COST ADDITIONAL YEARS OF SCHOOLING PER $100 SPENT
42. 42 CEA - Impacts of CCT on Education Sensitivity Results Source: Dhaliwal et al. 2012, p. 34
43. Discount Rate (%) Impact on Outcome* 5% 0.032 10% 0.031 15% 0.031 43 Source: Dhaliwal et al. 2012, p.41 (*) outcome represented in primary school attendance and time frame is 4 years, and values are discounted at the end of years. CEA - Impacts of CCT on Education Sensitivity Results
44. 44 Exchange Rate Impact on Outcome Standard Exchange Rate 0.03 PPP Exchange Rate 0.02 Source: Dhaliwal et al. 2012, p.45 SENSITIVITY OF COST-EFFECTIVINESS TO EXCHANGE RATE ADDITIONAL YEARS OF SCHOOLING PER $100 SPENT CEA - Impacts of CCT on Education Sensitivity Results
45. How about assessment of health impacts? CEA and Health Impacts (Cost-Utility Analysis) A. Disability Adjusted Life Years (DALYs) B. Quality-Adjusted Life Years (QALYs) Q: DALY averted rather than QALY gained (?) Reading Source with Background Papers and Applications in health: WHO GUIDE TO COST-EFFECTIVENESS ANALYSIS, World Health Organisation, 2003 45
46. Study conducted: Baird, S., Ozler, B. and McIntosh, C., 2011. Cash or Condition? Evidence from a Cash Transfer Experiment. Quarterly Journal of Economics, 126, p.1709-1753 Model Presented: Dhaliwal, I., Duflo, E., Glennerster, R., Tulloch, C. December 2012. Comparative Cost-Effectiveness Analysis to Inform Policy in Developing Countries: A General Framework with Applications for Education. Abdul Latif Jameel Poverty Action Lab (J-PAL), MIT Also, read McEwan, P.J., 2012. Cost-effectiveness analysis of education and health interventions in developing countries. Journal of Development Effectiveness, 4(2), p. 189–213 Malawian Program of Conditional Cash Transfers, targeted at adolescent girls in Malawi, from available in worksheets provided in this link: http://www.povertyactionlab.org/doc/cea-data-full-workbook 46 Introduction of Spreadsheet Model for Cost- Effectiveness Analysis (CEA) for a typical Cash Transfer Program