SPA The Social Protection Atlas Benchmarking social protection

SPA The Social Protection Atlas Benchmarking social protection systems with ADe. PT HDNSP – SSN team, May 2010

SPA Motivation and background 2 Need to build empirical evidence for the SN case Multiple efforts; databases of programs For protection and promotion/WB web: www. worldbank. org/sp Program level (WB supported); Database of public spending on Social policy (70+ countries) UNDP Poverty Center: www. undp-povertycentre. org Chronic poverty research center (ver. 4) www. chronicpoverty. org/publications/details/social-assistance-in-developing-countries-database-version -4 -0 These sources provide program-level information Programs are presented regardless of size

SPA Objective 3 Provide nation-wide assessment focusing on he social protection system and its elements the incidence of spending (coverage, amounts) and impact using official representative household survey data, processed into a comparable and harmonized datasets. Benchmark safety protection programs across countries and across time. Provide open and easy to access data for policymakers, civil society, World Bank staff and other stakeholders. Initiate the dialogue with statistical offices for improving household data collection on social protection programs.

Is SPA unique? 4 There is no other source of similar indicators for the developing world With the same coverage Archived and accessible Focused on cutting edge techniques to assess program performance (ADe. PT tool) For richer countries Luxembourg Income Study projects offers similar type of capabilities LIS hosts data from 200+ surveys from 36 countries, for some covering 1970 -2008 (www. lisproject. org/techdoc. htm) Poverty lines use relative poverty to allow comparisons (% of median) Detailed list of social transfers and taxes Facility to conduct analysis, produce tables or download key figures

What is ADe. PT? 5 Software platform for Automated Economic Analysis Free, stand-alone program available to everybody Accepts individual- and household-level data in Stata and SPSS format. Uses Stata numerical engine for computations. Minimal data preparation required from the users Extensive diagnostics of possible problems with the data ADe. PT is a tool for simulations and sensitivity analysis Intuitive user-friendly interface Tested on 100’s of datasets from more than 50 countries ADe. PT Poverty: Released – June 2007 ADe. PT Labor: Released – November 2007 ADe. PT Gender: Released – November 2008 ADe. PT Social Protection: Website: www. worldbank. org/adept Released – June 2009 ADe. PT Education: Released – September 2009 ADe. PT Health: Released – December 2009 ADe. PT Inequality: Planned Release – Spring 2010 ADe. PT Targeting: Planned Release – Summer 2010 ADe. PT MAPS: Released – November 2008 …. .

ADe. PT: From data to report User micro-level data: DHS, LSMS, LFS, … ADe. PT Inside ADe. PT: User interface Computational kernel (Stata) Print-ready output

SPA in LAC 7 Most recent household survey data available at CEDLAS (Centro de Estudios Distributivos Laborales y Sociales) with information on household income and social protection programs � � � � Argentina 2006 Bolivia 2006 Brazil 2006 Chile 2006 Colombia 2003 Costa Rica 2008 Rep Dominicana 2007 Ecuador 2008 � � � � El Salvador 2005 (2007) Guatemala 2006 Honduras 2007 Jamaica 2006 Mexico 2008 Nicaragua 2005 Panama 2008 Paraguay 2007 � Peru 2008 Suriname 1999 Uruguay 2008 Venezuela 2006 � More to come � � �

SPA in ECA 8 Most recent household survey data available at ECA targeting data base and ECAPOV) with information on household income and social protection programs � Armenia 2008 � Kosovo 2006 � Ukraine 2006 � Azerbaijan 2007 � Kyrgyzstan 2006 � More to come � Belarus 2008 � Latvia 2008 � Bosnia 2007 � Lithuania 2004 � Bulgaria 2007 � Macedonia 2005 � Georgia 2007 � Poland 2005 � Hungary 2004 � Romania 2008 � Kasakhstan 2007 � Serbia 2007

9 SPA in AFR, SAS, MENA and EAP Most recent household survey data available collected from AFR, MENA, EAP and SAS with information on household income and social protection programs � � AFR � MENA � EAP � Kenya 2005 � Egypt 2008 � Vietnam 2002 � Mauritius 2005 � Jordan 2003 � More to come � Morocco 2001 � Yemen 2005 � West Bank 2007 SAS � Bangladesh 2002 � Pakistan 2005 � More to come

SPA Impact of SP on inequality 10

11 SPA Impact on inequality (SI vs SA)

SPA Cross country analysis 12

SPA standard output (ADe. PT SP) 13 Sample and population size of survey data used Average transfer values by program (and all combined) Coverage (% of population in beneficiary households) Distribution of benefits (% accruing to each decile) Targeting accuracy (exclusion and inclusion errors) Relative incidence Generosity (% of income covered by program) Impacts on poverty and inequality (by program and all combined) Undercoverage, leakage and targeting differential Overlap across programs Coady-Grosh-Hoddinott indicator Cost-benefit ratios

SPA: Discovering gaps 14

SPA: Sorting out definitions 15

SPA benchmarking SI/SA 16

Uses of SPA 17 Social protection system of a country/region at a glance In terms of overall coverage of the population/poor By its elements/ overlaps/ gaps Benchmarking whole systems and programs In terms of coverage /incidence In terms of impact on poverty/inequality In terms of cost –benefit ratios Identifying data gaps In terms of programs/countries

SPA Next steps 18 Process of review with regional teams: expanding coverage of countries and time periods!!! Dialogue / partnership with LIS to extend the coverage to high-income countries Cross-validation with administrative data Add simulations capability …. . to mimic program rules (selective based on country data) (? ) to compare effect of the crisis with and without SP

SPA Welcoming cooperation 19 If we agree on the need to build a joint empirical basis for SP advocacy we can consider several options: Joint data platform Sharing and validating results Target specific groups (children, disabled) in the set of indicators Jointly producing and collecting data Joint use/development of software platform ADEPT surveys: it is a tool adoptable to any survey and can be customized It can greatly harmonize and streamline production of a rich set of indicators based on available surveys It informs collection of new data (obtain rich set of results by including 3 -5 new questions into the standard household surveys)

Data is a public good 20 We look forward to practical ideas for cooperation Thank you!