June, 2015 PRO-POOR TARGETING IN IRAQ TOOLS FOR POVERTY TARGETING Step 1: Exclusion of conflict-affected governorates (Nineveh, Anbar, and Salah ad-din) PRO-POOR TARGETING IN IRAQ Tools for poverty targeting Step 2: Universal eligibility in poorest nahiyas (headcount rate> 60 percent) Step 3: Proxy Means Test formula applied in remaining areas and households THE TWIN CRISES OF 2014 WORLD BANK AND GOVERNMENT OF IRAQ, 2015/16 Iraq was plunged into two simultaneous crises in the second half of 2014, one driven by a sharp decline in oil prices, the other, by Islamic State militants. Since June 2014, crude oil prices per barrel have fallen from around 112 USD to 97 USD in September and 62 USD by December. Given Iraq s heavy dependence on oil as a share of GDP and exports, and a source of government revenues, this decline in prices alone would have hit Iraq s fragile economy hard. In addition, since June 2014, Islamic State (IS) or Da ash militants extended their influence from Syria into Iraq s northern and western provinces of Anbar, Nineveh, Salahadin, and to a lesser extent, Kirkuk and Diyala. A total of 354,000 families were internally displaced between June and December of 2014 which represents about 2.1 million individuals (Source: International Organization of Migration Displacement Tracking Matrix (IOM-DTM) January to December 25, 2014); and those left behind have been cut off from the rest of the country. The internally displaced persons (IDPs) have sought refuge across Iraq and about half of those who have crossed governorates boundaries have moved to Iraqi Kurdistan. Civilian casualties in Iraq have increased to close to 2007 levels (Iraq Body Count) as a result of the ongoing violence. The prevailing insecurity has disrupted oil exports from the north, adversely impacted trade and investment, led to the destruction of infrastructure and impeded the flow of goods and services across the country, leading to a sharp contraction in oil and non-oil GDP. Moreover, the sharp physical fragmentation of the country has impeded normal mechanisms of adjustment and likely magnified impacts. A TIME OF CRISIS AND A TIME FOR REFORM Social protection programs in Iraq are inefficient and fragmented. The main social protection program is the Public Distribution System (PDS), a food subsidy program, which reaches the poor almost universally: more than 99 percent of the poor receive PDS benefits. The high coverage rate of the poor comes at the cost of significant leakage of fiscal resources to the non-poor: about 98 percent of the nonpoor receive PDS benefits as well. The only cash transfer scheme in the country is the Social Safety Net (SSN) transfer that provides cash to eligible households defined as households with orphans, married students, disabled, blind, paralyzed, families of the imprisoned and missing persons, and the unemployed. Due to the categorical targeting criteria, which is blind to the monetary needs of a household, considerable share of the transfer recipients are non-poor and many poor households do not receive the benefits: only 11 percent of the poor receive benefits while of those receiving benefits, only 26 percent are poor.
2 Despite a difficult political and economic context, Iraq has embarked on a gradual reform of its social protection sector. The new Social Protection Law for Iraq (Law 11/2014) enacted in April 2014 establishes poverty and means testing as the primary targeting mechanisms. In addition, the Iraq Social Protection Strategic Roadmap, developed in November 2014, aims to have a comprehensive, integrated, and efficient social protection system for Iraq covering social safety nets, social insurance, and labor market policies. The recently drafted Social Protection Strategic Framework for the Kurdistan Region recognizes the inefficiencies in the social assistance programs brought about by poor targeting of beneficiaries. The framework aims to improve coverage of the poor and reduce leakage to the non-poor by adopting targeting mechanisms and formula that are better able to identify the poor. The shift is in harmony with the goals of Kurdistan Region of Iraq 2020: A Vision for the Future, a strategic policy paper issued in 2013, that aims to redesign the [social assistance] programs so that by 2020, assistance is targeted to those in need and creates opportunity for all. The move from categorical targeting to poverty targeting makes it necessary to identify the poverty status of households. Complex and elaborate surveys are required to calculate household expenditure, poverty line, and poverty status. It is not practical to carry out expenditure surveys for the universe of potential beneficiary households; the surveys are expensive, they take time, and it takes specialists to design the survey and compute total household expenditure. What is needed instead is a short and simple tool that can be implemented by social workers and used by non-specialists to estimate household poverty status in a short time and under a small operating budget. Proxy Means Test is a simple and cost-effective tool that has been in use at least since mid-1990s to identify poor households in developing countries. PROXY MEANS TEST: METHODOLOGY The central idea behind proxy means test is simple: one uses proxies of welfare (income or expenditure) to rank households and judge their welfare status. To prevent households from strategically misreporting the information, the proxies must be observable or nearobservable and verifiable such as family size, dwelling type, and ownership of assets like car, motorbike, refrigerator, heater, etc. In particular, the variables must satisfy the following three criteria: i. Difficult to misreport or manipulate ii. Easy to observe, measure, and verify iii. Highly correlated to expenditure and explains a high share of the total variation in consumption The proxy variables and the corresponding weights are identified using ordinary least squares (OLS) regression. Once the data is collected from potential beneficiaries and the formula is applied, each household receives a score and is categorized as poor or non-poor depending on a predetermined threshold. Eligibility criteria The eligibility threshold that determines who gets the benefit and who does not depends on the intended level of coverage, the availability of fiscal resources, and the projected increase in resources required over time. Nevertheless, in order to achieve a reasonable coverage rate and rates of error, and considering that the proposed targeting reform focuses only on the existing SSN beneficiaries, it is recommended that a relatively generous threshold of 60th percentile from the 2012 expenditure distribution be used to determine eligibility. In other words, all households whose formula-predicted expenditure is less than the 60 th percentile value from the 2012 expenditure distribution will be eligible for the benefits. Non-eligible Eligible Non-poor Accurate prediction Inclusion error Poor Exclusion error Accurate prediction Errors Although the model is built to get as accurate prediction as possible, there are errors inherent to the targeting formula. Proxy means test suffers from two types of error: exclusion and inclusion. Exclusion error, also known as undercoverage, refers to incorrectly categorizing a poor household as non-eligible. Inclusion error or leakage points to the opposite problem identifying a household as eligible when it is in fact not poor. A good PMT formula is one that minimizes the errors. PROPOSAL: MULTI-STAGE TARGETING Applying various targeting methods in different combinations and sequences may be more effective and efficient in reaching the poorest. For example, in areas where poverty rate is extremely high, it is not efficient and cost-effective to carry out a PMT survey. In such areas, social assistance programs may be offered universally to all households. A hierarchical, multi-stage targeting structure is
3 recommended for governorates outside Kurdistan given the security situation in the country and prevalence of deep pockets of poverty. Conflict-affected governorates are excluded and residents of nahiyas with poverty rate of more than 60 percent are deemed automatically eligible for the cash transfer. The PMT formula is applied in remaining areas and among remaining households to determine eligibility. Geographically segregated areas with deep pockets of poverty do not exist in Kurdistan so the gains from first-stage geographical targeting in Kurdistan is minimal. Therefore, it is recommended that a PMT formula be applied to all households in the region to determine program eligibility. Step 1: Exclusion of conflict-affected governorates (Nineveh, Anbar, and Salah ad-din) Step 2: Universal eligibility in poorest nahiyas (headcount rate> 60 percent) Step 3: PMT formula applied in remaining areas and households PMT SCORECARDS To better capture the profile of the poor, separate scorecards are developed for Kurdistan, Baghdad, and Rest of Iraq. Household size is a strong predictor of household welfare in all three regions, with smaller households generally better off than larger ones. Ownership of durable assets like air conditioner, generator, PC, vacuum cleaner, car, fridge, freezer, and washing machine also indicate higher level of household welfare. Households that have access to water and sewage services through the public network are well off than households that do not. In Kurdistan, households that own their own house and live in urban areas or in Sulaimaniya are better off than other households. The PMT formula can be updated only so often when a new household survey becomes available so the scorecard has to retain its predictive power over the short run. The scorecard must also minimize the targeting errors with a parsimonious set of proxies. Several variations of the scorecard were tried and tested before choosing the optimal one. The quality and accuracy of the scorecards stand the test of international comparison when compared with the scorecards developed for other countries. An application of the proposed multi-step targeting scheme on current SSN beneficiaries in Qadisiya shows that it could cut total social safety net expenditure by a third. The savings comes from fewer household being eligible for cash transfers. When the hierarchical targeting is applied in conjunction to the new benefits schedule, it more than triples the average benefits per household while only doubling total annual expenditure. This becomes possible due fewer, but poorer, households becoming eligible for benefits. Testing the targeting proposal in Qadisiya Total annual expenditure (billion IDs) 51.03 34.29 102.41 Number of beneficiary households 37753 24301 24301 Current Multi-stage targeting Multi-stage targeting & old benefit scheme & new benefit scheme Current Multi-stage targeting Multi-stage targeting & old benefit scheme & new benefit scheme Annual benefit per household (million IDs) 1352 1411 4214 Current Multi-stage targeting & old benefit scheme Multi-stage targeting & new benefit scheme
TARGETING PMT scorecard for Baghdad 1 Number of household members 1 to 4 5 to 7 8 to 10 10 or more 2 Number of children aged 0-6 0-2 3 or 4 5 to 7 7 or more 3 Number of children aged 7-17 0-2 3,4 or 5 6 or more 4 Age of the household head 20-29 30-39 40-49 50-59 60-69 70 or more 5 Does the household own [ ]? Yes No Air conditioner Generator PC Vacuum Washing machine Car 6 Type of sewage Public network Septic tank Covered drain, open drain, other
5 PMT scorecard for Rest of Iraq 1 Urban 2 Household size 1 to 4 5 to 7 8 to 10 10 or more 3 Number of children aged 0-6 0-2 3 or 4 5 to 7 7 or more 4 Number of children aged 7-17 0-2 3,4 or 5 6 or more 5 Age of the household head 20-29 30-39 40-49 50-59 60-69 70 or more 6 Number of household members aged 60 or older 0 1 2 3 4 or more 7 Does the household own [ ]? Yes No Car Vacuum PC Freezer Washing machine Water heater Refrigerator Air conditioner Generator 8 Water Source Public network: housing unit connected, public network tap Tanker Other 9 Sewage type Public network Septic tank Covered drain, open drain, other 10 Type of household unit House/flat Clay house/bamboo house/other
6 PMT scorecard for Kurdistan 1 Number of household members 1 to 4 5 to 7 8 to 10 10 or more 2 Urban 3 Sulaimaniya 4 Number of children aged 0-6 0 1 2 or more 5 Number of children aged 7-17 0 1 2 or more 6 Age of the household head 20-29 30-39 40-49 50-59 60-69 70 or more 7 Does the household own [ ]? Yes No PC Water heater Vacuum Car Fridge Freezer AC Generator This note summarizes the work undertaken by the World Bank in collaboration with the Ministry of Planning and Central Statistical Organization, Iraq and the Kurdistan Region Statistics Office.