Utilizing household survey microdata obtained from the ECAPOV database, the team has made significant progress in assembling, harmonizing, and organizing SPEED performance data. In the past few years, following earlier improvements to the data management approach, the team has managed to process and assimilate performance indicators for nearly all available surveys. These changes (discussed below) ensure a wider array of up-to-date, consistent data on crucial performance indicators, disseminated in an easily digestible format via a user-friendly interface that allows World Bank colleagues to easily inform their projects.
While the team has long relied on outputs created using ADePT, a program supported using STATA infrastructure to produce standardized performance statistics— including adequacy, coverage, relative incidence, and social welfare impact— these outputs had previously been created largely as one-off products in an ad hoc fashion. This current project has streamlined and regularized the production of these data to allow for comparability across countries and time.
These data rely on a combination of consumption and income data from household surveys processed by the World Bank Group Poverty Global Practice in collaboration with local statistical offices. SPEED makes use of this harmonized data and follows the same methodology across countries wherever possible. We have regularly engaged with the Poverty team to better understand breaks in the available series, clarify methodological questions, and document disparities. As a result, SPEED was able to include indicators that allow for an international comparison of the poverty impacts of social programs using the World Bank poverty lines at 1.90, 3.20 and 5.50 PPP USD. In addition to using the harmonized data, the team also resorted to using raw survey data to reflect local peculiarities of social protection systems in specific instances. These developments were also shared back with the Poverty team and incorporated in the harmonization process whenever relevant. The SPEED team plans to continue collaborating with the Poverty team on data sharing and validation to avoid discrepancies in the definitions of programs going forward.
An essential part of the methodology used to develop harmonized cross-country performance data is the standardized classification of social protection program categories. By applying the same methodology the ECA team already uses to categorize expenditure programs, the SPEED team was able to harmonize the classification of performance data (including sub-programs) across all available county-years through an iterative validation process. While country-specific experts obviously understand a given country’s situation best, ensuring comparability of social protection programs and sub-programs was a challenge that required extensive coordination and collaboration throughout the entire regional team.
To date we have harvested standardized data from the following available countries: Albania, Armenia, Azerbaijan, Belarus, Bosnia and Herzegovina, Bulgaria, Georgia, Hungary, Kazakhstan, Kosovo, Kyrgyz Republic, Croatia, Lithuania, North Macedonia, Moldova, Poland, Romania, Serbia, Slovak Republic, Tajikistan, Türkiye, Ukraine for all available years. Because separate ADePT output files must be created for each demographic breakdown, assembling this panel data set has required the production of more than 1,100 Adept outputs. In these intermediate outputs, available age, gender and poverty files are produced alongside an output representing the results for the general population. To organize this new collection of resource documents, strict protocol was implemented regarding the structure of performance data folders and the naming of files. This new system ensures that the current extensive library of files does not grow unwieldy. Importantly, this system ensures that these files will be searchable, and so usable, for data users for years to come.