Document of The World Bank FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION AND RESULTS REPORT TF072201 ON A SMALL GRANT IN THE AMOUNT OF USD 1.9 MILLION TO THE Republic of Zimbabwe FOR Implementation support to the PICES survey in Zimbabwe (P158475) December 15, 2018 Poverty And Equity Global Practice Africa Region ABBREVIATIONS AND ACRONYMS PICES Poverty Income and Consumption Expenditure Survey SDG Sustainable Development Goal TA Technical Assistance WB World Bank ZIMREF Zimbabwe Reconstruction Fund ZIMSTAT Zimbabwe National Statistics Agency Regional Vice President: Hafez M. H. Ghanem Country Director: Paul Noumba Um Senior Global Practice Director: Carolina Sanchez-Paramo Practice Manager: Andrew L. Dabalen Task Team Leader(s): Robertus A Swinkels ICR Main Contributor: Robertus A Swinkels TABLE OF CONTENTS DATA SHEET ....................................................................... ERROR! BOOKMARK NOT DEFINED. I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES ....................................................... 4 II. OUTCOME ...................................................................................................................... 6 III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME .................................. 7 IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME .... 7 V. LESSONS LEARNED AND RECOMMENDATIONS ................................................................ 8 ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS ............................................................. 9 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) DATA SHEET BASIC INFORMATION Product Information Project ID Project Name Implementation support to the PICES survey in P158475 Zimbabwe Country Financing Instrument Zimbabwe Investment Project Financing Original EA Category Revised EA Category Organizations Borrower Implementing Agency Republic of Zimbabwe ZIMSTAT Project Development Objective (PDO) Original PDO The project development objective is to support Zimbabwe’s national statistics agency to collect high quality welfare data through the preparation and implementation of a household survey. Page 1 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) FINANCING FINANCE_T BL Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$) Donor Financing TF-A4121 1,900,000 1,900,000 1,900,000 Total 1,900,000 1,900,000 1,900,000 Total Project Cost 1,900,000 1,900,000 1,900,000 KEY DATES Approval Effectiveness Original Closing Actual Closing 13-Dec-2016 03-Feb-2017 15-Jun-2018 15-Jun-2018 RESTRUCTURING AND/OR ADDITIONAL FINANCING Date(s) Amount Disbursed (US$M) Key Revisions KEY RATINGS Outcome Bank Performance M&E Quality Moderately Satisfactory Satisfactory Substantial RATINGS OF PROJECT PERFORMANCE IN ISRs Actual No. Date ISR Archived DO Rating IP Rating Disbursements (US$M) ADM STAFF Role At Approval At ICR Regional Vice President: Makhtar Diop Hafez M. H. Ghanem Country Director: Ivan Velev Paul Noumba Um Senior Global Practice Director: Carlos Silva-Jauregui Carolina Sanchez Page 2 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) Practice Manager: Andrew L. Dabalen Andrew L. Dabalen Task Team Leader(s): Robertus A Swinkels Robertus A Swinkels ICR Contributing Author: Robertus A Swinkels Page 3 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES Context 1. Zimbabwe is a landlocked country with an estimated population of 16.5 million in 2017. Gross Domestic Product (GDP) that year was estimated at US$17.8 billion and Gross National Income (GNI) per capita at US$910. It is estimated that between 1999 and 2008, the economy contracted by more than 45 percent due to disruptions caused by the land reform program and flawed macroeconomic policies, which ultimately led to hyperinflation. The economic growth rebounded in 2009 and Zimbabwe’s GDP expanded by 25.2 percent over the period 2009-2011, at a constant annual rate of 7.8 percent. Despite the strong 2009-2011 economic performance, the country remains vulnerable to shocks and adverse weather conditions. GDP growth moderated to 4.6 percent in 2012, reduced to 2.9 in 2013 and went down to between 0.5 and 2.0 percent during 2014 -2016 only to pick up to 3.4 percent in 2017. 2. Zimbabwe is different from other fragile and conflict-affected states because it has the characteristics of a middle-income country. For example, backbone infrastructure and human capacity tend to be of higher quality that in other fragile states. Despite that, recent years have witnessed a loss of institutional capacity, especially in core government functions to deliver services to citizens. 3. The Poverty Income Consumption and Expenditure Survey (PICES) is conducted at five-year intervals and is the building block for national accounts data and for measuring poverty and inequality in the country . The survey is conducted by the Zimbabwe National Statistics Agency (ZIMSTAT) which is a corporate body that was established through the Census and Statistics Act of 2007. The last survey was conducted in 2011/12 funded by development partners with some World Bank technical assistance. A poverty profile based on these survey data was presented in a ZIMSTAT poverty report on 2013. Past estimates of poverty are not strictly comparable as the recall period for some non-food items has changed from 12 months to one month. No government funding is available for PICES surveys. 4. In 2015 the government requested the World Bank and other donors for financial and technical support to the implementation of a new PICES survey, to be implemented in 2016/17. The World Bank, together with other development partners, started engaging in discussions with the Zimbabwe national statistics agency on the design of the survey. The PICES survey has traditionally been used for national account estimation. During 2016 several changes were made to the PICES questionnaire to better reflect international good practice in consumption measurement and to ensure the Sustainable Development Goals (SDG) indicators can be tracked. Survey costs were estimated at US$ 3.5 million based on a sample size of 32,000 households. However, the World Bank proposed a considerably smaller sample size and accordingly agreed to fund US$ 1.9 million using funds from the WB managed ZIMREF Trust Fund. Other donors committed around US$ 400,000. The Bank agreed to fund the survey with a larger sample size only if donors were found to fill the funding gap, which consisted of US$ 1.2 million. In October 2016 such a donor was found. The World Bank also conditioned its support to the government agreeing to make the anonymized micro-data available to users. When the government of Page 4 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) Zimbabwe agreed to amend the legal framework for statistics to make this happen it was agreed that the survey would go ahead, and project preparation started. 5. Complementary Bank Executed technical assistance (P157466) was used to support survey design, data analysis, and improving the legal framework for data dissemination, while a Bank-executed activity P161054 funded a special survey-module on measuring agricultural productivity. Project Development Objectives (PDOs) 6. The project development objective was to support Zimbabwe’s national statistics agency to collect high quality welfare data through the preparation and implementation of a household survey: the PICES 2016/17. Key Expected Outcomes and Outcome Indicators 7. Achievement of the overall development objective was to be assessed through the following PDO level indicators: (i) A robust system for electronic data verification is designed and tested to ensure that it will immediately detect errors and inconsistencies. (ii) Proportion of sampled households for which complete data sets are collected and entered in the computer without major inconsistencies within 6 weeks after data collection. The Grant was structured in three Components 8. The project provided funds for the implementation of the PICES survey in conjunction with funding from other development partners. No government funding was available. As mentioned, it complemented World Bank technical assistance (TA) for improving data collection methodology including questionnaire design, sample design, data processing and analysis, data anonymization and dissemination. The project had three components: (1) survey design and preparation, and (2) data collection and data entry, and (3) project management. 9. Component 1: Survey design and preparation. This part included the finalization of the questionnaire through technical review, field testing, and validation with stakeholders as well as the development of the survey manual. This was followed by a training of trainers and enumerator training as well as questionnaire printing and design of data entry systems. Adaptations to be made to the questionnaire and the manual reflected agreements that will emerge from a range of technical discussions including on how long the household diary will be kept, defining the role of the enumerator in diary completion, ensuring consistency of data collection, flexible recall periods for non-food consumption, and transfer to the database. Possibilities for including a non-standard unit measurement survey were to be explored. 10. Component 2: Data collection and entry. This component was to cover the actual data collection as well as the quality control during field work. Data was to be gathered using paper questionnaires as well as through a Page 5 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) household consumption diary. Data from both instruments were to be entered in the computer in Harare, checked for errors and cleaned. 11. Component 3: Project management. This component aimed to provide support to ZIMSTAT with various fiduciary aspects of project implementation, including financial management, procurement and auditing activities. II. OUTCOME Assessment of Achievement of Each Objective/Outcome 12. The project ensured that a comprehensive set of welfare data were collected through the PICES 17 survey. As the questionnaire was amended to reflect many good practices on the collection of consumption data, and several quality control mechanisms were put in place, the data are likely to be of good quality and so the PDO was achieved. Overall Outcome Rating 13. The outcome rating is moderately satisfactory: Through the implementation of the PICES household survey, this project enabled the government of Zimbabwe to monitor the twin goals of ending extreme poverty and ensuring shared prosperity as well as facilitate measurement of SDG indicators. Data enable the measurement of poverty in a manner that is comparable to the previous PICES survey as well as the introduction of good practice measurement methods. The project also contributed to the availability of high quality data needed to inform policies for raising welfare of low-income households and evaluate the welfare impact of different subsidies and taxes. However, due to funding gap that remained after the start of the survey and the delays in procurement of the car and fuel, field supervision was less close then foreseen. This is likely to have somewhat impacted the quality of the data collection process. Also, not all data were entered in the computer within six weeks after collection due to funding shortages which may have somewhat impacted quality. Other Outcomes and Impacts 14. The funding of the project provided leverage for the World Bank to engage in many current and future in- depth technical discussions on data collection and analysis, micro-data anonymization and the release of anonymized micro-data to users. It strengthened the trust between the WB and ZIMSTAT and enabled the mobilisation of other trust funds resources (TFSCB) for attaching a detailed agricultural productivity module. It also provided an opening for testing alternative and likely more efficient food consumption measurement approaches that was conducted in the last month of the survey (December 2017). The project leveraged WB technical engagement with ZIMSTAT on survey approaches which has exposed ZIMSTAT to recent thinking on Page 6 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) consumption measurement, including on likely better practice approaches for both collecting and analyzing data. III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME 15. IMPLEMENTATION SUPPORT TO THE PICES SURVEY IN ZIMBABWE (P158475) was approved on December 13, 2016 and became effective on February 3, 2017. The project development objective was to support Zimbabwe’s national statistics agency to collect high quality welfare data through the preparation and implementation of a household survey: the PICES 2016/17. Overall, project’s implementation was Moderately Satisfactory. 16. There were some factors that affected the implementation of the project. The first one regards the funding gap of US$ 1 million. The donor that had committed to fill the remaining funding gap of that amount did not honor the commitment. The subsequent shortage of funds led to less field supervision by ZIMSTAT headquarter and provincial staff then foreseen. It also led to a shortage of funding for data entry which was later filled with additional BE ZIMREF trust fund resources. Second, the survey had started before the project was effective. This led to some enumerators having to advance their own travel costs. However, this impact was moderated by an advance of US$ 0.5 million of the Zimbabwean government. Third, considerable delays in the procurement of car and fuel for field visits made it necessary for ZIMSTAT provincial supervisors to advance the fuel costs of their supervision trips and to reduce their field supervision trips. The delays in procurement were caused by limited capacity of ZIMSTAT to follow World Bank procurement regulations, by the shortage of hard currency which led to some suppliers to request advances which World Bank procurement rules do not allow. The currency shortage also led to the quotation for the care purchase to become quickly outdated. Procurement delays were also caused by the limited internet capacity of ZIMSTAT to load the WB STEP program on their computers which led to ZIMSTAT procurement staff having to travel to the WB office to use the internet there. IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME 17. No safeguard policies were triggered. The grant’s activities were rated Category “C� for Social and Environmental Screening as it was not expected to have any direct social or environmental impacts. 18. Frequent country field visits (almost monthly) by the TTL ensured that survey technical issues were dealt with rapidly and support to fiduciary issues was organized rapidly. 19. The World Bank provided intense fiduciary and procurement support through field office FM and procurement specialists. Delays were entirely caused by ZIMSTAT’s lack of experience with World Bank financial management and procurement policies and procedures. The World Bank responded well to requests for support which prevented further delays. Page 7 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) 20. The World Bank provided hands-on training and follow-up support on World Bank procurement regulations. Frequent hands-on support was provided by local WB procurement specialists to ZIMSTAT procurement staff. Procurement plans had to be continuously updated to reflect delays in implementation. V. LESSONS LEARNED AND RECOMMENDATIONS 21. While agreeing to fund a less than perfect survey can be risky, it can also build considerable trust and subsequently provide opportunities to engage in subsequent technical discussions and gain influence to introduce technical improvements in various aspects of the survey including analysis as well as future data collection and other collaboration activities. This can be important in an environment where the relationship between the WB and the statistics agency needs strengthening. 22. Lack of previous experience with World Bank procedures, in an environment with weak institutional capacity, can pose significant challenges for the implementation of World Bank-financed projects. In hindsight it may have been better if the WB had hired procurement consultants to assist ZIMSTAT with procurement matters. 23. Although its institutional capacity still requires strengthening, the activity had a small but positive impact on the ZIMSTAT’s capacity to implement World Bank-financed projects, especially with respect to fiduciary and procurement procedures. 24. It would also have been better if the survey would have started a few months later, after the project was declared effective and not before. This would have prevented ZIMSTAT supervisory staff having to advance their own travel funds. 25. Agreeing to fund a survey based on a verbal agreement from another donor to fill the remaining funding gap is unwise and in hindsight it would have been better if an official communication from that donor had been obtained before a decision was made to go ahead. . Page 8 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS A. RESULTS INDICATORS A.1 PDO Indicators Objective/Outcome: Complete the PICES 2017 household survey in line with international standards Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Complete sets of field data are Percentage 0.00 100.00 100.00 100.00 collected and entered in the computer without major 16-Dec-2016 16-Dec-2016 28-Feb-2018 29-Jun-2018 inconsistencies and within one month after data are collected Comments (achievements against targets): All data were collected conform the schedule and entered in the computer within a few months A.2 Intermediate Results Indicators Component: Appropriate quality control mechanisms have been applied and data have been cleaned. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Page 9 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) A robust system for electronic Yes/No N Y Y Y data verification is designed and tested 16-Dec-2016 16-Dec-2016 28-Feb-2017 31-Jul-2018 Comments (achievements against targets): Page 10 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) B. ORGANIZATION OF THE ASSESSMENT OF THE PDO Objective/Outcome 1 1. A robust system for electronic data verification is designed and tested to ensure that it will immediately detect errors and inconsistencies. Outcome Indicators 2. Proportion of sampled households for which complete data sets are collected and entered in the computer without major inconsistencies within 6 weeks after data collection. 1.Survey preparation. Questionnaire reflects good practice and was field tested. Intensive enumerator training was conducted. 2. Data collection and data entry. Data were collected from all Key Outputs by Component households was completed as scheduled. Data were entered in the (linked to the achievement of the Objective/Outcome 1) computer in Harare with four months. 3. Project management. Field arrangements and payments to enumerators were made in time. No fiduciary issues were found in the audit. Page 11 of 11 The World Bank Implementation support to the PICES survey in Zimbabwe (P158475) . Page 12 of 11