Mid-Term Evaluation Report on the Enhanced Data Dissemination Initiative Project

This report presents the results of the mid-term evaluation of the Enhanced Data Dissemination Initiative (EDDI) financed by the United Kingdom’s Department for International Development (DFID) covering the period April 1, 2010 to September 30, 2012. The evaluation was conducted internally by the IMF in consultation with DFID. EDDI is a five-year project (April 2010–March 2015) implemented by the IMF to improve macroeconomic statistics in 25 African countries. The project includes modules for sub-groups of countries covering national accounts, monetary statistics, government finance statistics (GFS), balance of payments statistics (BOP), and harmonization of statistics in several regional organizations. The mid-point of a five-year project is an appropriate time for all stakeholders of the project to step back and take stock of what has been accomplished in the first half of the project, what has gone well, what aspects have been disappointing, and what might be adjusted or changed to make the remainder of the project more effective in achieving its objectives. To facilitate this process, questionnaires were developed to obtain feedback from three groups: counterparts in participating countries, IMF module managers and experts, and DFID country and regional advisors. Recommendations made by the stakeholders that will be followed up in the second half of the project are listed as bullets in italics below.

Abstract

This report presents the results of the mid-term evaluation of the Enhanced Data Dissemination Initiative (EDDI) financed by the United Kingdom’s Department for International Development (DFID) covering the period April 1, 2010 to September 30, 2012. The evaluation was conducted internally by the IMF in consultation with DFID. EDDI is a five-year project (April 2010–March 2015) implemented by the IMF to improve macroeconomic statistics in 25 African countries. The project includes modules for sub-groups of countries covering national accounts, monetary statistics, government finance statistics (GFS), balance of payments statistics (BOP), and harmonization of statistics in several regional organizations. The mid-point of a five-year project is an appropriate time for all stakeholders of the project to step back and take stock of what has been accomplished in the first half of the project, what has gone well, what aspects have been disappointing, and what might be adjusted or changed to make the remainder of the project more effective in achieving its objectives. To facilitate this process, questionnaires were developed to obtain feedback from three groups: counterparts in participating countries, IMF module managers and experts, and DFID country and regional advisors. Recommendations made by the stakeholders that will be followed up in the second half of the project are listed as bullets in italics below.

Executive Summary

This report presents the results of the mid-term evaluation of the Enhanced Data Dissemination Initiative (EDDI) financed by the United Kingdom’s Department for International Development (DFID) covering the period April 1, 2010 to September 30, 2012.1 The evaluation was conducted internally by the IMF in consultation with DFID. EDDI is a five-year project (April 2010–March 2015) implemented by the IMF to improve macroeconomic statistics in 25 African countries. The project includes modules for sub-groups of countries covering national accounts, monetary statistics, government finance statistics (GFS), balance of payments statistics (BOP), and harmonization of statistics in several regional organizations.

The mid-point of a five-year project is an appropriate time for all stakeholders of the project to step back and take stock of what has been accomplished in the first half of the project, what has gone well, what aspects have been disappointing, and what might be adjusted or changed to make the remainder of the project more effective in achieving its objectives. To facilitate this process, questionnaires were developed to obtain feedback from three groups: counterparts in participating countries, IMF module managers and experts, and DFID country and regional advisors. Recommendations made by the stakeholders that will be followed up in the second half of the project are listed as bullets in italics below.

It was encouraging that the response rate of country counterparts to the 34 questionnaires sent out was 85 percent, and that 89 percent of the participating countries sent at least one response. Considering that the questionnaire was two pages in length and contained numerous verbal questions that required some time to complete, it is fair to say that the response rate is indicative of the level of commitment and ownership by the countries.

The results of the quantitative ratings questions (on a 1–4 scale, 1 most negative and 4 most positive) varied from 3.5 for effectiveness to 3.8 for relevance, with 3.6 recorded for each of the three sections efficiency, impact, and sustainability.

The responses to the verbal questions were consistent with the quantitative ratings. Accordingly, for the highest rated category, relevance, there were no recommendations for changes, indicating that the EDDI modules were consistent with the countries’ own national objectives and priorities. Many examples were provided.

The lowest rated category, effectiveness, not surprisingly had the largest number of suggestions for improvement. That said, it must be noted that 3.5 is still a very good rating. Of the four questions under effectiveness, the two questions accounting for the relatively lower rating were related to whether the TA advice has been timely and whether the work plan has been realistic. The other two questions on the quality of TA advice and that TA advice was easily understood scored quite high. Most of the verbal responses said that they did not have any suggestions for improvement, or complimented one or more aspects of effectiveness, and two countries suggested improvements that they themselves should make to improve accountability and responsibility for implementing the project’s work plan.

The suggestions for improvement of effectiveness were almost all related to improving the project’s flexibility to accommodate countries that feel they are not getting all the assistance they need to achieve the project’s objectives. These countries comprised a fairly small group of the less advanced and needier EDDI countries, mostly in the national accounts area. They recommended increased flexibility to provide more assistance in the following ways:

  • More frequent missions to countries especially in need.

  • More lengthy missions (e.g., 3–4 weeks instead of 2) to countries especially in need.

  • More workshops or training courses to increase capacity to absorb TA (national accounts and GFS).

  • Timing missions when new source data are ready.

The IMF appreciates the seriousness of the concerns expressed by these suggestions. Indeed, they reinforce an awareness that began to develop as early as the opening workshops, where it was apparent that a few countries in national accounts were starting from much lower baselines than the others. One country insisted from the beginning that the proposed work plan was not enough and that they needed a full-time resident expert to progress as quickly as they needed to satisfy the demands of their policy makers. Unfortunately, the budgetary limits of EDDI and overall project design cannot accommodate such extensive needs by any one country.

Regarding ratings for efficiency, most of the responses again said overall that they had no issues and that the project is going well. Of the countries that did provide suggestions for improvement, most were the same countries repeating their suggestions made under effectiveness on the frequency, length, and number of missions and workshops. Two countries made suggestions to:

  • Extend EDDI’s TA and training beyond the main government agencies to include data reporting institutions.

The category impact received no ratings below 3.5 indicating the countries that responded thought they were already achieving positive results in the project. This was borne out in the verbal responses. There were no suggestions for changes or improvements, and nearly all countries listed examples of results achieved, often multiple results.

The category sustainability earned no ratings of 2 or below, indicating that all responding countries either agreed or agreed strongly that their results would be sustainable. The most common response related to the increased capacity of their institutions created by the project in terms of the skills and numbers of staff, and in terms of new systems that are being put in place (databases, methodologies, and work processes) and are beginning to be used as part of standard operating procedures.

Increased awareness and demand for data by policy makers was mentioned as a factor. Countries also pointed to their own efforts, in terms of increased dedicated staff, creation of new departments or divisions, and their own financial support of the project, in terms of training their staff, funding surveys, and budgetary support.

A number of countries described the potential risks that may threaten sustainability. Most of these responses said that sustainability depends on continued budgetary resources to maintain and train staff and continued budgetary resources for surveys and data collection.

The feedback from IMF module managers and experts was generally consistent with that of country counterparts on the good progress made in the first half of the project. In looking to the second half of the project, IMF module managers and experts said that they would reinforce adjustments already made, such as continuing more intensive assistance to countries in relatively greater need. In addition, since inadequate human resources and budgetary support for project implementation was identified as a major factor impeding progress in some countries, they recommended in the second half of the project to:

  • Sensitize high-level management of the main statistical agencies (statistical offices, central banks, and finance ministries) of the importance of continued funding for implementation of the project. This can be done by IMF African Department missions to countries, high-level meetings in Washington at the annual and spring meetings, and communications by senior IMF Statistics Department staff as needed with particular countries.

The feedback from DFID regional and country advisors was also generally consistent with the country counterparts on the good progress made in the first half of EDDI and the concrete results achieved. The project’s relevance was rated as high by 100 percent of responses, who also unanimously rated the IMF experts’ quality of advice as high. IMF advice was rated as timely by 80 percent of responses. A number of good examples of relevance and impact in countries were provided. Most of the concerns and suggestions for adjustments were in two areas, better communication/collaboration by IMF experts with local stakeholders and lengthier missions to certain countries in need. It was emphasized that this was important for ensuring that relevance and high priority are maintained during the intervals between missions, and another suggested that it was useful for organizing support for required funding of activities complementary to EDDI.

  • More IMF experts will meet with DFID country advisors while on missions, and with other donors that are known to be engaged in work closely related to EDDI. DFID advisors will where possible assist in organizing meeting with significant other donors/stakeholders.

In conclusion, the positive overall results of the mid-term evaluation of EDDI and feedback from major stakeholders are consistent with the performance of EDDI with respect to annual project milestones. The significant number of concrete impacts and results already achieved and noted by many countries testifies to the project’s value to the recipient countries. Although no major adjustments or changes need to be made, a number of expressed concerns will be followed up and adjustments will continue to be made as needed in specific cases. Both the IMF and DFID project managers need to continue to push to maintain the positive momentum created in the first half of the project. A danger is that successes so far have partially been achieved by picking “the lower hanging fruit” and that the more difficult and challenging achievements are those that remain for the second half of the project. Some of these more difficult challenges may well demand some reallocation of EDDI resources to support needier and less developed participants, which hopefully will be freed up by some of the successful participants requiring fewer resources in the second half and also by increasing collaboration with other sources of funding.

I. Introduction

1. The purpose of this report is to present the results of the mid-term evaluation of the Enhanced Data Dissemination Initiative (EDDI) financed by the United Kingdom’s Department for International Development (DFID) covering the period April 1, 2010 to September 30, 2012. The evaluation was conducted internally by the IMF in consultation with DFID. EDDI is a five-year project (April 2010–March 2015) implemented by the IMF to improve macroeconomic statistics in 25 African countries.2 It builds on the successes of two preceding DFID-funded statistics projects for Africa implemented by the IMF and the World Bank between 2002 and 2009. These resulted in the widespread participation of African countries in the General Data Dissemination Standard (GDDS) and the improvement of members’ economic and financial statistics.

2. Section II provides a brief background on the EDDI project. Section III presents feedback obtained from the three main stakeholders: country counterparts in recipient countries, IMF managers and experts, and DFID country and regional advisors. Section IV presents the recommendations for follow up.

II. Background

3. EDDI employs a modular approach in providing technical assistance (TA) and training, focusing on relatively small groups of countries with similar needs and grouped according to sub-regions where possible. The effectiveness of the modular approach was confirmed by the results achieved in the previous GDDS Phase 2 project. Included are seven topical modules (three monetary statistics modules, one balance of payments (BOP) module, two national accounts modules, and one government finance statistics (GFS) module), and four functional modules (three data dissemination modules for East, South, and West Africa plus a module focused on data harmonization in the Southern Africa Customs Union (SACU)). Table 1 presents the country participation in each module.

Table 1.

EDDI Modules and Countries3

article image

4. The first monetary statistics module focuses on expanding the coverage of these statistics to include a larger share of the total banking and financial system, and ensuring that the compilation of these statistics is consistent with the IMF Monetary and Financial Statistics Manual and companion Compilation Guide. The second monetary module assists two regional organizations, the East African Community (EAC) and the Southern African Development Community (SADC), and their member countries to harmonize their monetary statistics in preparation for forming monetary unions. The third monetary module is helping a group of African countries vulnerable to the current financial crisis to develop financial soundness indicators (FSIs) to assess and monitor the soundness of their financial systems.

5. The BOP statistics module is assisting participating countries to develop and improve surveys used to gather data on private capital flows and stocks for use in producing international investment position (IIP) statistics, as these statistics are growing increasingly important for many African countries.

6. The first national accounts module is focusing on improving annual national accounts in three West African countries. The second national accounts module is concentrating on helping a group of relatively more advanced countries to produce quarterly national accounts (QNA), which greatly increase the ability of policy makers to monitor developments in the economy. The QNA module is being implemented jointly with the AFRITAC East.

7. The main objective of the GFS module is to improve the compilation and dissemination of GFS statistics within a group of countries in east Africa, in order to be consistent with the IMF Government Finance Statistics Manual 2001.

8. The three data dissemination modules (for eastern, southern, and western Africa) include all project countries. The focus is on assisting two countries (Burundi and South Sudan) to join the GDDS, and also to assist a small group of relatively advanced countries to reach the stage where they can subscribe to SDDS. In addition, the project is helping as many countries as possible to begin to produce and publish national summary data pages (NSDPs) and advance release calendars (ARCs).

9. The SACU module is assisting the SACU and its member countries harmonize their statistical treatment of SACU transactions. This module is being implemented jointly with the AFRITAC South.

10. Finally, the EDDI project is supporting the production of manuals and guides specifically designed to address the statistical compilation issues faced by developing countries. Such a guide has already been produced in the GFS area, and another is in process in the BOP area.

11. A summary of the key results achieved in the first half of the project is provided in Box 1.

Highlights of Key Results Achieved in First Half of EDDI

  • Burundi joined the GDDS on August 9, 2011, following the satisfaction of all requirements.

  • Mauritius subscribed to the SDDS on February 28, 2012, following a successful BOP mission that verified it had satisfied the last remaining requirement regarding coverage of offshore enterprises.

  • All countries in the BOP statistics module completed private capital flows surveys (PCFS) and have begun to analyze and incorporate results, and Ghana has begun to publish international investment position (IIP) statistics for the first time based on PCFS data.

  • Ghana, Rwanda, and Uganda published for the first time QNA, and Zanzibar4 also released its first QNA.

  • Six additional countries have begun to publish National Summary Data Pages and eight additional countries have begun to publish Advance Release Calendars.

  • Following the preparation of institutional tables and bridge tables, provisional GFS data for central government according to the Government Finance Statistics Manual 2001 were compiled for recent years for Burundi, Kenya, Malawi, Mozambique, Rwanda, and Zimbabwe.

  • The Government Finance Statistics Compilation Guide for Developing Countries was published on September 14, 2011.

  • Monetary statistics TA focused on expanding coverage to leasing and mortgage companies in Botswana, savings and loan companies, rural banks, and finance houses in Ghana, money market unit trusts and microfinance institutions in Kenya, financial cooperatives and collective investment schemes in Lesotho, savings and credit cooperatives and money market funds in Swaziland, and savings and credit cooperatives in Uganda.

  • Three additional countries (Mauritius, Seychelles, and Uganda) have begun to publish FSIs.

III. Feedback from Stakeholders

12. The mid-point of a five-year project is an appropriate time for all stakeholders of the project to step back and take stock of what has been accomplished in the first half of the project, what things have gone well, what aspects have been disappointing, and what looking ahead to the second half might be adjusted or changed to make the project more effective in achieving its objectives. To facilitate this process, questionnaires were developed to obtain this feedback from three groups: Counterparts in participating countries, IMF module managers and experts, and DFID country and regional advisors (see Appendices I, II, and III).

A. Counterparts in Participating Countries

13. Questionnaires (questions are listed below) were sent to 18 countries and two regional organizations that participated actively in the first half of the EDDI project.5 The countries that were not sent questionnaires had either no or very little activity (Eritrea, Ethiopia, Seychelles, South Africa, Sudan, South Sudan, and Zambia). Some of the countries participated in more than one module and therefore received more than one questionnaire. A total of 34 questionnaires were sent, and of these, 29 responses were received for an 85 percent response rate. Of the 18 countries that were sent one or more questionnaires, 16 sent at least one response, for an 89 percent country representation rate.

14. All modules were well represented. For national accounts (including QNA), 11 out of 11 responses were received, for BOP, 5 out of 6 were received, for monetary statistics, 7 out of 10 were received, for GFS, 4 out of 6 were received, and for the regional organizations, 2 out of 2 were received.

15. The questionnaire comprises five sections (relevance, effectiveness, efficiency, impact, and sustainability). At the beginning of each section are one or more statements that were responded to with ratings (1 = I strongly disagree, 2 = I disagree, 3 = I agree, 4 = I strongly agree, na = not applicable to me). At the end of each section are one or two questions to be responded to verbally. Higher ratings indicate stronger performance of EDDI. The overall quantitative averages of the ratings are shown in Table 2. Responses showed that the average section ratings ranged from 3.5 for effectiveness to 3.8 for relevance, with 3.6 recorded for each of the three sections efficiency, impact, and sustainability. Chart 1 demonstrates more strikingly how strongly positive the results were.

Table 2.

Average Ratings from Questionnaires

article image
Chart 1.
Chart 1.

Share of Respondents that Agree or Disagree with Statements in the Questionnaire

Citation: Policy Papers 2013, 022; 10.5089/9781498342117.007.A001

16. The average ratings for modules varied between 3.4 to 3.8, with modules for regional organizations at 3.8, monetary statistics at 3.7, national accounts at 3.6, GFS at 3.5, and BOP at 3.4. The questions that earned the highest average ratings were related to relevance to the country at 3.9, consistency with the agency’s priority needs at 3.7, high quality advice of IMF experts at 3.7, and helpfulness of sharing lessons with other countries at 3.7. Also ranking high at 3.6 were IMF experts providing advice that is easily understood, positive results already being achieved, and sustainability of the results.

17. There were only four 2 ratings, three for national accounts and one for regional organizations, and no 1 ratings. The 2 ratings for national accounts were in relation to the questions involving timing of expert’s advice, work plan being clear and realistic, and project is well organized. A regional organization also gave a 2 rating for the question on the clarity and realism of the work plan.

18. The verbal responses to the questions that followed the ratings in each section generally provided reasons or evidence to support the ratings given. In trying to summarize all of these responses, it is not possible of course to mention every point made by every respondent. Emphasis will be placed on responses that were repeated multiple times. Some exceptions to this will be for criticisms or suggestions that are made for particularly difficult cases, or where the suggestions made seem worth mentioning in terms of reasonable adjustments that might be made in the project. Selected direct quotes will be provided in cases where they seem to summarize nicely a point made by a good sample of respondents, or where they are considered especially meaningful to a better understanding of how the project has been experienced and judged by those receiving the assistance and training.

Relevance—To what extent does the project fit with your institutional and national objectives and policies? Please expand on your response.

19. Just as this question received the highest average rating at 3.9, the verbal responses were overwhelmingly positive. Most emphasized the importance of EDDI’s objectives to improving policy making and their consistency with their national strategies and plans. In national accounts Sierra Leone pointed to the fact that three objectives in their National Strategy for the Development of Statistics (NSDS) have already been achieved by EDDI (rebasing national accounts, improved CPI, and a new PPI). Rwanda, who last year published its first QNA, said that EDDI provides “indicators that are very important not only for monitoring the economy but also for economic policy and decision making.” Ghana said that its first publishing of QNA was “a fulfillment of our institutional objective,” and that these data are “in great demand by policy makers, especially the Ministry of Finance and Economic Planning and the Bank of Ghana, as well as researchers.” In the monetary module, Sierra Leone summed up nicely the views of a number of other countries in the module by saying that EDDI “fits perfectly well the institution’s objective of achieving and maintaining price stability,” which “is a prerequisite for proper monetary policy reporting.” Kenya mentioned the project’s relevance to the regional objective of harmonization of statistics in the EAC. The GFS module countries also emphasized the link to policy making. In the BOP module, Nigeria emphasized how EDDI was helping in the achievement of the objectives of the country’s Vision 20 2020 where strategic planning and informed policy decisions are emphasized. Mauritius mentioned that the project “came at the right time for us when we were thinking about moving from indirect to direct collection of BOP data.” Mozambique said that EDDI was consistent with their objective of expanding coverage and improving the accuracy and timeliness of their statistics.

Effectiveness—Do you have any suggestions on ways to improve the project’s effectiveness in the remainder of the project?

20. The effectiveness average rating of 3.5 was the lowest of the five sections. Of the four ratings questions on effectiveness, the two questions on timeliness of experts’ advice (3.4) and the clarity and realism of the work plan (3.3) were among the lowest three of all questions (i.e., only three questions rated below 3.5 out of fifteen questions). These two questions accounted for three of the four 2 ratings given for all questions. A good number of the responses that made suggestions for adjustments (many did not make any suggestions) involved requests for more frequent or more lengthy missions, and/or more frequent training and workshops.

21. The 2 rating for timeliness of advice is at least partly a result of a lengthy interruption of TA to Nigeria because of security problems. Nigeria also suggests longer missions, from 3 to 4 weeks long. One of the two 2 ratings for clarity and realism of the work plan was by Liberia, a country that had the greatest need in national accounts of all the module countries. Because progress was challenging and difficult in the first two years, the IMF requested and received approval for IMF funding of a 3-month expert, who completed his work in August 2012. Real progress was achieved during this period and Liberia is now expected to produce more complete and reliable GDP estimates. This was a good example of the IMF providing funding for a necessary supplement to the EDDI project, as it is not within the terms of reference or budget of EDDI to fund longer-term experts for 3 months or more.

22. The other 2 rating came from a concern that the objective of harmonization of monetary statistics may not be achieved by the end of the project because a number of countries are lagging behind in implementation. For this reason they are suggesting support for follow up individual missions to these countries. Provision for this was not made in the EDDI budget, so the scope for such follow up missions are limited.

23. Two of the three West African countries in the national accounts module with relatively large TA needs, The Gambia and Liberia, asked for a training course. The IMF is aware of this need and is planning such a training course in the spring of 2013 which will also include Sierra Leone, the third country in the module.

24. Three countries suggested providing relatively more time to countries in need, rather than always doing two-week missions or at set intervals. To a certain extent this has occurred in the first half, with a number of instances of having three-week missions rather than two-week missions because of particular needs (especially in national accounts for West African countries), and conducting missions with shorter intervals in between because of particular needs.

25. Two countries made suggestions about the content of the TA, with one in national accounts suggesting that more demonstrations with actual country data would be helpful, and one country in GFS suggesting that they would like more practical tools. Another country in GFS, on the other hand, complimented the TA for providing good tools and templates.

26. Two countries suggested adjustments that should be made by the countries themselves. One country in national accounts said that its statistical office should designate particular people who are responsible for following up on the recommendations of missions, and another in monetary statistics said that mandatory requirements should be put on countries to do regular progress reports on their implementation of EDDI TA.

Efficiency—Do you have any suggestions on ways to improve the administrative arrangements or flexibility of the project to better meet your needs?

27. The average rating for the questions on efficiency was 3.6, with the value of sharing lessons with other countries at 3.7 and the organization of the project at 3.4. The relatively lower score given to organization (including the remaining 2 rating) resulted partly because a few respondents reiterated their desire for greater frequency and length of missions (which they had already mentioned under effectiveness) as a negative aspect of the project’s work plan and organization.

28. Most of the responses either provided no verbal comments or paid a compliment to the project’s efficiency indicating it was going well so far. Two countries suggested that the TA and training be extended beyond the main government agencies, primarily data reporting institutions. Botswana in monetary statistics recommended extending training beyond the central bank to other financial institutions that are intended to report data to the central bank. In fact, this training beyond the central bank is planned for the next mission. Nigeria in BOP similarly recommended “a team approach” where reporting institutions could be included in training. These are both good recommendations that will be taken on board during the rest of the project.

Impact—If positive results are being achieved, please provide examples.

29. The responses to this question were very encouraging. The average rating for this section was 3.6, and no ratings below 3.5 indicate that all of the countries that responded thought they were already achieving concrete results in the project. This was borne out in the verbal responses. Nearly all countries listed examples of results achieved, often multiple results. Many of these results are not listed in the EDDI project logframe, demonstrating the value of this kind of feedback that provides a greater texture and context to complement the broad strokes of the logframe.

30. Ghana remarked that the newly published QNA “is now a major input in analyzing short-term performance of the economy by the Ministry of Finance and Economic Planning and the Central Bank of Ghana” and two other countries made similar comments about their new QNA. Sierra Leone reported their GDP was rebased and Liberia had its first post-war national accounts survey analyzed and new GDP estimates linked to updated benchmark figures for 2008. Rwanda said that better use of survey data significantly reduced the use of proxy indicators such as population growth. Kenya, Tanzania, and Uganda remarked that the Work in Progress method has significantly improved agriculture statistics. The Gambia noted that it had migrated successfully from the 1968 SNA to 1993 SNA and that its new estimates were now widely accepted by users. Sierra Leone pointed to the importance of the TA on price statistics as it has led to a PPI ready to be published for the first time and an improved CPI. Zanzibar also commented on the value of assistance to improve its CPI.

31. A number of countries reported significant improvement in the coverage of monetary statistics (Lesotho, Namibia, Sierra Leone, and Uganda) to include some important additional depository and other financial institutions. Kenya and Swaziland noted improved classification of monetary statistics. Botswana remarked on better data collection instruments adopted and the good results of training that was provided to data reporting institutions. The GFS module is not yet at the stage where results are apparent in published statistics, except for Mozambique that reported on the publication on the ministry’s website of new GFS data.

32. For the BOP module, Ghana noted the “tremendous improvement in our BOP statistics compilation, especially in the coverage and reporting on the financial account.” Nigeria, Mozambique, and Mauritius also commented on the value of the surveys conducted and the expansion of coverage in BOP statistics that resulted. Nigeria also mentioned improved classification and interagency collaboration, while The Gambia noted improved compilation methodology and BOP statistics.

33. The EAC and SACU both commended the fact that agreement had been reached by the membership of both regional organizations on the way forward to improve harmonization and had both agreed on work plans that were now underway.

What are the most important changes in your agency as a result of the project?

34. By far the most common response to this question was the increase in skills and numbers of both compilation and management staff (Ghana, Liberia, Namibia, and Uganda in national accounts, Sierra Leone in monetary statistics, Mozambique and Rwanda in GFS, and The Gambia and Uganda in BOP).

35. Better communication and collaboration with other agencies and institutions was also mentioned by several countries. Kenya (national accounts) noted improved communication and interaction with data providers. Botswana said that its central bank Research Department is now working more closely with the Bank Supervision Department to increase the coverage and quality of data reported by financial institutions.

36. Greater awareness and changes of institutional attitude was pointed to in several countries. Nigeria (national accounts) said that it was now more aware of what needs to be done to improve statistics. Rwanda said its central bank is now more aware than it was before of the policy value of GFS data, and as a result its demand for GFS data has increased. Mauritius noted that “We have become more ambitious. Since we have gathered expertise on survey design, we are now thinking of leapfrogging to try to develop a web-based survey form linked with in-house data management software to process the data.”

37. Better institutional organization was mentioned by Nigeria (BOP) saying that it now had “better organized groups for the task management and management support.” Rwanda (national accounts) mentioned that it is now doing regular quality assessments of QNA and monitoring of the project work plan.

38. A number of countries responded by describing results that belonged under the previous question rather than changes in their agencies. They were therefore mentioned in the text under the previous question, to avoid confusion.

Sustainability—To what extent do you think the project will lead to a sustainable improvement in the quality and dissemination of data by your agency?

39. The average rating for whether the project would be sustainable was 3.6. There were no ratings of 2 or below, indicating that all responding countries either agreed or agreed strongly that their results would be sustainable. The most common response related to the increased capacity of their institutions created by the project in terms of the skills and numbers of staff, but also in terms of new systems that are being put in place (databases, methodologies, and work processes) and are beginning to be used as part of standard operating procedures.

40. The countries that specifically mentioned the importance of skills of staff for national accounts were Namibia, Rwanda, Tanzania, and Uganda, and for monetary statistics—Lesotho, Namibia, and Sierra Leone.

41. In terms of new systems put in place, Namibia (monetary databases), Rwanda (national accounts), and Swaziland (monetary databases) referred to the systems as a whole which presumably include databases, methodologies, and work processes. Kenya (GFS) and Mozambique (GFS) described it as adopting “best practices,” which is commonly understood to mean internationally adopted standards of statistical methodologies and dissemination practices. A number of countries emphasized the new working processes and procedures. Nigeria (national accounts) said “We have instituted transmission mechanisms to ensure that knowledge is transferred to other staff in the division.” And Nigeria (BOP) stated that “most of the processes being established will become routine and budget lines provided to fund them. Time lines for data releases have been established.” Kenya (national accounts) noted the fact that they have “developed quarterly monitoring systems” that will ensure the maintenance of the higher quality data achieved. Uganda (national accounts) mentioned new working methods as a main reason why results will be sustainable. Namibia (national accounts) noted the value of having the documentation (TA reports) to serve as a continuing reference especially for new staff. SACU emphasized the importance of the database the secretariat has established for the member countries that will enable maintenance on a regional level of the statistical harmonization achieved in the project.

42. The increased demand for the improved statistics by policy makers and accordingly by the management of the agencies producing the statistics was mentioned by Kenya (national accounts) and Rwanda (GFS) as an important factor ensuring sustainability.

43. Finally, a number of countries described the potential risks that may threaten sustainability. Most of these responses said that sustainability depends on continued resources to maintain and train staff (Tanzania, national accounts; Lesotho and Sierra Leone, monetary; and Ghana, BOP). The Gambia (national accounts) said that sustainability depended on continued resources for data collection, which implies provision of adequate staff for data collection. Ghana (BOP) made a similar point that sustainability depends on continued resources for the enterprise survey that has been adopted as a result of the project. Botswana (monetary statistics) emphasized the importance of the data providers’ understanding and awareness of the importance of providing the data to the central bank.

To what extent has your organization demonstrated commitment/ownership of the project (e.g., commitment of staff, resources, etc.)?

44. The most common responses stated that they have demonstrated commitment/ownership by committing the necessary staff to work on the project, in national accounts (The Gambia, Namibia, and Uganda), in monetary statistics (Namibia), and in GFS (Kenya and Rwanda). Some countries said that they actually increased their staff for the project, often responding to the recommendations of the experts, in national accounts (Kenya, Rwanda, Sierra Leone, and Tanzania), and in monetary statistics (Swaziland and Uganda). Many countries noted that they have used their own resources to support the training of their staff, in national accounts (Uganda and Zanzibar), in monetary statistics (Botswana, Kenya, Namibia, and Swaziland), in GFS (Mozambique), and in BOP (Nigeria).

45. In some cases, new departments, divisions, or work groups have been established that can effectively address the demands of the project. In Mozambique (GFS) the finance ministry has set up a new Department of Statistics. In Nigeria (BOP) the central bank has established the External Sector Statistics Division that has an earmarked budget for surveys. Rwanda (BOP) said that it has set up “focal persons appointed to work on GFS.” Kenya (GFS) stated that “a team of staff from all the relevant departments has been set up to implement the recommendations of the project.”

46. A few countries focused on their commitment of resources more generally to the project. Ghana (national accounts) said that its Economic Statistics Division now has the highest priority in the Ghana Statistical Service and it is well equipped to support the demands of the project in terms of resources and equipment. Nigeria (national accounts) likewise remarked that it has devoted substantial resources for the Macroeconomic Department. Ghana and The Gambia (BOP) noted their budget support for surveys, and Mozambique (BOP) said that the project has been included in the central bank’s action plan. Mauritius (BOP) noted that it is using its own resources to develop in-house software to enhance the efficiency of processing the survey data resulting from the project.

47. Regarding regional organizations, the EAC said it is following up on implementation of the project’s action plan by seeking funding for a one-year expert to perform the work. SACU has put the project activities in its work plan along with performance contracts.

48. One country in monetary statistics admitted that it has not yet demonstrated adequate commitment to the project, as it has not followed the recommendation to increase staff.

Other comments—Comment on any aspect of the project that you found particularly beneficial, or any aspect of the project that you think could be improved in the second half of the project.

49. Most did not provide any response to this opportunity to make additional comments. A few repeated suggestions they had made earlier in the questionnaire, such as requesting more training or workshops (especially in national accounts), or more follow up missions or workshops (EAC and SACU). A few took the opportunity to express appreciation for the individual expert who has been providing assistance during the first half of the project.

B. IMF Module Managers and Experts

50. Fifteen IMF module managers and experts completed the questionnaire or were interviewed by the project manager on the basis of the questionnaire. The responses cover all of the modules of the project. Their responses will be summarized in a way that provides a good overview of the thrust of all responses, but highlighting particular areas of emphasis or areas with implications or recommendations for adjustments in the second half of the project. In two cases below, two questions were combined because it was found that the responses overlapped considerably with each other.

Project on track?—Is module on schedule? For experts, is your country on schedule?

51. Responses indicated that all modules were broadly on track, and five said that they were ahead of schedule at the mid-point of the project. The modules ahead of schedule are the modules for QNA, FSIs, BOP, and data dissemination, and the GFS module for the part dealing with manuals and guides. In the first three of these, the achievements were related to more countries than expected beginning to produce and publish new statistics (QNA and FSI) or produce new source data (cross-border enterprise surveys for BOP). For manuals and guides, the GFS compilation guide for developing countries was published earlier than expected, and a draft for a second manual (BOP) is already completed. In data dissemination, more countries have begun to produce NSDPs and ARCs than expected, Burundi joined the GDDS, and Mauritius subscribed to the SDDS.

Main areas of progress—What have been the main areas of progress and/or lack of progress in your module/country? Please provide reasons and evidence for your response. Suggestions for changes or adjustments for remainder of project? Country by country review of experiences during the first half of the project in the module, areas where progress has been on track or better, and areas where we or the authorities need to adjust and change.6

52. For monetary statistics, the workshop and missions have assisted all module countries to set up new reporting systems to expand coverage and trained staff both in central banks and reporting institutions. Some countries have already begun to produce preliminary expanded data. No major adjustments are seen as necessary. For monetary harmonization, two workshops to the EAC have succeeded in agreeing on a methodological framework for the member countries to harmonize their monetary data. A similar workshop was conducted in 2012 to begin work on monetary harmonization for the SADC countries. For FSIs, three countries have begun to produce and publish FSIs (Mauritius, Seychelles, and Uganda), and assistance will continue to be provided to other countries in the second half of the project.

53. For BOP statistics, all six countries have begun to conduct cross-border enterprise surveys to provide source data for compiling IIP and improved BOP statistics. Three of the countries (Ghana, Mauritius, and Mozambique) have succeeded in setting up regular funding mechanisms for the surveys to be done on a regular basis, reinforcing the sustainability of this effort. The second half of the project will shift to focus more on integrating the survey data into IIP and BOP, to assist the remaining three countries to set up regular funding mechanisms to conduct the surveys, to address database issues (The Gambia and Nigeria), and to increase the frequency of the surveys from annual to quarterly. To better assist The Gambia and Nigeria with their database issues, the IMF has facilitated study visits with Ghana for both countries to learn from Ghana’s system. In one country, the authorities need to adjust to better enforce reporting requirements by parastatals and foreign trade zones. In another country, new leadership will require high level sensitization in order to take ownership of the EDDI objectives, which may be accomplished at the IMF/World Bank annual meetings.

54. For annual national accounts, Sierra Leone has been able to conduct both a household survey and an enterprise survey to provide much improved source data for national accounts, and in addition, TA under EDDI is assisting them to produce and publish for the first time a producer price index (PPI), and an improved CPI, both of which are expected to be released by the end of 2012. In the first part of EDDI, Liberia had serious problems in producing reliable source data for GDP. After several EDDI missions, it was decided that an adjustment was needed, and the IMF agreed to fund an intensive three-month mission to focus on resolving difficult issues in how to make use of a faulty survey to compile GDP estimates. The mission was successful and it is now expected that Liberia will soon begin to produce more reliable GDP estimates. The Gambia has experienced similar problems as Liberia, and it is possible an adjustment like the one conducted in Liberia will be done for The Gambia in the second half of the project. But new staff have been hired, and new source data from a VAT will be available in 2013, so an EDDI mission will then visit The Gambia and assess the situation with these new resources to decide on the best strategy.

55. For GFS, all countries have produced institutional tables, used new bridge tables to convert national data to GFS classifications, and produced preliminary data using these new systems. Mozambique and Zimbabwe have made the most progress, because they have received two missions. Second missions to the other countries (Burundi, Kenya, Malawi, and Rwanda) have been delayed because staff in the countries were busy with pressing work on their budgets. But missions to these four countries are expected now to take place in the next six months. The focus in the second half of the project will shift to expanding the basic GFS coverage to include extra budgetary accounts, social security, and local governments, along with the refining of GFS estimates.

56. For data dissemination, NSDPs and ARCs are ahead of schedule and the participation in GDDS objective is on schedule, with Burundi having joined. South Sudan has been substituted for Eritrea for the second country to assist to join before the end of the project, and a mission for this purpose is expected to visit South Sudan in March 2013. The SDDS part of this module was on track after Mauritius subscribed to SDDS, but it will soon be behind schedule as other prospective SDDS countries have yet to take the necessary actions, including agreeing to the transparency needed to take the last steps to subscribe to the SDDS.

57. For the SACU module, the second workshop narrowed its focus to only national accounts, the most critical area of statistics for the SACU countries to harmonize. A study is now planned that will provide better information to the SACU governments on the impact of statistical harmonization on their GDP statistics, which will help them decide whether they wish to participate in harmonization.

Effectiveness of project design—Has the overall design of the module and work plans proved to be a sound basis for achieving the expected results? Have any adjustments already been made, or is there a need for adjustments now? Has the mix of missions and workshops proved to be a good balance in terms of both effectiveness and efficiency?7

58. All responses confirmed that the basic module design was working well in EDDI, not surprisingly, since it was tried and proven in the previous GDDS Phase II project, and the staff and experts are well experienced and versed in how to deliver TA using this design. The topical modules (national accounts, BOP, monetary statistics, and GFS) all reiterated that the basic design of an opening workshop followed by a series of individual country missions has worked well, with additional workshops held during the project as deemed necessary for training in specific areas or to share experiences and lessons learned from other module countries. The modules relying on workshops as their main vehicle of TA delivery (monetary harmonization, FSI, data dissemination, and SACU) also said that design was overall working well.

59. There were, however, a number of minor adjustments made and some adjustments indicated as possible for the future. For example, in national accounts it was decided that Liberia needed more than the standard two-week missions or it was never going to make sufficient progress. The IMF therefore decided to fund a three-month expert to assist with obstacles to making good use of survey data to produce GDP estimates. Occasional more intensive attention may be needed in cases like Liberia. It was indicated that The Gambia may be another such case for the second half of the project in national accounts. In the BOP module, it was decided to facilitate a third TA vehicle, study visits, in order to provide special assistance to The Gambia and Nigeria. Arrangements were made for both countries to work with the Bank of Ghana to demonstrate to them how to develop a well-functioning database management system. A positive aspect of ownership in both cases was demonstrated as both The Gambia and Nigeria central banks agreed to fund these study visits. Both the FSI and the data dissemination modules have added another vehicle by increasing the use of follow-up by email in order to get more countries to use the training they received in workshops to produce FSI statistics and NSDPs and ARCs.

60. For QNA, a comment was made that greater high level sensitization at the beginning of the project might have helped develop greater ownership by some countries (Kenya and Tanzania). The BOP module had the same issue in the previous GDDS project, and as a result started the EDDI project with initial missions to each country for the purpose of getting high level buy in to the project, and it was only after these initial missions that the opening workshop was held to develop work plans. For any future projects, this lesson might be kept in mind as a possibility, to start modules with initial missions to obtain ownership and commitment at a high level. That said, the QNA module has turned out to be one of the most successful modules, so it is not clear that initial missions would have made the critical difference with Kenya and Tanzania.

Choice of module countries—Has the choice of countries for the module proved to be appropriate? Have there been any adjustments made, or is there a need to make adjustments now?

61. Comments generally confirmed that the choice of countries in modules has worked out well, although a few countries have been dropped (invariably for lack of interest and commitment) and other more enthusiastic countries added. For QNA, Rwanda requested permission to join and was accepted as the seventh country in that module. In GFS, Tanzania dropped out and Zimbabwe requested to join and has done well so far. In monetary statistics, Ethiopia dropped out. In data dissemination, South Sudan has joined and will be assisted to join GDDS in the second half of EDDI.

62. The choice of countries has worked out well because the basis of selection was sound. In all cases the countries originally expressed interest and exhibited need. In QNA and FSI, countries had to be relatively statistically advanced. For several modules, country selection was based on being located in the same region or belonging to the same regional organizations (annual national accounts, monetary harmonization, GFS, and SACU).

63. In the second half of the project, if a few countries reach their objectives before the end of the project, it is possible other countries will be able to join the particular module of their interest.

Choice of experts—Has the choice of experts and IMF staff for TA delivery worked well and smoothly? Any adjustments already made, or need to make any now? Reasons? Has the balance between IMF staff and expert involvement been appropriate and enough to monitor closely the quality of TA being provided?

64. Overall the choice of experts and IMF staff to work on the various modules has worked out well, as all experts were experienced, chosen from the IMF STA Panel of Experts and staff were senior staff and economists in the relevant topical divisions. Very few changes have been made so far, and those were for various reasons (transfers by staff, expert preferences, etc.). The balance between experts and staff has been good, with the typical situation being that the experts conduct the missions and staff backstop their work and monitor its quality and help keep them up to date with current methodology. The stability of experts working on particular countries is highly valued by the countries because they get to know the experts and trust them and work with them even by email after they establish a good rapport on missions. In the BOP module, there was some redundancy in backstopping by both the module manager and IMF staff in the first year of the project, but this was addressed successfully and is no longer a problem. Two modules (FSI and data dissemination) have used exclusively staff so far, but this has mainly been due to availability and interests of particular staff.

Administration of project—Has the overall administration of the project been accommodating to a smooth delivery of the necessary TA inputs? Suggestions for improvement?

65. Responses generally stated that the administration of the project has gone smoothly and they have experienced little or no problems. This is not surprising in view of the fact that the administrative system and procedures were not new to these respondents with the EDDI project, but had been well established with the previous DFID-financed project. Particularly commended was the flexibility in administration to changes needed during the project to adjust to the changing needs of countries and circumstances encountered. A couple of responses mentioned about the delays caused by the temporary budget tightness of the project in the beginning of its second year, but appreciated that these were able to be worked out without much delay thanks to DFID’s flexibility to bring forward a disbursement. It was also mentioned that at times experts were not informed in a timely way about changes in procedures, but this was addressed and there have not been any such issues recently.

Most difficult country issues—What are the main issues in countries with inadequate progress: lack of resources, commitment/ownership, adequate skills, security, staff turnover, other?

66. The two issues mentioned most frequently were inadequate human resources (both in terms of numbers of staff and level of skills) and lack of commitment/ownership. Mentioned almost as frequently was the lack of adequate funding for hiring staff, purchasing adequate equipment and software, and conducting surveys. Of course, funding and level of commitment and ownership are often closely related, although not necessarily as in the relatively poorer countries (e.g., The Gambia, Liberia, and Sierra Leone). Security delays were only mentioned for one country, for the BOP and QNA modules. Delays in agreeing to the scheduling of missions were mentioned for several GFS module countries, and political will was indicated as an important factor in the data dissemination and SACU modules.

67. It was apparent from the different responses that the reasons behind the cases where insufficient staff or effort were devoted to the project varied. For example, delays in agreeing to scheduled missions in several GFS module countries were related to competing work pressures that required the staff to be involved in high priority work such as budget preparation. In the West African countries mentioned above, commitment of the few staff working on the project has been shown to be very high and enthusiastic, often requesting more TA missions than the project can afford to send. These countries typically simply cannot afford to hire more staff.

68. It is often difficult to sort out the reasons why follow up missions find that the progress that was to have been achieved since the previous mission has fallen short. The reasons provided usually involve that the staff were forced to be involved with higher priority work, but in some cases the staff simply did not have the capacity and required more intensive hands on training and assistance than the EDDI project could provide. The national accounts module will address this issue by conducting a special training course for the three West African countries in early 2013, and may also request the IMF to fund a longer mission to The Gambia similar to the one sent to Liberia earlier this year.

69. In general, modules working with central banks (monetary statistics and BOP) tended to fare better in terms of adequate staff and funding for implementing the recommendations of the EDDI experts. Finance ministries and particularly statistical offices were generally less well funded and staffed.

70. Political will was mentioned by the data dissemination and SACU modules as being the most important obstacle to progress. In the data dissemination module, political will is a problem mostly with respect to transparency issues. The production and publication of NSDPs and ARCs require an increase in transparency regarding data and timeliness that the authorities may not be willing to commit to, although staff may also not feel confident about meeting deadlines. Similarly, fulfilling all the requirements to enable subscription to SDDS also requires greater transparency. Often the final requirements remaining for a country to fulfill relate to sensitive areas such as international reserves or offshore accounts. For the political authorities of the five SACU countries to all agree to harmonize their statistics will require them to accept the impact and modifications on their own national statistics, around which a good deal of uncertainty still exists.

Need for any IMF adjustments?—Are there any country issues that require IMF adjustments: adequacy of expert, project administration, design and realism of work plan, frequency of missions/workshops, flexibility, other?

71. There was a good deal of overlap in the responses to this question to the second question listed above. As already mentioned, the national accounts expert for Nigeria is being changed for the second half of the project. For the QNA module, new countries may be added if some of the existing countries in the module reach their objectives early. The annual national accounts module stressed the importance of continued flexibility of the project to the particular needs of the more under-resourced West African countries, especially with respect to the type of TA to be provided (e.g., longer missions, more training, etc.). The national accounts module and BOP modules also suggested continued flexibility to changes in objectives, for example to address the rebasing of national accounts in Nigeria and Namibia, and the serious BOP database issues in The Gambia and Nigeria. The GFS module mentioned the probable need for a training course in 2013. Some high level sensitization of certain countries was suggested by both the BOP and national accounts modules, which can possibly be accomplished in part at the IMF/World Bank annual and spring meetings.

External factors—Can you point to examples of external factors that have contributed to the success of the module in particular countries? (e.g., consultation/coordination with DFID, other donors, or regional organizations, change of government/ministers, etc.).

72. Several responses mentioned that collaboration and communication with other donors working on database issues in particular countries was very helpful. This type of collaboration is particularly productive, because the IMF’s work on compilation, methodology, and dissemination improvements complements very well other donors’ work on developing source data and database issues. Indeed, the total success of both types of TA could not be achieved one without the other. In national accounts, this kind of collaboration took place with the World Bank in The Gambia and Liberia, Statistics Denmark in Tanzania, and with UNDP in Rwanda. In addition, valuable communications on DFID/World Bank database work and other issues occurred with the DFID country advisors in Ghana and Nigeria on national accounts and BOP. The DFID advisor in Nigeria was very helpful in advising on the security situation and helping get EDDI back on track following security disruptions.

73. The collaboration between IMF headquarters TA and AFRITAC East TA in the QNA module has also been a critical factor in that module’s success so far. In the BOP module, consultations with the DFID/World Bank advisor in Kenya were helpful in addressing the issue of the funding and timing of surveys. Also in the BOP work on Kenya, collaborating with the Macroeconomic and Financial Management Institute (MEFMI) on database issues proved very useful. In the GFS module’s work in Zimbabwe, frequent contact with the British Crown Agents working on Zimbabwe’s computer system and database was helpful. The monetary statistics harmonization module is critically dependent on successful collaboration with two regional organizations, the EAC and SADC. Similarly, the SACU module’s progress will be facilitated by the SACU secretariat and its assistance in organizing workshops and following up with countries on progress. Collaboration with the new AFRITAC South will also assist work on SACU.

C. DFID Regional and Country Advisors

74. Questionnaires were sent to all DFID country and regional advisors in Africa. The response rate was 90 percent. All country advisors responded. One regional advisor did not respond, but the senior statistical advisor for the same region responded, ensuring a comprehensive set of feedback from all advisors in Africa in terms of coverage. The questions are listed below in italics followed by summaries of the responses and comments.

Communication—Are you kept well informed of EDDI activities in your country/countries by the DFID project manager and/or IMF experts? Have you had any contact with IMF experts or have knowledge of their work in your country? What additional information would you like to have?

75. Almost all responses (89 percent) said that the DFID project manager had kept them well informed of EDDI project activities. This is the agreement worked out between DFID and the IMF at the start of the project that the IMF project manager would funnel information through the DFID project manager, who would then distribute to the DFID advisors. Accordingly, only 22 percent of responses said that they were kept well informed by IMF experts, although 56 percent of responses said that they had contact or involvement with IMF experts or have knowledge of their work in their respective countries.

76. The only country advisor not kept well informed and who felt that more information was needed was in South Sudan. This response is understandable, as EDDI activities in newly independent South Sudan are only scheduled to begin in March 2013, when a GDDS mission is planned. This mission will do an overall assessment of macroeconomic statistics, help the authorities to prepare a comprehensive set of metadata, and agree on medium-term plans for improvement, including IMF TA. It is expected that the mission will pave the way for South Sudan to join the GDDS in the second half of 2013.

77. Of the countries not having had contact with IMF experts, a good part of this is explained by the fact that many EDDI missions work with the ministries of finance in GFS or with the central banks in monetary statistics or BOP statistics. These are usually not a focus of DFID regional and country advisors, who mostly work with the national statistical offices. The result is that almost all of the experts who have contacted DFID advisors in countries are interested in national accounts, a subset of the activities of the total EDDI project. Nevertheless, the responses indicated some interest by DFID advisors to begin to have greater contact with central banks and ministries of finance, realizing that they are important parts of the overall statistical systems. The IMF and DFID project managers will work together to encourage more contact with IMF experts working in ministries of finance and central banks where there is interest.

Relevance—In your view has the EDDI project been of relevance to the country/countries you deal with, either within the national statistical offices, central banks, finance ministries, or the wider statistical system?

78. All responses affirmed the relevance of the EDDI project, consistent with the strong ratings given to relevance by country counterpart responses. The advisor in Kenya pointed to several EDDI activities that supported national objectives, including improving BOP statistics, developing a module for maintaining an effective and up-to-date business register, and the regional objective for harmonizing monetary statistics among the EAC countries. The advisors in both Ghana and Rwanda noted the importance of EDDI activities to the success in publishing QNA for the first time.

79. Some concern was expressed about the difficulty of maintaining relevance and high priority in countries during the time in between missions when countries are easily distracted by other non-EDDI activities undertaken by the national statistical offices. One suggestion was to try to use the local IMF representative to sustain awareness and relevance during these times. This suggestion is consistent with suggestions made by feedback from IMF experts, who suggested in addition to promoting high level awareness and commitment on occasions like the IMF/World Bank annual and spring meetings. On the other hand, the module approach has worked well in many countries, where they have used the time in between missions productively to implement the recommendations and achieve results. Too much “hand holding” by experts who are in country as residents or on extended missions can result in the other problem of the authorities never learning how to compile the statistics on their own, always expecting the experts to do it for them, which is not a sustainable result.

Concerns and suggestions—Do you have any concerns with the program or its implementation? Have you discussed these concerns with Global Statistics Partnership Team and have they been addressed?

80. Some concern was expressed that missions seem “too short to get real traction,” and momentum can be lost if there is no sustained follow up. As mentioned above, our experience with the module approach in EDDI and the previous GDDS project has been that the series of sustained standard two-week mission’s work well in many countries, with countries implementing the last mission’s recommendations before the next mission is sent. Normally the interval between missions is between 6 and 9 months, depending on how quickly the countries implement the recommendations. However, in cases of the more under-resourced developed countries it may be necessary to lengthen missions from time to time or provide additional training, a point that was also made by IMF expert feedback. The IMF is already following up on some of these cases with additional training and occasional longer missions.

81. Mention was also made of the lack of complementary funding to TA being a constraint (e.g., funding for surveys and database development). The IMF does not have the capacity for this type of funding, and therefore the EDDI budget does not provide this kind of support. It must depend on the IMF working with donors who can provide complementary funding, such as the World Bank. In fact, in EDDI and the previous GDDS project there have been numerous examples of this kind of funding, such as World Bank and the West African Institute for Financial and Economic Management support for complementary surveys in several West African countries, MEFMI database/software support in Kenya for BOP, and the British Crown Agents database/software support in Zimbabwe. The IMF has worked closely with all of these. Nevertheless, doubtless more opportunities for this type of collaboration exist and this provides further support for the argument for closer communication with other donors.

82. More communication by IMF experts while on mission with the wider donor community was suggested specifically as a recommendation. The IMF will continue to make efforts to encourage its experts to visit the DFID country advisors while on mission along with the World Bank and other donors when it is known that they are engaged in related work. DFID country advisors can contribute to this effort by including specific donors in meetings with the IMF expert when they are known to be engaged in related work. Indeed, the DFID country advisor in Ghana organized a meeting with the World Bank team for a visiting IMF expert in February 2013.

Quality of advice—How would you rate the quality of advice that IMF experts have provided to the relevant organizations in the countries you deal with?

83. All of the DFID advisor responses rated the quality of advice by IMF experts as high.8 Experts were described as “subject experts in their fields with years of practical experience,” “one of the best internationally,” and as having “a very good technical understanding and a valuable flexibility.” The advice was also described as “pragmatic recognizing the constraints of data collection and availability in many African countries.”

Timeliness of advice—In your view, have the IMF provided timely advice to the relevant organizations in the countries you deal with?

84. Most of the responses (80 percent) said that the advice was provided in a timely way. The one negative response appears to have been a partial no and recognized that this could partly be due to the national statistics office’s planning and partly due to the IMF being unable to field support quickly for large jobs. It is true that the IMF requires a minimum amount of time to arrange for a mission to take place. If the expert is available (many are in demand for other countries and have busy schedules), a minimum lag of two months is possible, allowing time for the hiring and contracting of the expert and travel arrangements to be made. However, most EDDI missions are scheduled as part of a regular series of missions that both the countries and experts can plan for. It is the unexpected scheduling or disruptions (such as the security disruption prohibiting IMF missions for a lengthy period in one country in the second year of EDDI) that cause problems.

Impact—What difference, positive or negative, has the project made to the relevant organizations in your countries? Please give examples where possible.

85. Responses generally confirmed that EDDI was having a positive impact in their countries. Examples cited the improved timeliness of BOP statistics in Kenya, improved quality, and perception of quality by users, of national accounts in Rwanda, and publication of QNA for the first time in Ghana. One response cautioned that occasionally officials in hierarchical organizations can feel their authority or knowledge questioned by external experts and therefore resist it. In such cases, it would be very helpful if the relevant DFID advisor, who is more intimately aware of these kinds of circumstances, could discuss with the IMF expert and provide advice about the most effective and pragmatic course of action. Indeed, DFID country advisors have provided this kind of helpful advice to IMF experts during the first half of EDDI in a number of countries, where personal and institutional peculiarities may not be so evident to an external advisor visiting only for brief periods.

Other comments—Please comment on any aspect of the project that you found particularly beneficial, or any aspect of the project that you think could be improved in the second half of the project.

86. Two of the four responses repeated the suggestion about better communication by IMF experts with local donors and organizations engaged in related work (e.g., DFID and World Bank). Again, the IMF and DFID will work together in the future to improve coordination and communication. The suggestion about providing “complementary funding” was also repeated. As stated above, this is not part of the IMF “tool box,” but improved communication with donors will hopefully identify more opportunities in the donor community for funding activities complementary to EDDI. One point should be mentioned here, that the best source of complementary funding is the host government itself, which is the preferred way to ensure sustainability. And in this area, EDDI has done quite well in promoting the inclusion of activities complementary to the project to be included in the annual budgets of host countries, as the many examples provided by country counterparts’ feedback attest (e.g., BOP surveys, hiring and training of additional staff, and software development).

87. One suggestion was for the IMF to consider having in-country IMF advisors. If a country thinks that it needs an in-country IMF advisor, it should make the formal request either through its executive director on the IMF Board of Directors or directly to the Director of the IMF Statistics Department. The IMF may fund a limited number of in-country advisors each year depending on the availability of funds. More intensive TA can also be provided through the IMF regional TA centers. The AFRITAC East and AFRITAC South are already operating and providing assistance to many EDDI countries, and a new AFRITAC West is being planned and organized to begin in the coming year or two.

IV. Recommendations

  • More frequent missions to countries especially in need;

  • More lengthy missions (e.g., 3–4 weeks instead of 2) to countries especially in need;

  • More workshops or training courses to increase capacity to absorb TA (national accounts and GFS);

  • Timing missions when new data are ready;

  • Extend EDDI’s TA and training beyond the main government agencies to include data reporting institutions;

  • Sensitize high-level management of the main statistical agencies (statistical offices, central banks, and finance ministries) of the importance of continued funding for implementation of the project. This can be done by IMF African Department missions to countries, high-level meetings in Washington at the annual and spring meetings, and communications by senior IMF Statistics Department staff as needed with particular countries; and

  • More IMF experts will meet with DFID country advisors while on missions, and with other donors that are known to be engaged in work closely related to EDDI. DFID advisors will where possible assist in organizing meeting with significant other donors/stakeholders.

Mid-Term Evaluation Report on the Enhanced Data Dissemination Initiative Project
Author: International Monetary Fund