1. Overview

Thomas Morrison
Published Date:
November 2005
  • ShareShare
Show Summary Details

Thomas K. Morrison*

Purpose of Studies

1.1 The importance of statistical information in the public arena is well established. Increasingly in the past decade, events and developments have propelled efforts to improve the quality of statistics. The financial crises of the 1990s increased attention to the timeliness and reliability of statistics. In addition, a heightened emphasis worldwide on good governance, transparency, and accountability underpinned the drive to improve data quality.

1.2 In this context, statistical capacity building became a focus of governments of developing countries and of donors that provide technical assistance (TA) in support of this effort. Growing out of crisis prevention and an emphasis on good governance and transparency, demands arose from a broad range of countries for TA. Demands also came from particular groups of countries and regions—for example, the transition countries and a growing list of postconflict countries. Both these groups required fundamental reforms and rebuilding of the statistical infrastructure.

1.3 Responding to these needs, developing country governments and TA providers alike focused scarce resources on increasing delivery of TA and implementing plans for improving statistical capacity. Over the past decade, TA in statistics provided by the International Monetary Fund (IMF), for example, increased more than fourfold.

1.4 Sufficient experience has now been gained to permit stepping back and examining the effectiveness and impact of capacity-building efforts and attempting to learn lessons about how to improve these efforts. The country case studies in this report are based on the experiences of the four authors providing IMF TA. Both Mr. Abbasi’s and Mr. Atcherley’s experiences have been as resident long-term multisector statistics advisors, in Cambodia and Bosnia and Herzegovina (BiH), respectively. Both Mr. Kučera’s and Mr. Slack’s experiences, on the other hand, have been as monetary statistics experts involved in sustained medium-term programs of short-term missions to the Ukraine and BiH, respectively.

Scope and Structure

1.5 The case studies were selected to cover different TA delivery mechanisms, sectors targeted, and country situations. In the cases of Cambodia (see Chapter 2, Abbasi) and BiH (see Chapter 3, Atcherley), the mechanism was IMF multisector TA, focusing on the statistical system as a whole and including statistical organization, the main macroeconomic statistics sectors (balance of payments, monetary, government finance, national accounts, and price statistics), and selected sociodemographic statistics. In the cases of BiH (see Chapter 4, Slack) and Ukraine (see Chapter 5, Kučera), on the other hand, the mechanism was more narrowly focused single-sector (i.e., monetary statistics) TA, delivered by a series of short-term missions over a medium-term period.

1.6 All three countries are transition countries, having transformed over the past decade from planned to market economies, but in the different regional contexts of Eastern Europe, the former Soviet Union, and Asia. The two case studies on BiH provide the opportunity to examine two different approaches to TA in the same country. In addition, while the authors’ experiences originate from their IMF assignments, the authors endeavor to recognize the capacity-building efforts in the countries in a broader context, including not only IMF activities but also those of the country authorities and other TA providers.

1.7 The four case studies are similarly structured, covering similar medium-term time frames in the late 1990s and early 2000s. They generally begin by describing the setting, giving a snapshot of the country’s starting point in terms of general background, institutional setting, capacity, and statistical outputs. They then review the medium-term strategies and plans for improvement that guided the capacity-building activities. Following that is a description of the implementation and outcomes, focusing on concrete outputs and quantitative indicators. Obstacles and shortcomings in implementation are also noted and explained. The closing sections present conclusions and lessons learned. They emphasize factors that contributed to achieving objectives and, in cases where implementation fell short, discuss the reasons and identify lessons learned.


1.8 The four studies describe the settings in the three countries, first in terms of general sociopolitical and economic background—thereby providing the general context in which capacity-building efforts were launched—and second in terms of the status of statistical capacity building. In the first context, as noted earlier, all three countries were transition countries, although they were moving from different centrally planned systems and also from different types and levels of sociopolitical and economic development. Ukraine was moving from the former Soviet Union model, BiH from the Yugoslav model, and Cambodia from an Asian version of central planning. Ukraine had the advantage that it was not postconflict, compared with both BiH and Cambodia, where basic infrastructure had been destroyed and had to be rebuilt. As a result, poverty and a lack of resources were relatively greater in BiH and Cambodia than in Ukraine. All three countries benefited from large amounts of foreign assistance—which was most pervasive in BiH, where the Dayton Agreement set up a complex and divided political structure to rule the country. Ukraine and Cambodia were fortunate in experiencing their transitions with relative political stability and unity.

1.9 In terms of statistical capacity, all three countries began their transition needing fundamental reforms of their statistical systems, formerly based on centrally planned models. Interestingly, decentralization itself created serious problems to be overcome in setting up new systems, as traditional sources of data collection disappeared and the dispersion of responsibilities caused confusion and problems of coordination. Staff in statistical agencies needed basic training in the new methodologies. New institutions had to be created or rebuilt in the two postconflict cases. In all cases, the situations in the beginning were better for statistics produced by the central banks (e.g., monetary and balance of payments statistics), where relatively strong leadership and resources provided better bases on which to build. Nevertheless, in all country cases, statistical outputs at the beginning of the period were fairly limited. Statistical legislation was also lacking in all three countries.

Strategies and Plans

1.10 The authors discuss strategies for statistical capacity building and TA either for an entire statistical system (Chapters 2 and 3, Abbasi and Atcherley) or for a sector or project (Chapters 4 and 5, Slack and Kučera).

1.11 Not surprisingly, the difficulties described in the case studies dealing with systemwide strategies are greater than those in the studies dealing with sector-specific strategies. In both Cambodia and BiH, Abbasi and Atcherley describe a lack of initial overall strategies to guide the capacity-building efforts of the authorities and the donors. As a result, progress in the early years in both cases was seriously hampered by gaps in coordination and coverage, confused and overlapping responsibilities, and a lack of effective leadership.

1.12 Coordinated strategic planning in Cambodia improved over time as authorities and donors gained experience, but coordination in BiH has been more uneven. Cambodia made greater progress, because it was not hindered by the divided political structure experienced by BiH. A critical event in Cambodia was the convening by the authorities in October 2002 of a workshop, “Partnerships in Statistics Capacity Building for Cambodia,” where a Statistics Master Plan for Cambodia was agreed upon by the various statistical agencies of the government and the donor community. This workshop probably would not have been possible had it not been for (i) the agreement of the major donors in 2001 to finance jointly and work together on a medium-term Technical Cooperation Action Plan (TCAP) that paved the way for closer donor cooperation and (ii) Cambodia’s decision in 2001 to participate in the IMF’s General Data Dissemination System (GDDS), which provided a comprehensive planning framework. Even in the more difficult situation of BiH, in Chapter 3 Atcherley describes four stages of the evolution of statistical strategy, leading finally to a situation where statistical legislation is in place and an empowered state statistical office is ready to embark on more prioritized strategic planning.

1.13 At the sector- and project-specific level in Chapters 4 and 5, Slack and Kučera point out that strategies were agreed upon fairly early by the authorities and the IMF to improve monetary statistics over the medium term in Ukraine and BiH. However, it became more complicated than simply using a kind of generic blueprint or “boilerplate” approach. Slack describes how flexibility and a sensitivity to logistical and pragmatic constraints were needed in deciding an appropriate phasing of TA delivery. The development of a unified strategy in both cases was simplified because the IMF was the major donor assisting with monetary statistics, and relatively little coordination with other donors was necessary. Although the IMF had developed an analytical framework specifically for improving the monetary statistics of the former Soviet Union, Kučera explains that this framework had to be adjusted to the particular circumstances in Ukraine, given that it takes time to change the culture of institutions.


1.14 The discussions of implementation in the four case studies demonstrate clearly the importance of having a well-planned and coordinated strategy from the start.

1.15 In the two cases of TA focused only on monetary statistics, clear initial strategies and definite objectives led to a logical progression of outputs. While some unexpected setbacks and delays inevitably occurred, the existence of a “road map” served to keep the projects broadly on track. In both cases, the principal mode of TA delivery was a sustained series (over several years) of short-term missions, supplemented by training courses. The missions, staffed by the same experts to the extent possible, could effectively evaluate progress since the last mission, advise about appropriate next steps, and then leave the officials to manage the implementation largely on their own during the intervals between missions. This served to promote “ownership” of the capacity-building efforts and to better ensure their sustainability. Examples of some key actions that facilitated implementation were (i) the formation of a working group in Ukraine to improve internal coordination within the central bank and (ii) an effort in BiH to bring the commercial banks more effectively into the process through seminars explaining why the reforms were necessary and through actions to reduce their reporting burden.

1.16 On the other hand, in the two cases where TA focused on the statistical system as a whole, implementation suffered from a lack of initial overall strategy. This problem was most serious in BiH where the fragmented political structure and a huge but uncoordinated foreign donor presence resulted generally (except for the financial sector) in a number of uncoordinated and sometimes inconsistent strategies, despite the expenditure of substantial resources. In addition, the scarce staff resources of BiH were stretched thin by the demands of so many projects and were spread over many activities, some less critical than others. Another serious problem was the phasing of inputs, exemplified by the lack of resources provided to develop source data.

1.17 These problems had the greatest impact on the BiH real sector statistics, whereas financial sector statistics (i.e., balance of payments, monetary, and government finance statistics) benefited from the fact that a single donor (i.e., the IMF) could take the lead in TA and agree with one national institution (i.e., the central bank) on unified and coordinated strategies. One telling example of the difference strong unified leadership can make was the personal intervention of the BiH Office of the High Representative (OHR), with strong support from the governor of the central bank, in 2002 to impose the passage of statistical legislation, thereby strengthening the national statistical office.

1.18 Cambodia suffered as well from a lack of overall strategy in the early years, but both the donors and the authorities recognized and began to deal with these problems at a relatively early stage. Eventually, with the critical support of “champions” for statistics (see Chapter 2, Abbasi), a coordinated and unified strategy was developed. A number of factors served to highlight statistics in Cambodia and stimulate the demand for better statistics by policymakers. Statistics were given prominence in the Millennium Development Goals (MDGs), the National Poverty Reduction Strategy (NPRS), and the second Socioeconomic Development Plan (SEDP).

1.19 Nevertheless, an important problem of phasing in Cambodia—the failure for so long to enact statistical legislation despite the completion of many drafts and preliminary steps—hindered statistical capacity-building efforts. This lack of legislation adversely affected the implementation of TA, undermining the authority of statistical agencies and leading to problems in response rates and registration by businesses.

Lessons Learned

1.20 A main purpose of this project was to see whether a series of country case studies on statistical capacity building, conducted within a single framework, would yield some common threads and patterns of conclusions and lessons learned. If so, the lessons involved could be more generally applicable than if they were associated only with an individual case study. This is not to say that the findings of each case study are not rich enough by themselves to provide provocative lessons for TA practitioners. Indeed, much of the value of the individual case studies is the flavor and the texture of the lessons presented as real-world experience.

1.21 When the case studies are examined together, they do indeed provide common lessons. The most common are listed below under four headings: (i) lessons for country authorities, (ii) lessons for TA donors, (iii) lessons regarding prerequisites, and (iv) lessons for transition and postconflict countries. These categories may be somewhat artificial (for example, the lessons for authorities and for donors are not mutually exclusive), but they serve to focus attention on where the lessons may be the most relevant.

Lessons for Country Authorities

  • The importance of high-level and unified leadership in the government to push statistical capacity building was emphasized especially in the multi-sector cases of BiH and Cambodia (see Abbasi’s focus on “champions” in Chapter 2), but also in the single-sector cases (i.e., leadership within the central banks) of BiH and Ukraine. Unified leadership in BiH and Cambodia was complicated by inadequate attention to the impact of decentralization on the statistical system. In BiH the fragmented political structure presented serious challenges, whereas in Cambodia high-level champions emerged to promote statistical capacity building. Some examples of successful leadership in BiH are noted, however, such as the strong efforts by the governor of the central bank to take responsibility for and improve financial statistics, and the pushing through of statistical legislation in 2002.
  • Taking ownership of statistical capacity-building activities and reforms was key in all cases, particularly with respect to sustaining the reforms. Interrelated with the leadership lesson above, this lesson also bespeaks the willingness of authorities to provide the necessary staff and budgetary resources to commit to the process. The Ukrainian authorities were relatively more successful in taking ownership, whereas in BiH the largely uncoordinated manner in which TA was delivered by donors discouraged ownership in the early years, and in Cambodia progress has continued to be hindered by inadequate budgetary resources.
  • Taking charge of internal coordination of statistical activities in the government was critical. In Ukraine, for example, internal coordination within the central bank was a serious obstacle until an interdepartmental working group coordinated the compiling of monetary statistics within the bank. On the other hand, in BiH, internal coordination driven by the governor was a positive feature in the central bank’s efforts to improve statistics.
  • Giving responsibility for donor coordination to a single government institution was important in the two multisector cases. In BiH, donors improved effective coordination gradually but never fully achieved it, whereas in Cambodia high-level support of a forum for donor coordination could serve as a model for other countries. If the authorities do not take the lead in this area, experience shows that donors by themselves find it difficult to coordinate their respective activities. The identification of responsibility for donor coordination and the related responsibility for producing an overall strategy and work plans should be part of statistical legislation.
  • Attention to user consultation was found to raise the profile of statistics by fostering demand for statistics and making products more relevant. The case study on Cambodia shows how the government used seminars and other means to consult with users.

Lessons for Technical Assistance Donors

  • Regarding donor coordination, donors should engage in projects in accordance with the strategy and priorities of the host country, and not propose projects that are tied to their own strategic interests and priorities if these are at odds with the priorities of the host country. If the host country does not have a clear overall strategy, then donors should do the best they can to coordinate among themselves. BiH (See Chapter 3, Atcherley) has been an example of a host country in which donors have at times not allocated project resources in accordance with the country’s priority needs.
  • Donors should pay attention to the appropriate phasing of project inputs in view of the absorptive capacity of the country and the most effective sequence of inputs. The logical phasing for improving monetary statistics in BiH provided in Table 4.2 (see Chapter 4, Slack) is a good example of the effective phasing of project inputs. BiH (see Chapter 3, Atcherley), on the other hand, showed how too much attention to technical training initially at the expense of developing source data can prevent significant progress in improving national accounts.
  • It is important to complement strategies and recommendations with hands-on TA and to work with the authorities to show them how to implement recommendations in the day-to-day, nitty-gritty activities. The effectiveness of the hands-on approach is particularly noted in Cambodia and BiH, which provide examples of how important it is for experts to be willing to work with the authorities on the mundane aspects of improving statistics, and not only on the big picture issues (see Chapter 2, Abbasi, and Chapter 4, Slack). This hands-on approach is facilitated where donors use resident long-term advisors rather than only short-term missions. Similarly, it is also critical that donors focus on direct data production activities and not solely on supporting activities such as seminars, study tours, and provision of office space, important though they may be (see Chapter 3, Atcherley).
  • Closely related to the last point is that donors should be careful in their choice of experts. Several examples are noted in Cambodia (Chapter 2, Abbasi) and BiH (Chapter 3, Atcherley) where the effectiveness of projects was compromised by experts who were not adequately qualified.
  • The choice of mode of delivery by donors (e.g., long-term advisor, short-term mission, series of short-term missions, training, workshops) should match the particular needs of the country’s situation. The two cases of TA in monetary statistics (Chapters 4 and 5, Slack and Kučera) are good examples of how a sustained series of short-term missions matched well the need to provide patient medium-term assistance that allowed for the difficult process of culture change to take place and learning/technology transfer to occur. Resident advisors can complement short-term missions in countries requiring intensive sustained TA (which is especially the case in transition and postconflict countries).
  • The ultimate success or failure of a project will depend on its sustainability and on whether the donors have designed an exit strategy. This is an important complement to the ownership lesson listed above under “Lessons for Country Authorities.” Again, the Slack and Kučera case studies (Chapters 4 and 5) are good examples of how the TA was geared consistently toward the objective of eventual total transfer to the authorities.

Lessons Regarding Prerequisites

  • Many examples are provided, particularly in Chapter 2 (Abbasi) and Chapter 3 (Atcherley), of how important statistical legislation is for progress in improving statistics in a variety of areas, such as coordination, adequate authority of the statistical agency, clear designation of responsibilities, and professional autonomy. Nevertheless, in BiH (Chapter 4, Slack) a culture of data reporting and goodwill among data providers and compilers was an effective short-term substitute for statistical legislation.
  • Chapter 2 (Abbasi) and Chapter 3 (Atcherley) also provide many examples of how critical adequate resources are, both in terms of staff and budgetary resources, for achieving objectives in capacity building. This lesson is closely related to the ownership lesson for country authorities, as it is essential in cases like Cambodia and BiH, where large donor assistance is provided, that the country authorities gradually take over more and more of the ownership and financing for statistical capacity building. Resource shortages, however, need not prevent the production of at least limited yet reliable statistics in the near term. In BiH (see Chapter 4, Slack), careful staff selection, planning of workloads, and dedicated local officials compensated for initial shortages in staff numbers and technology.

Lessons for Transition and Postconflict Countries

These lessons are not exclusively for transition and postconflict countries, but they are of particular importance for these cases (and also perhaps for a number of very poor, resource-starved countries).

  • To be effective, the scale of assistance has to be significantly larger than in typical cases of TA, because fundamental changes in methodology and culture are often involved, as well as the basic building or rebuilding of statistical infrastructure.
  • Donors must be extremely flexible in their approach and must adapt their strategies and modes of delivery to unusually difficult circumstances and low absorptive capacity.
  • Both donors and country authorities should be patient and prepared for sustained efforts over a multiyear period, and should be focused consistently on eventual total ownership by the authorities and on donor exit strategies.
  • Institutional and statistical organization issues are critical for these countries. For example, it was shown for both Cambodia and BiH that a decentralized statistical system seriously challenges post-conflict countries engaged in fundamental reform of the statistical system. A strong central statistical office can be a great help in coordinating and implementing large statistical improvement programs.

The author would like to take this opportunity to thank Carol S. Carson, who was director of the IMF Statistics Department when she approved my one-year sabbatical in 2003, giving me the time required to organize this project. I would also like to express my appreciation to the four authors of the case studies, who did not have the luxury of a sabbatical but somehow managed to carve out the necessary time from their busy schedules. Joan Gibson provided valuable editorial assistance. Thanks also to the various country officials who provided valuable comments. The views expressed in this paper are those of the authors and myself and do not necessarily reflect those of the IMF.

    Other Resources Citing This Publication