NUMBER 95 * ED Precis Operations Evaluation Department September 1995 Designing Technical Assistance Projects: Lessons from Ghana and Uganda Technical assistance (TA) success stories political turmoil and economic de- complement the reform efforts. But are hard to find, especially in countries cline. By the early- to mid-1980s, the with Uganda still suffering from the emerging from long periods of internal economies of both countries had col- effects of years of civil strife, condi- strife. Ironically, these are the countries lapsed, the countries' physical infra- tions at the start of TA III were far most in need of technical assistance, yet structures were in disarray, and the worse than those in Ghana. least able to use it effectively. Years of institutions and government services political turmoil have weakened their were nonfunctioning. Within a few Project goals public administration, leaving them with years of each other, both countries little institutional capacity to absorb experienced a change in government. Ghana's $10.8 million SAIS credit such assistance. Often they are weighed In both cases, the new governments, project was approved by the Bank in down by an institutional culture steeped emboldened by a broad consensus 1987 and closed in 1993. Uganda's in patronage and rent-seeking, creating for change, initiated far-reaching eco- $18 million TA III project, approved an environment ill suited to viable insti- nomic reforms. in 1988 and closed in 1994, was one tutional development. of a series of five free-standing TA In 1983, the Ghanaian government operations mounted by the Bank be- Experience from World Bank-sup- launched the Economic Recovery tween 1980 and 1995. But only in the ported TA projects in two African coun- Program (ERP). Soon after, both second half of TA II (approved in tries-one a relative success in Uganda the government and the Bank be- 1983 and closed in 1992) did the in- and the other a relative failure in came convinced that Ghana's weak stitutional environment become con- Ghana-bring to light the essential re- economic management institutions, ducive to reform. quirements for designing and imple- insufficient policy analysis capabili- menting successful TA projects in ties, and shortages of qualified ex- The central objectives of the SAIS countries with weak or nonfunctioning perts were serious constraints on the project in Ghana and TA III in government infrastructures. The lessons, government's capacity to implement Uganda were the same: (1) to help captured in two OED audits,*focus on the reform program. The Structural ministries of finance and planning four points: (1) the importance of an en- Adjustment Institutional Support and, in Uganda, the central bank de- abling institutional environment condu- (SAIS) Project was initiated in an ef- cive to technical assistance; (2) borrower fort to reform public sector manage- commitment, strengthened by the sup- ment in support of the ERP. The Pertrmnantce atudil reporls. port of a powerful decision-maker who is project was designed to complement 'Ghi7na: Structural 4djustment willing to become the project's advocate; the Bank's first structural adjustment lnstitiutionial Support Proiecl," (3) a flexible "process" project design credit to Ghana. Report No. 13262, June 1994, and tempered by strong, disciplined project 'Lg'anrda: Secol7d and Third Tech- management; and (4) careful Bank su- Real change came to Uganda in nical Assistance Projects." Report pervision by staff with substantial field 1986, with the new government of No. 14706, June 1995. Perfor- and country experience. Yoweri Museveni. After a brief at- mnance auidit reports are avnailable tempt at a statist approach, the gov- to Batik executive director s anad Background ermnent in 1987 reversed course, slaff front thte Inlerntal Documients launching reforms to stabilize and Unii and froml Regional Ilnforma- Both Ghana and Uganda had suf- liberalize the economy. Technical lion Services Centers. fered through almost two decades of Assistance (TA) III was initiated to Box 1: Major project goals Policy unit nomic Planning to manage public * Ugiantda: Help the NMinistry investment programming, budgeting, of Public Service and Cabinet * Ghana: Create a policv advisory expenditure forecasting and control, Affairs prepare and implement a unit in the Ministry of Finance and revenue collection, debt management, comprehensive civil service reform Economic Planning (MFEP); create and aid coordination. program. an economic liaison unit in the office * Uganda: Improve the Ministry of of the head of state. Finance's budgetary management and Reseiirc/i anid ado,sory ier.-icesf * Uganda: Strengthen the Economic expenditure control; restore govern- statistics ca;pacaty biilding Advisory Unit established under TA ment accounts; improve Bank of IT; strengthen policy analysis in the Uganda's external debt and financial * Ghana: Carry out studies and Agricultural Secretariat under the management. pro, ide advisory services in sup- Bank of Uganda. (The UN Food and port of critical structural adjust- Agriculture Organization executed Civil service refornm ment tasks, where possible by TA III's component to strengthen mobilizing skilled Ghanaians trom agricultural policy analysis in the * Ghana: Execute staffing, functional, outside the public service. (This Agricultural Secretariat.) and salary reviews of the civil service; component became known as the establish a sound administrative and Skills Mobilization Scheme.) Ca rry Financial management personnel management system; de- out a living standards measure- sign a program and framework for ment survey. * Ghana: Strengthen the capacity of cofinancing redeployment activities. * ULgan,da: Strengthen the capacitv the Ministrv of Finance and Eco- (See Box 2.) of the Statistics Department. sign and implement a progression of son unit. Both components were soon project completion, nearly 70 percent economic reforms (short-term goal) abandoned. Moreover, attempts to of project funds had been spent on and (2) to strengthen the institutional strengthen the economic policy unit equipment compared with the 31 per- systems and capacities of the agen- proceeded only haltingly. Monies cent planned at appraisal. cies themselves (long-term institu- "freed up" from those components tional development goal). (See Box 1.) were largely diverted to low-priority In contrast, TA III's contribution to uses such as purchase of vehicles. At Uganda's economic reforms and long- Findings OED found the objectives of both projects to have been highly relevant Box 2: Highlights of ODA's 1993 evaluation of Ghana's and timely. Moreover, the projects SAIS civil service component were designed and implemented in countries with similar backgrounds The UK Overseas Development freeze on recruitment waas i neffec- and in support of similar economic Administration CODA) fma nced and tive. The substantial gros red uction reforms. But Ghana's SAIS project, provided the aid management for in the civil ser vice *worktorce despite the country's better environ- the civil service component of the (invok ing compensation pay- ment, largely failed to achieve many SAIS project. Its major find ings are ments) was largely offset by newv of its important objectives, whereas as follows: recruitment. Uganda's TA III project largely suc- * Sele,ted le4sns: (nI) A more stra te- ceeded. The reasons for the different * Overall conclusion. The project was gic and long-term approach is project outcomes provide the main partially successful. Mlany compo- needed for civil service reform lessons of the OED audits. nents took considerablv longer than t2l Conditionality should cover the planned and many reforms have vet overall size of the public sector and Outcomes to be implemented. w%age bill, as well as recruitment. O Ownership: Reforms were essen- (3) Project design needs to take into tially externally driven and lacked consideration the social dimensions Although SAIS had some success the full com mitment of senior mem- of Civil service reform (traditional in reforming Ghana's National Rev- bers of the Ghanaian civil service. authoritv structures, famillt loyalties. enue Service (see Box 3), it was un- Project design gave little consider- gender). C(4 Emphasis needs to be on able to get its major institutional ation to Ghana's history of the civil implementation and ownership, development components off the service and the country's political rather than on the proviiion oi ground. It failed, for example, to and cultural context. consultancy studies. And (5l precise. implement the Skills Mobilization * Gross ani net reiretrchment: A quantifiable targets are needed. Scheme or to create the economic liai-_I OED Precis term institutional development was Approach substantial. Although the project Box 3: Ghana's National made little progress in strengthening In both Ghana and Uganda, Revenue Service: a reform the Ministry of Finance's budgetary project designers realized that they enclave? and financial management system, it could not predict the form and tim- succeeded in most of its other institu- ing of every contingency that might One of the more successful tional development efforts. Among arise. They thus rightly adopted a components ot the SAIS project its achievements: flexible project design that incorpo- was its su pport of the National rated both "blueprint" and "process" Revenue Service (N RS), the Helping the central bank, whose elements, with emphasis on the latter agency responsible for customs accounting and debt management (see Box 4). Organizational changes. including systems had collapsed, reestablish imation andiwr both, including developing an inven- As a result of the process ap- conditions, led to notable im- tory of Uganda's debt. An immediate proach, several components in both provements in efficiency within benefit of this effort was enabling SAIS and TA III were left under- de- NRS. Several factors explain the Uganda to initiate Africa's first debt signed at appraisal. In TA III, for ex- component's success: buy-back arrangement. ample, a sizable component was devoted to "special advisory services * Good management-an out- * Assisting the Economic Analysis for the economic reform program." standing leader at the top. Unit (EAU), created under TA II, to This component allowed the project * MIanagement autonomy-nota- become a highly competent island of to finance studies, activities, and bly in setting salary levels and in excellence within Uganda's poorly other inputs as the need arose. being able to use a share of tax functioning civil service bureaucracy. collections for agency spending. Payments from project funds enabled A positive feature of the process * Abetter incentive framework EAU to offer substantial salary approach in both countries was the for employees. Base salaries were supplements to staff the unit exclu- introduction of the annual work comparable to those in the ci6 iI sively with Ugandans. The unit was plan, which served as a combination service, but allowances (in cash later merged into the combined Min- planning, implementation, and External training w*as made avail- istry of Finance and Ministry of Plan- monitoring tool to ensure adjustment able as an incentive device. ning and Economic Development. of project activities to changing cir- * Better working conditions. Although the supplementary pay- cumstances. In Uganda, the ap- * Selective recruitment to replace ments were necessary to attract proach worked well, but in Ghana, it incompetent and unmotivated Ugandan experts, the practice became a liability, primarily because employees. proved problematic because of ten- of poor project management and * Effectve use of information- sions created by the dual benefit weak Bank supervision. processing technologies. system. Management * Helping establish the basis for comprehensive civil service reforms, For the process approach to work, faced considerable pressure to allow including a 50 percent reduction in it needs strong project management. the funds to be used for equipment the size of the service between 1991 In Uganda, the risks associated with and travel abroad. Often, the re- and 1994. the process approach were recog- quests were barely related to SAIS nized and procedures were gradu- activities. Instead of resisting such * Carrying out agricultural studies ally developed to ensure that pressures, the Ghanaian project man- and policy analysis for the Agricul- activities funded through the project agers tended to comply and the Bank tural Secretariat. These efforts helped were consistent with national priori- supervisors tended to be permissive. convince the government to liberalize ties. By the time TA III was initiated, Poor record-keeping, inadequate the coffee sector and strengthened the project management unit had monitoring, and lack of performance the government's move to a more acquired substantial experience indicators further weakened man- market-oriented strategy. through its association with TA II. agement and supervision. This continuity helped strengthen One important reason for the rela- project management. Ownership tive success of TA III was that many of the project's potential weaknesses In contrast, one of the major weak- The governments of Ghana and had already been identified in TA II nesses of SAIS was its lax project Uganda actively participated in de- and addressed in the design and management, which allowed some fining the policy reform objectives implementation of TA III. Other key of the project to be highjacked by and were committed to implement- factors were the high level of govern- rent-seekers. When funds from ing the economic reforms. But com- ment ownership, strong project man- project components became available mitment to institutional reform, agement, and close Bank supervision. for other activities, project managers including reform of the civil service, September 1995 enhance learning and mutual trust, Box 4: Blueprint versus process design increasing prospects for the project's success. The 'blueprint" approach defines liberatelv designed to build and main- set objectives and specific compo- tain ownership among participants, to * Under the right conditions, foreign nents for achieving objectives, re- emphasize learning and capacity long-term (or "coming and going") quires firm cost estimates (including building, and to cope with changes in advisers can be effective in helping contingencies), and sets a predeter- the operafing environment. While the strengthen institutional capacity mined timeframe. blueprint approach is often preferred without creating excessive depen- for projects involving large physical The "process" approach offers investments. the process approach has dency (see Box 5). But when manage- more flexibility in its choice of objec- proven more suitable for those involv- ment and supervision are weak, no bves, means, and priorities. It is de- ing institutional development. modality will work. * In highly aid-dependent countries, was higher in Uganda. TA III was Lessons such as Ghana and Uganda, care championed and overseen by the must be taken to minimize donor- top officials in the Ministry of Fi- * The more unsettled a country's in- created distortons in local incenssve nance and Economic Planning, who stitutional environment, the greater Such distortons can be especially knew what they wanted from the the need for a flexible, process-ori- severe when donors pay varying sal- project and worked actively to ented project design. ary supplements to civil servants and promote it.arsupeettocvlsvnsad prome . . . compete with each other to recruit . A flexible project design demands skilled employees. Bank performance strong project management. The main challenge in TA operations in coun- The Bank had an important role tries emerging from serious gover- in defining the projects' highly rel- nance crisis is establishing disciplined Box 5: Innovative use of evant objectives and in introducing project management to ensure con- foreign consultants in TA 111 the process approach that character- stant reassessment of issues and ized the projects' design. In Uganda, progress toward long-term objectives. One method s.ucces,fullV used Bank supervision was intensive, of The main lesson of TA III is that in Liganda's central bank is the high quality, and geared to a practi- "management matters most." ' 'coming and going of outside ex- cal, results-oriented management perts." An expatriate technical ad- approach. Moreover, the continuity * There are no substitutes for owner- viser spends an initial period in the and cohesiveness of the Bank's ship on the borrower's side and in- country becoming acquainted with country team helped strengthen depth country knowledge and careful the local env ironment and the na- project implementation, adding supervision on the Bank's side. Bor- tu resLsthe problemts to he ad - further discipline to the process rower ownership of institutional de- dressed. During this time, the first further discipline to the process rower s~~~tepu are taken in putting in place approach. velopment objechves must be the new systemand procedUres. established before appraisal, includ- T training ot local stafi begins and Bank supervision of the SAIS ing commitment to civil service re- work plans are agreed on. The con- project, although substantial in form. Moreover, experienced staff sult3nt tht-n leaves, but visits at quantity, was poor in several areas: with substantial field and supervision regular intervals to help solve it was too permissive, lacked formal experience should be assigned to ap- problemi and monitor progress review of the project's development praise and supervise TA projects sup- l effectiveness, and lacked Bank man- porting institutional development. The method avoided the risk ot agement attention. It was therefore dra ing the expatriate experts too unable to identify and confront de- * When government commitment is I much into operational a,signments mands to transfer project funds from backed by the support of a high-level and put pressure ofn national cOLin true institutional development ac- project advocate, the likelihood of the terpa rts to take responsibility for tivities to equipment purchases. A project's success increases markedly. cairrying out the unit's functions. set of performance indicators could At the same time, it kept in place a have helped the Bank identify Continuity and cohesiveness of the support -system and made the project deviations early on, but none Bank's country team and the bor- tocued and co.t-efiectie . had been established. rower's project management team f OED Precis is produced by the Operations Evaluaon Department of the World Bank to help disseminate recent evaluaion findings to development professionals within and outside the World Bank. The views here are those of the Operations Evaluation staff and should not be attributed to the World Bankorits affiliated organ-izations. Please address comments and enquiries to the managing editor, Rachel Weaving, G-7137, World Bank, telephone 473-1719. Internet: rweaving@worldbank.org OED Pr6cis