Human Development 4 April 2002 Findings reports on ongoing operational, economic, and sector work carried out by the World Bank and its member govemments in the Africa Region. It is published periodically by the Knowledge and Leaming Center on behalf of the Region The views expressed In Findings are those of the author/s and should not be attnbuted to the World Bank Group Ghana: Tracking Public Resource Flows in Schools and Clinics he World Bank report, Can Af- The GPETS was designed to track rica Clazm the 215' Century the extent to which public re- (April 2000), called attention sources flow from the district of- to the urgent need for investing in fices to service delivery points. The human capital in Sub-Saharan Af- records of public expenditure allo- rica (SSA). The Government of cation at the central government Ghana (GOG) is committed to in- level will help the tracking survey crease the access to public health to provide a measurement of re- and education services, especially source flows between three points: for the poor. One of its main mea- central government, district offices, sures is to increase the social sec- and public service provision facili- tor budget. However, a higher bud- ties. This survey is based on a get does not necessarily guarantee sample of basic education and pri- better outcomes. It is crucial to mary health care facilities, as well ensure that the allocated resources as the district offices in charge of | - ~ =_ are channeled efficiently to the in- education and health. Due to its tended facilities, reaching the ser- moderate scale and experimental vice users, especially the poor. nature, the results of the study are 4 _ There are two broad categories of meant to be more provocative than reasons which might cause ineffi- conclusive. ciencies in the distribution of pub- The distribution systems of pub- lic expenditures. The first is cor- lic financing in Ghana are divided ruption, namely leakage of re- into two main categories: the dis- sources to individuals or unin- tribution of salary expenditures ;____ tended organizations. The second and of non-salary expenditures. is the inefficient use of resources, The majority of non-salary recur- such as a mismatch between what rent expenditures are distributed --- U t facilities need and what the govern- to schools and clinics through dis- ment distributes The Ghana Pub- trict offices in the form of materl- lic Expenditure Tracking Survey als, while salaries are paid directly (GPETS) measures the first type of by a central government agency, inefficiency, the leakage between the Controller Accountant General different points of public resource (CAG), to public employees, distribution. through the banking system. Al- Table 1 Primary school recurrent expenditure subsidies 1997/98 1998/99 Non- Salary Total Non- Salary Total salary salary Based on MOE estimates 12081 108724 120805 N/A N/A N/A Based on District Education Office estimates N/A N/A N/A N/A N/A N/A Based on school budget estimates 5730 65828 71558 5104 71577 76681 Adjusted for 8% overhead cost 6188 71094 77283 5512 77303 82815 As % of MOE estimates 51 65 64 N/A N/A N/A Source: Ghana Public Expenditure Tracking Survey, 2000. though the central government has Primary and junior secondary clude, based on other referencing a precise budget to allocate public education and public resource information, that the percentage of expenditures, there is little infor- flows leakage is about the same in 1998/ mation at the district offices or fa- 99 as in 1997/98. The results from cilities regarding what is actually The tracking survey at primary the tracking survey also indicate allocated to them and how much schools provided somewhat differ- that GOG's support for the non-sal- they actually receive, due to the ent resource flow information than ary expenditure has decreased, lack of knowledge regarding the indicated by the statistics of the while the non-governmental/pri- monetary values of the materials MOE. Table 1 includes a compari- vate funding has increased signifi- received. This knowledge gap be- son between the expenditures al- cantly. This, however, was mainly tween the monetary values and the located by the MOE to primary the result of a large increase in a material form of the resource flows schools and the resources received few wealthier regions. Non-public may create a much more serious by primary schools. In 1997/98, support tends to be weak in the leakage in the health than in the about 64 percent of the public ex- poorer regions. education sector, due to a much penditures that left the MOE In 1997/98, only about 75 per- larger share of salary expenditure reached schools, after taking into cent of the salary expenditure and in the former and the more resal- account overhead administrative less than 50 percent of the non- able nature of medical supplies. costs. Specifically, 51 percent of salary recurrent expenditure allo- non-salary and 65 percent of sal- cated to Junior Secondary Schools ary recurrent expenditure reached (JSS) reached them. The expendi- the primary schools. Although we ture per student has increased do not have budget data from the slightly between 1998 and 1999 in MOE for 1998/99, we can con- real terms. In the context of decen- Table 2 JSS government recurrent expenditure subsidies 1997/98 1998/99 Non- Salary Total Non- Salary Total salary salary Based on MOE estimates' 19665 176984 196649 N/A N/A N/A Based on District Education N/A N/A N/A N/A N/A N/A Office estimates Based on school budget 8765 123500 132265 11292 133380 142145 estimates Adjusted for 8% over head cost 8765 123500 132265 12195 144051 153517 As % of MOE estimates 48 75 73 N/A N/A N/A Source: Ghana Public Expenditure Tracking Survey, 2000. ' Based on MOE 1998 budget sheet, the salary is about 90 percent of total recurrent expenditure. tralization, GOG has delegated fi- Table 3 nancial resources to the District Financial resource flow in health sector Assemblies Common Fund (DACF) to invest in the projects Government non-salary expenditure of per out patient visit 1998 1999 and programs of local priority, in 1998 cedis The tracking survey, however, ob- served only a small amount of re- Based on MOH estimates 3875 N/A Based on District Health Office sources flowing from DACF to estimates 1231 1599 schools. As % of MOH estimates 31.8 N/A Overall, the government sup- Based on Sub-district clinic estimates 830 991 ports a large proportion of non- As % of District office estimates 67.4 62.0 salary recurrent expenditures - about 70 percent. The capital in- These estimates do not include the overhead administrative expenditures. vestments remain very low on av- erage. In reality, only a few schools resources received by the clinics.( amount to over half of the total received resources for rehabilita- The average exchange rate in 1999 medical cost. Financial resources tion and capital investment. In ad- was 2,669 cedis to the US dollar). from NGOs comprise 7 percent of dition to public resources, schools As shown in Table 3, the track- the total budget. also supplement their resources ing survey shows that by the Dis- with parents' contributions and pri- trict Health Office (DHO) account, Conclusion vate donations. It is common in expenditures per outpatient Ghana that, at the request of their amounts to only about one third of An accurate estimate of public ex- teachers, each student carries a what had been indicated in MOF's penditure flows must start from the small amount of building material, budget. From the DHO to the clin- distribution and recording systems such as stones or sands, to his/ ics, 67 and 62 percent of funds which would permit accurate track- her school. These private contribu- reached clinics in 1998 and 1999, ing. While the strategies to improve tions were most likely not ac- respectively. Overall, only 20 per- these systems in Ghana are beyond counted for. cent of non-salary recurrent expen- the scope of this study, it presents ditures reached the clinics. here the problems encountered Primary health care and public Based on MOF's account, at the while trying to track public expen- resource flows sub-district clinic level, GOG and ditures. Hopefully, this will provide donors provide a total of 87 per- an entry point for relevant parties pubi heal care Gain Gha is sup- cent of medical care cost for pa- to discuss the best ways to increase porte by fudin fromtw bradi tients. However, the survey of the the efficiency of public expenditure soures, publicd health eOG ndio- facilities reveals that only 40 per- distribution. Estimated resource tures, financed by GOG andrdo- cent of the budgets of the sub-dis- flows are also presented. Although nors, and internally generated . .. trict clinics are paid for by the GOG the accuracy of estimates is not funds (IGF), paid by the patients. (Table 4). The internally generated claimed in absolute terms, the pat- It is estimated that total govern- funds, which are paid by patients, terns of the public expenditure dis- ment recurrent expenditures to clinics in 1998, including donors' Tabl 4 funds, amounted to about 5200 A comparison of percent of public expenditure at ministry and facility level cedis per person (Canagarajah and Ye, World Bank, 2001). About 35 Percent of financial resources from percent of 5200 cedis is the salary GOG and donor Intemally NGO's poled fund* Generated Fund assistance expenditure. Since the tracking Based on MOH estimates 87 13 survey does not have salary expen- for sub-district clinics ditures, it only compares the 65 Based on sub-district clinic 39 54 7 estimates percent of 5200 cedis, the non-sal- --__--- ary recurrent expenditure, with the *Including MOH program funding. Data sources: Ghana PER and GPETS 2000 calculation. tribution revealed by these esti- the health sector, where a large The Ghana PETS pointed to an mates were consistent with the per- proportion of expenditures are non- important area where public ser- ceptions of district level education salary related. In addition, the sys- vices could be improved. Enhanced and health officers, to whom these tems have hardly any feedback accountability in public expendi- results were presented at a work- mechanisms. There were instances ture distribution systems could in- shop in Ghana. when more books than needed on crease resources at the facility level The Ghana PETS reveals a dis- one subject were shipped to a by 30 to 100 percent without in- connect between the measurements school, while there was still unmet creasing the current public spend- of the resource flows at the central demand for books in other subjects. ing envelope. Indeed, a similar government and at the facility lev- There were also complaints about study in Uganda has demonstrated els. Consequently, at the facility publicly-procured drugs being that by simply publicizing the bud- level, there is no accurate account more expensive at clinics than get allocation at the district and of the labor costs of rendering ser- those sold at private pharmacies. school levels, the schools received vices, nor of the monetary value of There was sufficient evidence to almost all the resources allocated the materials received. If a consis- demonstrate that significant re- to them, instead of the mere 20 per- tent measurement of the resource source leakage existed both in the cent when such information was flow is in place, there must also be health and education sectors with not available. It is obvious that SSA systematic record keeping and a certain patterns becoming appar- countries need increased resources record filing procedure that would ent. The major leakage occurred to invest in people, but a signifi- ensure an institutional memory. between the central government cant improvement in internal effi- However, the PETS survey found no and the district level offices. There ciency is essential. Otherwise, more standard forms to keep financial was also leakage between the dis- resources could mean greater leak- records at the facility level. Further, trict education offices and schools, age, not better outcomes. people who kept records often had but to a much lesser extent. The no relevant training. There was also leakage between the district health no standard procedure for filing offices and clinics seemed to be records, which were often filed at much more significant. different places and kept by differ- This leakage pattern coincides _____ ent people. with the distribution procedure. The complete version of this The public resource distribution The majority of non-salary public article is forthcoming as a World systems in Ghana are based on a expenditures are transformed from Bank publication. For more infor- considerable degree of control from cash to in-kind at the central gov- mation please contact Xiao Ye the central government. This ernment level, and the distribution (xyecworldbank.org) or method works to some extent, es- of these materials is uncountable Sudharshan Canagarjah pecially in education sector where in terms of their monetary values. (scanagarajah9worldbank.org) at salaries comprise a large propor- This means that there are far fewer the World Bank. tion of the budget. However, such opportunities to divert public funds systems, providing very little local at the district level than at the cen- accountability by the central gov- tral government level, and also ernment, do not work too well in fewer opportunities in the educa- tion than in the health sector.