Assessment of reallocation warrants in Tanzania

Transcription

1 ANALYSIS OF REALLOCATION WARRANTS Final report: Assessment of reallocation warrants in Tanzania July 2014 Scanteam: Team leader Torun Reite and team member Erlend Nordby

2 ANALYSIS OF REALLOCATION WARRANTS Table of Contents Abbreviations and acronyms...3 Executive Summary Background and introduction Background for the assignment Introduction Quantitative analysis Total budget reallocations as share of actual expenditure Total number of reallocations transactions Reallocations on recurrent and capital budget votes Ministry of Agriculture Ministry of Education and vocational training Ministry of Lands and Human settlement Ministry of Health and Social Affairs Ministry of Works Qualitative analysis Dialogue with key stakeholders Examine existing laws, regulations, guidelines CAG audit reports PEFA performance in the region Recommendations...39 Annex A: Terms of Reference...41 Annex B: Documents consulted...44 Annex C: Stakeholders met...46 Annex D: Inception report methodology and approach...47 Annex E: Overview of PEFA performance in selected countries...50 Annex F: Checklist for a reallocation policy...59 Scanteam: Team leader Torun Reite and team member Erlend Nordby

3 Abbreviations and acronyms CAG DfID EAC FY GoT MDA MoF PEFA PFM PFMRP PI ToR TPDC TRA - Controller and Auditor General - Department for International Development - East African Community - Financial year - Government of Tanzania - Ministry, Department, Agency - Ministry of Finance - Public Expenditure and Financial Accountability - Public Financial Management - Public Financial management Reform Program - Performance Indicator - Terms of Reference - Tanzanian Petroleum Development Cooperation - Tanzania Revenue Authority 3

4 Executive Summary DfID commissioned this study of reallocations in Tanzania on behalf of Development partners and in close collaboration with the GoT, Ministry of Finance. Highlights from the quantitative assessment are - An increase in actual expenditure amount reallocated over the three FYs: 2010/11, 2011/12 and 2012/13 also accompanied by an increase in budgets and actual expenditure over the three FYs. - A relatively stable but substantial reallocation as share of actual expenditure that varies between 27% in 2010/11 and 31% in the two following FYs. - Reallocations between votes represent the largest amounts and shares of total actual expenditure and show an increase in the most recent two years of study: 24% (FY ) and 21 % ( ) compared to 14% in The number of reallocation transactions varies considerably from one vote to another, from less than 10 per year to more than 900. The majority of reallocation transactions are within votes, but with lower amounts per transaction than between votes transactions. - The number of reallocations transactions increases substantially in second half of the FY, from January to June. -The three years of study show a consistent reallocation from the capital budget to the recurrent budget. Interviews with stakeholder in the Ministry of Finance and selected Line Ministries reveal four main root causes for the many reallocations. The first is the lack of realistic budgeting and too low budget ceilings and resource envelopes, particularly for the recurrent budget. The second is the constant cash constraint and insufficient cash releases having an adverse effect on budget execution. The PEFA performance and our analysis point to the weaknesses in cash forecasting and the lack of active and systematic use of the information on external financing as the two most important areas to address. The third aspect brought up was the very detailed budgeting on sub-item level, particularly in the case of the recurrent budget. Moreover it was pointed out that the administrative process required to apply for warrants to the Ministry of Finance/Government budget division (MoF/GBD) represented a time-consuming task both for LMs and above all for the MoF/GBD although the added value in terms of effective controls was perceived as limited, particularly in the case of the within vote reallocation warrants. Reassessing the procedures combined with the possible reassessment of the authorities related to reallocation within vote could be considered. Finally, the budget classification in itself was presented as a potential constraint, providing limited flexibility in the use of funds, particularly for Line Ministries that have new activities and projects that have not been properly catered for during budget preparation. It should be noted, however that stakeholders put different emphasis on this aspect. In this context the amount of arrears and relatively limited management of arrears is also an area that contributes to reduced credibility of the budget. It is a clear conclusion from the analysis that budget controls/commitment controls seem to be too detailed and fragmented to be effective and a readjustment of the level of detail and a redesign of the administrative process should be considered. Overall there is a need 4

5 for strengthened quality of budget preparation and budget and commitment control. In the context of a reallocation policy the emphasis should be put on two actions: i) Reducing the number of administrative reallocation processes and strengthen the overall coordination to allow for a more strategic reallocation process ii) Reassessing the classification and coding to reduce the level of detail on budget authorisations. It should be pointed out that the reassessment of classifications and codes would require further analytical work. In the Tanzania context there is an expressed intention by the MoF, Budget Division, to adhere to the principles and internationally accepted good practices of having fewer administrative reallocation processes a year and to aim at capturing more of the movements/changes and individual transactions in each of these administrative reallocation processes. This should be strongly supported, guided and put into practice through a reallocation policy. Based on the this review of the root causes to the numerous reallocation transactions and considerable reallocation as share of actual expenditure a reallocation policy will not in itself be a sufficient to solve or address the issue and a more comprehensive strategy that addresses the 4 main root causes identified in this report (based on discussions and document review and analysis and confirmed by MDAs, MoF and CAG reports) would be required to strengthen budget credibility, commitment controls and predictability of warrants in budget execution. The assessment of PEFA performance in Tanzania (and in other country in the region) supports such an analysis and also points to strengthened cash forecasting and management, arrear management and a more active use of available data on external financing as important areas of improvement. For the two recommended actions within a reallocation policy to become effective, the cash forecasting and management probably needs to be strengthened. 5

6 1 Background and introduction 1.1 Background for the assignment The Government of the United Republic of Tanzania (GoT) has been implementing the Public Finance Management (PFM) reform agenda since Three phases have so far been implemented, covering the period from 1998 to 2012 with significant successes. In order to maintain the achievements recorded in the previous phases, since July 2012, the Government is implementing Phase IV of the Public Financial Management Reform program (PFMRP) focusing on: revenue management; planning and budget management; budget execution transparency and accountability; budget control and oversight and program management, monitoring and communication including change management. In particular, the program strives to foster a culture of financial discipline, attain value for money and combat poverty with the ultimate aim of delivering effective services to citizens. This year the Public Expenditure & Financial Accountability Assessment (PEFA) Report for Mainland Tanzania was finalized. One of the identified weaknesses pertains to the lack of budget credibility. Budget credibility requires actual budgetary releases to be similar to voted budgets; budget spent on approved activities and items and requires appropriate fiscal discipline to be in place. However, the PEFA notes that there is significant variance in budget out-turn when compared to the approved budget over the last three years. It further notes, The reasons lie in the budget preparation system (over-budgeting in the development budget, under-budgeting in some areas of recurrent expenditure) and the budget execution system (cash rationing due to resource uncertainty, lack of commitment controls and insufficient access to excess liquidity in GoT bank accounts held in commercial banks). Based on this Department for International Development (DfID), on behalf of the joint PFMRP financiers and in collaboration with the GoT commissioned this assessment of budget credibility, with an emphasis on a reallocation analysis. 1.2 Introduction The first part of this report contains a quantitative assessment of reallocation transactions and amounts for the three previous financial years 2010/11, 2011/12 and 2013/13 at an aggregate and vote level for the five selected votes agreed with Ministry of Finance. These include: Ministry of Agriculture, Vote 43 Ministry of Education and vocational training, Vote 46 Ministry of Lands and Human settlements, Vote 48 Ministry of Health and Social Affairs, Vote 52 Ministry of Works, Vote 98 The consultancy team was provided with data from different databases managed by the Ministry of Finance but had no direct access to the databases to allow for crosschecks and controls and possible reconciliation with other sources. There are some discrepancies between the data provided and the expenditure and reallocation data publicly available. Nonetheless the quantitative analysis provides an overall picture that illustrates the amount of reallocations at vote and sub-vote level and informs a deeper understanding of the 6

7 composition and amounts pertaining to the reallocation process than what is available in other documents. The findings are considered to be valid even with some variation in amounts. The qualitative analysis presented in this report is based on limited resources and was targeted towards the selected sample of ministries. A follow-up process would have allowed a verification process providing an even deeper understanding of the rationale and root causes of the reallocations processes. This should be considered as a follow-up exercise. 2 Quantitative analysis This section 2 presents the quantitative analysis of reallocations. The first section presents total budget reallocation in absolute terms and expressed as a share of total expenditure. Thereafter the analysis provides an overview of the number of transactions. The latter provides a picture of the administrative processes related to the reallocations. The remainder of the quantitative analysis and the last sections provide an overview of reallocations for the votes of the five ministries selected as a sample to deepen the analysis. 2.1 Total budget reallocations as share of actual expenditure The box summarizes the key findings on total budget reallocations related to aggregate budget and each element is presented in further detail in following sections (Sections ). - There has been an increase in actual expenditure amount reallocated over the three FYs. - There has been an increase in actual expenditure over the three FYs. - Reallocation as a share of actual expenditure is substantial but relatively stable and varies between 27% and 31% - Growth in expenditure amount reallocated as compared to previous FY is largest in 2011/12, with a 40% increase. - Reallocations between votes represent the largest amounts and shares of total actual expenditure and increased in the most recent two years of study: 24% (FY ) and 21 % ( ) compared to 14% in The number of reallocation transactions varies considerably from one vote to another, from less than 10 per year to more than The number of reallocations transactions increases substantially in second half of the FY, from January to June. - The three years of study show a consistent reallocation from the capital budget to the recurrent budget - The reallocation from the capital to the recurrent budget represented 1%, 6% and 3% of actual expenditure in the three years of study. - 7

8 The following table presents the total budget reallocation amounts in absolute terms and as a percentage of total budget for the Financial years: FY , FY , and FY Table 2.1. Total amount of reallocation in absolute terms, all votes included* Financial year Amount in 1000 TZS Share of actual expenditure* % 31% 31% *Excluding vote 20, 30, 38, and 96, due to lack of available information on total expenditures on these votes. There has been an increase in the nominal amount reallocated during the last three financial years. The nominal amount reallocated is based on gross reallocation and movements between and within votes and can thus be higher than the net amounts sometimes reported. There has also been an increase in total budget and total expenditure in the same period. The gross reallocation as a percentage of total expenditure as a share of actual aggregate expenditure has been quite stable from 27% of expenditure in the financial year to 31 % of total expenditure in both 2011/12 and 2012/13. The increase as a percentage of total budget and expenditure in the previous financial year is shown in table 2.2. The table also shows the increase in gross reallocation as compared to the previous financial year. Table 2.2. Increase in budget, budget expenditure and gross reallocation From tables 2.1. and 2.2. above it can be noted that the financial year 2011/12 stands out in terms of showing the greatest increase in gross reallocation, the increase amounting to 40 % of total expenditure. 8

9 In table 2.3 total amount of reallocation between vote and within vote in absolute terms is presented. The analysis is deepened and compares the amounts reallocated between votes with the gross amounts reallocated within votes. The data show the following trend: Table 2.3. Total amount of reallocation between vote and within vote in absolute terms Financial year Between vote Within vote From 2010/11 to 2011/12 there was a great increase in the amount reallocated between votes. Whereas Table 2.2. above revealed a 40% increase in the total gross reallocation, the increase for the between votes is even greater and amounts to about 115% whereas the reallocations within votes decreased. For all three financial years the amounts reallocated between votes are greater than the gross amounts reallocated within votes. This is also illustrated in the following table where the reallocation between vote and the gross reallocation within vote are shown as a percentage of total actual expenditure. Table 2.5. Reallocation between vote and within vote as percentage of actual expenditure Financial year Between vote 14% 24% 21% Within vote 13% 7% 10% * Excluding vote 20, 30, 38, and 96, due to lack of available information on total expenditures on these votes. The two last completed financial years have reallocations between votes amounting to 24% (FY ) and 21 % ( ) of total actual expenditure respectively compared to 14% in For the gross reallocations within votes the amounts are smaller and represent 13% of total expenditure in , 7 % in and 10% in Total number of reallocations transactions This section focuses on the number of transactions, defined as each individual movement or/and change to a specific budget item (sub-item). This includes all movements/changes in the reallocation processes, both within votes and between votes. For every administrative reallocation process there are several transactions involved. The minimum number of transactions per administrative reallocation process is two transactions: one debit and one credit. International good practices indicate that there should be a minimum of movements/changes unless unforeseen incidences occur. Moreover the administrative processes should be kept to a minimum and capture as many of the movements/changes ( reallocations transactions ) as possible to enable a strategic perspective and enable the Ministry of Finance (MoF) and the Line Ministry (LM) to get an overview of total reallocation needs across all budget votes, to enable prioritisation and assess alternative 9

10 sources of financing, measure additional needs up against each other and assess reductions and potential budget cuts - much in the same way as is ideally done during budget preparation. Table 2.5. Number of reallocation transactions between vote and within vote Financial year Between vote Within vote Table 2.5. reveals a great amount of transactions representing movements and changes in votes and sub-items. There is also an increase over the three FYs. The number of transactions within votes is most striking. As shown in Table below both the median and the average number of reallocation transactions have increased from 2010/11 to 2011/12, and from 2011/12 to 2012/13. The median value for FY 2012/13 is 150, whereas the average value is 200. Table Median and average, all votes It is important to note that there is no clear and unified picture across votes when comparing the number of reallocation transactions. Figure 1.1. below reveals that the number of reallocation transaction vary considerably depending on vote, from less than 10 per year to more than 900. The example is provided for the last financial year completed. Figure 1.1 : Total number of reallocation transactions per vote, FY 2012/ Total reallocations per vote FY 2013/ Transactions Figure 1.2. provides a comparison between number of transactions between votes and within votes across votes and gives a visual illustration of what is shown in Table 2.6. The figure confirms that most transactions are within votes. Moreover the large majority of transactions are small amounts linked to within votes and subvotes reallocation processes. 10

11 However, even the within vote reallocation processes need to be formalized through written correspondence from the LM to the MoF/Budget Division and go through a formal assessment and approval process. In view of the enormous amount of transactions involved this represents a great administrative burden not only on the LM but also on the MoF/Budget division. In practice the suggested within vote reallocations are rarely rejected. All LMs interviewed stated that the MoF/Budget Division provided feedback on the decision and formalized the within vote reallocation requests without much delay and normally within 15 days. Figure 1.2 Number of within vote and external vote transactions, FY 2013/ Numbers of within vote reallocations Another important aspect is to see how the number of reallocations transactions is distributed in the financial year. Figure 1.3 in the following shows the number of reallocation transactions per month for the financial year 2012/13. There is a clear trend showing an increase in the number of reallocation transactions in the second half of the financial year and towards the last months, reaching a peak in May and June, the two last months of the financial year. Figure 1.3 shows data for financial year 2012/13 but the pattern is the same for all financial years assessed. Figure 1.3 Number of reallocation transactions per month, FY 2012/13 11

12 Reallocations per month Transactions In the Tanzania context there is an expressed intention by the MoF, Budget Division, to adhere to the principles and internationally accepted good practices of having fewer administrative reallocation processes a year and to aim at capturing more of the movements/changes and individual transactions in each of these administrative reallocation processes. However, based on available information it seems that in practice both between votes reallocation processes and within votes reallocation processes are to a great extent treated on a case by case basis that does not allow for a strategic overview and prioritisation of the use of scarce resources. The MoF/Budget Division publishes regular reports that provide an overview of total administrative reallocation processes, whereas this report provides and overview of total number of reallocation transactions. What is most striking is the enormous number of reallocation transactions within votes. There are 3-4 times more reallocation transactions within votes than between votes. In contrast the gross amount reallocated within votes is considerably lower than the between votes reallocated amounts indicating a clear pattern of lower amounts reallocated per transaction within vote than between votes. This will be assessed in more detail for the individual votes for the 5 sample Line Ministries in following sections. 2.3 Reallocations on recurrent and capital budget votes Table 2.1 below presents the total reallocation between votes in the budget, financial years 2010/ /13. The table reveals that there has been a significant reallocation from the capital budget to the recurrent budget. 12

13 Table 2.1 Total between votes reallocation on capital and recurrent budget 2010/ , all votes The table demonstrates that reallocation from the capital budget to the recurrent budget was moderate in 2010/11, representing 1 percent of the total budget and actual expenditures. In 2011/12, however, the reallocation represented 5 percent of total budget and 6 percent of actual expenditures. In 2012/13 the reallocation from the capital to the recurrent budget amounted to 3 percent of the budget and of actual expenditures. Table 2.2 below presents the net allocation on the capital and recurrent budget in 2010/ /13 on the five ministries sampled for more detailed analysis, i.e. votes 43, 46, 48, 52 and 98. Table 2.2 Net reallocation on capital and recurrent budget 2010/ /13, vote 43, 46, 48, 52 and 98 The table demonstrates that Vote 46 Ministry of Education and vocational training, Vote 48, Ministry of Lands and Human settlements, Vote 52 Ministry of Health and Social Affairs and Vote 98 Ministry of Works, had an overall reduction in their capital budget when the three years are summarized. The reallocations from the capital budget were particularly significant on vote 52 and 98 in 2011/12, and on vote 48 and 52 in 2012/13. Vote 43 Ministry of Agriculture did however receive an overall net increase in the both capital and recurrent votes. 13

14 2.4 Ministry of Agriculture - Number of transactions varies substantially between FY s, from 50 to Moderately higher number of external transactions than within vote transactions, caused by high number of external transactions in 2012/13. - Number of reallocations varies substantially between subvotes - Level of reallocations as share of total expenditure relatively consistent between all three FY s. - Significantly higher level of external reallocations than within reallocations all three FY s. - Significantly higher number of reallocations on recurrent than on capital budget subvotes, in particularly within vote. - No correlation between number of reallocation transactions and level of reallocation as share of expenditures. Within vote transactions were about 24 times lower than between vote transactions in FY 2010/11, but with a higher number of reallocation transactions Number of reallocation transactions Table below shows the total number of reallocation transactions within vote 43, Ministry of Agriculture, and between vote 43 and other budget votes. The highest number of reallocations was in FY 2012/13 with 699 transactions, a sharp increase from 2010/11 with 50 transactions and 2011/12 with 368 transactions. Table Number of allocation transactions, vote 43 Table also shows a moderately higher sum of external (between) vote reallocation transactions than within vote reallocations. The explanation for this is a high number of external vote transactions in 2012/13. Both FY 2010/11 and 2011/12 did however experience a higher number of within vote transactions than between vote transactions. Figure below illustrates the number of within and between vote reallocations on subvotes for FY 2012/13. The figure demonstrates a significant level of variance in number of reallocations both between subvotes and in terms of within and between vote reallocations. Figure Number of within vote and external vote transactions 14

15 2.4.2 Level of reallocation transactions Table below shows the level of reallocations within and between votes 43 in absolute terms for the period 2010/ /13. Despite the relatively large variance in number of transactions between FY s, column 9 shows that the total level of reallocations as share of total reallocations is relatively consistent between the three FY s, with 29 percent both in 2010/11 and 2012/13, and 21 percent in 2011/12. Table Reallocation amounts and shares for Ministry of Agriculture Columns 1 and 2 in the table show the originally approved budget and total expenditures. All three FY s total expenditures were significantly lower than the approved budgets. Column 3 shows the net level of external (between votes) reallocations. The column demonstrates that the lower level of expenditures compared to approved budgets cannot be explained by reallocations to other votes. Only in 2012/13 resources were transferred from vote 43 to other votes. Both in 2010/11 and 2011/12 Ministry of Agriculture received additional allocations from other votes. Columns 4-7 in the table demonstrates that the gross share of external reallocations as a proportion of total expenditures was substantially higher than within vote reallocations all three FY s. There is however significant variance between the FY s. Whereas the level of external vote reallocations was 15 percent in 2011/12, the external reallocations were 28 and 27 percent of total expenditures in 2010/11 and 2012/13. The level of reallocations within vote 43 was low both in 2010/11 (1 percent) and 2012/12 (2 percent), but 7 percent in 2011/12. 15

16 Column 8 shows the total level of reallocations, and column 9 the total gross reallocations as share of total expenditures. Figure below shows the net level of reallocations for FY 2012/13 per subvote. The figures reveal that there were reallocations transactions at all subvotes. However, some subvotes experienced significant higher levels of reallocation transactions than others. Figure Net level of within vote and external vote transactions (in TZS 1000) The figure also demonstrates the substantially higher level of the external transactions. In 2012/13 within vote reallocations accounted for approximately a quarter of total transactions, but the level of within vote reallocations was only 2 percent as share of total expenditure compared to 27 percent on external reallocations. The average amount on each within vote transaction is therefore substantially smaller than the amount on each external vote transaction Capital and recurrent budget transactions Table below decomposes column 3 in table into recurrent and capital budget transactions. The table shows that the Ministry of Agriculture had an increase in capital budget both in 2011/12 and 2012/13, receiving reallocations from other budget votes. The level of recurrent reallocations was however higher than the level of capital reallocations both in 2010/11 and 2012/13. Table Net capital and recurrent reallocations, between votes Figure below presents reallocations on subvotes separated between capital and recurrent budget items, including both within vote and between votes reallocations, in FY 2012/13. As clarified above, Ministry of Agriculture received an overall increase the capital budget, in terms of reallocations. The figure below shows that in particular vote 16

17 2001 received a substantial increase in the capital budget. In addition, within vote reallocations on recurrent budget items were credited. Figure Capital and recurrent reallocations, including within vote and subvote and between votes reallocation transactions, FY 2012/ Review of documentation In the Statement of reallocation warrants between votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided between votes in the financial year 2012/13 showing four reallocations processes between votes: In the first case, the available budget for the Ministry of Agriculture was reduced on a number of sub-items and the amounts used to finance salary adjustments, domestic debt servicing and contingencies. In the second case, the budget was increased to finance expenses within the National Food Reserve Agency. In the third case the budget was reduced for consultancy services to finance a budget increase in the Contingency vote managed by the Treasury. In the fourth case a number of expenditure items within the Ministry of Agriculture were increased and budget items related to consultancy work was additionally reduced. There are no narratives providing information or justifications beyond the budget numbers given and the details on which votes and sub-items were increased and which votes and sub-items were reduced to finance these increases. Special attention should be drawn to the fact that the Treasury contingency vote was increased without pointing out the use of these funds. This reduces transparency of allocations and resource use. It should be noted that there is a discrepancy in the amounts provided in the Statement of reallocation warrants and the data provided from the MoF/Budget division. In the Statement of reallocation warrants within votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided within votes in 17

18 the financial year 2012/13 showing the number of reallocations transactions within votes for the recurrent budget and the development budget respectively. The report also illustrates the great level of detail of the budget commitment controls at sub-item level. The largest numbers of reallocations within votes are within the recurrent budget whereas for the development budget the numbers of transactions are limited. 2.5 Ministry of Education and vocational training - Number of reallocation transactions varies between FY s from 676 to Significantly higher number of within vote transactions than between vote transactions. - All subvotes have reallocations, but the number of reallocations varies substantially between subvotes - Level of reallocations as share of total expenditure varies significantly between the three FY s, with a peak in 2011/12 and relatively moderate levels in 2010/11 and 2012/13 - High level of external reallocations in 2011/12, moderate level in 2010/11 and 2012/13 - Moderate level of within vote reallocations all three financial years, 3 6 percent to total expenditure. - In average each internal reallocation transaction is a limited amount whereas the average external reallocation transactions are substantially larger amounts Number of reallocation transactions Table below shows the total number or reallocation transactions within vote 46, Ministry of Education and vocational training, and between vote 46 and other votes. The total number of transactions was lowest in 2010/11 at 676 transactions. The number of transactions was substantially higher in 2011/12 and in 2012/13, respectively at 975 and 931 transactions. The table also shows that the number of within vote reallocation transactions was substantially higher than between vote reallocations all three FY s, with a total of 2287 within vote transactions and only 295 between vote transactions. Table Number of reallocation transactions 18

19 Figure below shows the number of reallocation transactions per subvote in 2012/13. The figures demonstrate that all subvotes were subject to within vote transactions, whereas only five of the 15 subvotes were subject to external vote transactions. Figure Number of within vote and external vote transactions Numbers of within vote reallocations Number og external vote transactions Level of reallocation transactions Table below shows the level of reallocations within and between vote in the Ministry of Education and Vocational training in absolute terms. Column 8-9 shows that the levels of reallocation, as shares of total expenditure, varies significantly between FY s, at 10 percent in 2010/11, 19 percent in 2011/12 and 7 percent in 2012/13. Table Reallocation amounts and shares for Ministry of Education and vocational training Column 1 and 2 demonstrates that total expenditures were significantly lower than approved budget all three FY s, with variance peaking in 2010/11. Column 3 shows that both in 2010/11 and 2012/13 resources were reallocated from vote 48 to other votes. In 2011/12 however, 91 billion shillings were allocated to the ministry form other votes. Columns 4-5 shows that the between votes reallocations peaked in 2011/12 with 16 percent of reallocations to total expenditures, whereas the level of between vote reallocations remained reasonable low in 2010/11 and 2012/13, at 4 percent and 2 percent to total expenditures. 19

20 Columns 6-7 shows that within votes reallocations were at a moderate level all three financial years, at 3 6 percent to total expenditures, with the highest level of within reallocations, both in nominal terms and as share of expenditures, in 2010/11. Figure below shows the net level of reallocations for FY 2012/13 per subvote. The figure demonstrates that all subvotes were subject to reallocations. With the exception of one subvote, subvote 7001, the net amounts of reallocations were moderate. Figure Net level of within vote and external vote transactions Capital and recurrent reallocations Table below show the net level of external vote reallocations on capital and recurrent budget votes. The table shows that all external reallocations made to / from the ministry in 2010/11 and 2011/12 were to recurrent budget subvotes / items. The net level of reallocations on capital budget votes was zero both in 2010/11 and 2011/12 and -3 billion in 2012/13. Table Net capital and recurrent reallocations, between votes Figure below presents reallocations on subvotes separated between capital and recurrent budget votes for FY 2012/13. The figure confirms that there was an overall net reallocation from both the recurrent and the capital budget. The figure demonstrates that in particular subvote 7001 was credited. 20

21 Figure Capital and recurrent reallocations, including within vote and subvote and between votes reallocation transactions, FY 2012/ Review of documentation In the Statement of reallocation warrants between votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided within votes in the financial year 2012/13 showing three administrative reallocations processes between votes: In the first case, the available budget for the Ministry of Education was reduced on a number of sub-items and the amounts used to finance salary adjustments, domestic debt servicing and contingencies. In the second case, the budget was increased to finance operational expenditure within the Ministry of Education. In the third case the budget was reduced for colleges and universities to finance a budget increase in the Contingency vote managed by the Treasury. There are no narratives providing information beyond the amounts and the details on which votes and sub-items were increased and which votes and sub-items were reduced to finance these increases. In the Statement of reallocation warrants within votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided within votes in the financial year 2012/13 showing the number of reallocations transactions within votes for the recurrent budget and the development budget respectively. In contrast to the Ministry of agriculture presented above, the Ministry of education and vocational training has a substantial number of reallocations within votes within the development budget, although by far the largest number of changes is found within the recurrent budget. 21

22 2.6 Ministry of Lands and Human settlement - Substantial increase in number of reallocation transactions, from 302 to times as many within vote reallocation transactions as between vote transactions. - Number of reallocations varies substantially between subvotes - Relative consistent level of between votes and within vote reallocations to share of expenditure in 2010/11 and 2012/13, at 7 10 percent. - Very high level of reallocations from vote in 2012/13, which cannot be explained by available data. Subvotes 2001 and 3001 credited. - Moderate level of between vote reallocations on both recurrent and capital budget items in 2010/11 and 2011/12. - No correlation between number of reallocation transactions and level of relocation as share of expenditures Number of reallocation transactions Table below shows the total number or reallocation transactions within vote 48, Ministry of Lands and Human settlement, and between vote 48 and other votes. The total number of reallocation transactions has increased substantially over the period, with 302 transactions in 2010/11, 512 transactions in 2011/12 and 610 transactions in 2012/13. Whereas the total level of external vote transactions for the three years were 164, the total number of within vote transactions were almost 8 times as many, at 1260 transactions, showing a gradual increase from 2010/11 to 2012/13. Table Number of within and between vote reallocation transactions Figure below shows the distribution of transactions between subvotes in financial year 2012/13. Whereas there were within- vote transactions on all subvotes, only 8 of 15 subvotes had external vote transactions. Figure Number of within vote and external subvote transactions 22

23 2.6.2 Level of reallocation transactions Table below shows the level of reallocations within vote 48, Ministry of Lands and Human settlement, and external reallocations between the ministry and other budget votes. Column 9 shows that the total level of gross reallocations were significant in 2010/11 and 2011/12 at 16 percent and 19 percent to total expenditures. In 2012/13 the level of reallocations were exceptional high at 208 percent. Table Reallocation amounts and shares for Ministry of Lands and Human settlement Column 1 and 2 shows that total expenditures were significantly lower than the approved budgets all three FY s, in particular in 2010/11 and 2011/12. Column 3 shows a low level of reallocations to external votes in 2010/11 and a moderate reallocation to vote 48 in 2011/12. The extreme level of reallocation from vote 48 in 2012/13 cannot be explained with the data made available to undertake the current analysis. Columns 4 and 5 show the gross level of between vote reallocations and the share of gross reallocations to total expenditures. Gross reallocations were moderate both in 2010/11 and 2011/12 at 9 percent. Columns 6-7 show the gross level of within vote reallocations. The level was moderate in 2010/11 and 2011/12 at 7 10 to total expenditures, and high in 2012/13 at 30 percent. Figure below shows the net level of reallocations for FY 2012/13 per subvote. The figure shows that the extreme level of external reallocations is from vote

24 Figure Net level of within vote and external vote transactions (in TZS1000) Capital and recurrent reallocations Table below shows the net level of external vote reallocations on capital and recurrent budget votes. The table shows that the capital budget was credited with moderate amounts in 2010/11 and 2011/12, but that substantial reallocation transactions were credited in 2012/13. On recurrent budget items the sum of reallocation transactions were relative moderate all three FY s. Table Net capital and recurrent reallocations, between votes Figure below presents reallocations on subvotes separated between capital and recurrent budget items for FY 2012/13. The figure shows that the high level of capital reallocations were credited subvote 3001and Figure Capital and recurrent reallocations, including within vote and subvote and between votes reallocation transactions, FY 2012/13 24

25 2.6.4 Review of documentation In the Statement of reallocation warrants between votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided within votes in the financial year 2012/13 showing two administrative reallocations processes between votes: In the first case, the available budget for the Ministry of Lands and Human settlement was reduced on a number of sub-items and the amounts used to finance salary adjustments, domestic debt servicing and contingencies. In the second case, the budget was reduced within the Ministry of Lands and Human settlement to finance an increase under the National Identity Cards Agency. There are no narratives providing information beyond the information given on which votes and sub-items were increased and which votes and sub-items were reduced to finance these increases. Special attention should be drawn to the fact that the contingency vote is increased without pointing out the use of funds. This reduces transparency of allocations and resource use. It should be noted that there is a discrepancy in the amounts provided in the Statement of reallocation warrants and the data provided from the MoF/Budget division. In the Statement of reallocation warrants within votes published by the Ministry of Finance in Tanzania for the in the financial year 2012/13 showing the number of reallocations transactions within votes for the recurrent budget and the development budget respectively. The Ministry of lands and human settlements has a substantial number of reallocations within votes within the recurrent and a limited amount of reallocations within the development budget. 25

26 2.7 Ministry of Health and Social Affairs - Number of reallocation transactions varies between FY s from 220 to Three times more within vote transactions than between vote transactions - All subvotes subject to reallocation transactions, but the number of transactions varies substantially between subvotes - Level of reallocations as share of total expenditure varies significantly between all three FY s with a peak in 2011/12 (29%) and relatively moderate level in 2010/11 (4%) and moderate in 2012/13 (16%). - High level of external reallocations in 2011/12 and 2012/13, moderate level in 2010/11. - Moderate level of internal reallocations all three financial years, 0-4 percent to total expenditure. - Significant reallocations from capital budget items in 2011/12 and 2012/13, net increase in recurrent budget items except 2012/ Number of reallocation transactions Table below shows the total number or reallocation transactions within vote 52, Ministry of Health and Social Affairs, and between vote 52 and other votes. The table shows significant variance in number of reallocation transactions between FY s, with the highest number of transactions in 2011/12 with 793 transactions and the lowest number of transactions in 2010/11 with 240 transactions. The total number of within vote reallocation transactions is three times as high as between vote reallocation transactions. There is however large variance between FY s. In 2010/11 there were 202 within vote transactions and 38 between vote transactions. In 2011/12 there were 752 within vote transactions and 41 between vote transactions. In 2012/13 however, the large majority of transactions were between votes. Table Number of within and between vote transactions Figures and below shows the number of between and within vote transactions on subvotes in 2012/13 and 2011/12. Whereas all subvotes had some reallocations to external votes in 2012/12, only three subvotes were subject to external reallocation transactions in 2012/13. The figures also demonstrates that it is not predictable from one FY to next which subvotes will be subject to reallocations. However, four subvotes, i.e. 2001, 3001,

27 and 5001, have been subject to a significant number of reallocations both in 2011/12 and in 2012/13. Figure Number of within and between vote transactions on subvotes in 2012/13 Figure Number of within and between vote transactions on subvotes in 2012/ Level of reallocation transactions Table below shows the level of reallocations within vote 52, Ministry of Health and Social Affairs, and external reallocations between the ministry and other budget votes. Column 9 shows that the total level of gross reallocations was substantial both in 2011/12 and 2012/13. The total level of reallocations increased from 4 percent of total expenditures in 2010/11 to 25 percent in 2011/12 and 16 percent in 2012/13. Table Reallocation amounts and shares for Ministry of Health and Social Affairs 27

28 Columns 1 and 2 show that the share of expenditures to approved budget has sharply increased during the period 2010/ /13. In 2012/13 the level of expenditures moderately exceeded the approved budget. Column 3 shows that the level of expenditure compared to approved budget cannot be explained by reallocation transactions. In 2012/13 80 billion TZS were reallocated to other votes. The level of between vote reallocations was also high in 2011/12, reaching 59 billion, but less than 1 percent of total expenditures in 2010/11. Columns 4 and 5 show the gross level of reallocations and share of gross reallocations between votes to total expenditures. Gross reallocations were low in 2010/11, at 2 percent, but relatively high in 2011/12 and 2012/13, at 25 percent to total expenditures in 2011/12 and 16 percent in 2012/13. Columns 6 and 7 show the gross level of within vote reallocations. The level was low all three FY s, between 0 and 4 percent to total expenditures. Figure below shows the net level of between and within vote reallocations per subvote in FY 2012/13. The figure demonstrates that the level of within vote reallocations were at an insignificant level on all subvotes. It also demonstrates that the great share of between votes reallocations were on two of the subvotes, i.e. subvote 2001 and Figure Net level of within vote and external vote transactions (in TZS1000) Capital and recurrent reallocations Table below shows the net level of external vote reallocations on capital and recurrent budget votes. In total for all three FY s vote 52 was credited 158,6 billion from capital budget items, and debited 23.8 billion on recurrent budget items. The table shows that the capital budget was credited with a moderate amount in 2010/11 and with a significant amount 2011/12. Recurrent budget subvotes and items were debited with a moderate amount in 2010/11 and with a significant amount 2011/12. In 2012/13 both the capital and recurrent budget were credited. 28

29 Table Net capital and recurrent reallocations, between votes Figure below presents the reallocations on subvotes separated between capital and recurrent budget items for FY 2012/13. The figure confirms that the bulk of credit allocation transactions were related to capital budget items in subvote 2001 and Figure Capital and recurrent reallocations, including within vote and subvote and between votes reallocation transactions, FY 2012/ Review of documentation In the Statement of reallocation warrants between votes published by the Ministry of Finance in Tanzania there is an overview of reallocation warrants provided within votes in the financial year 2012/13 showing two administrative reallocations processes between votes: In the first case, the available budget for the Ministry of Health and Social Affairs was reduced on a number of sub-items and the amounts used to finance salary adjustments, domestic debt servicing and contingencies. In the second case, the budget was increased within the Ministry of Health and Social Affairs to finance an increase under sub-items for Medical Treatment abroad and expenses for Civil Servants. As a general observation it can be reiterated that there are no narratives providing justifications for the changes made. 29

30 In the Statement of reallocation warrants within votes published by the Ministry of Finance in Tanzania for the in the financial year 2012/13 showing the number of reallocation transactions within votes for the recurrent budget and the development budget respectively. The Ministry of Health and social affairs has a limited number of reallocations within votes within the recurrent as compared to the other LMs presented above. No reallocations within the development budget were identified. 2.8 Ministry of Works - Number of reallocation transactions varies between FY s from 220 to Three times more within vote transactions than between vote transactions - All subvotes subject to reallocation transactions, but the number of transactions varies substantially between subvotes - The level of total reallocations as share of total expenditure is relative consistent between FY s, at percent. - Moderate level of external reallocations in 2010/11, high level in 2011/12 and 0 in 2012/13. - High level of internal reallocations in 2010/11 and 2012/13, and moderate level in 2011/ percent to total expenditure. - Significant reallocations from capital budget items in 2011/12, increase in recurrent budget items 2010/11 and 2011/ Number of reallocation transactions Table below presents the total number or reallocation transactions within vote 98, Ministry of Works, and between vote 98 and other votes for 2010/11, 2011/12 and 2012/13. The table shows that the number of reallocation transactions has increased, from 237 in 2010/11, to 399 in 2011/12 and 437 in 2012/13. The table also demonstrates that the number of within vote transactions were substantially higher than between vote transactions all three financial years, with a total of 806 within vote reallocation transactions and 267 between vote reallocation transactions. Table Number of within and between vote allocations 30