Document of The World Bank IMPLEMENTATION COMPLETION AND RESULTS REPORT (IDA-H8890, IDA-0340, AND IDA-0900) ON A

Transcription

1 Public Disclosure Authorized Document of The World Bank Report No: ICR Public Disclosure Authorized Public Disclosure Authorized IMPLEMENTATION COMPLETION AND RESULTS REPORT (IDA-H8890, IDA-0340, AND IDA-0900) ON A GRANT IN THE AMOUNT OF SDR 2 MILLION (US$ 3.0 MILLION EQUIVALENT) AND A GRANT IN THE AMOUNT OF SDR 1.1 MILLION (US$ 1.5 MILLION EQUIVALENT) AND A GRANT IN THE AMOUNT OF SDR 2.2 MILLION (US$ 3.0 MILLION EQUIVALENT) TO TUVALU FOR A FIRST DEVELOPMENT POLICY OPERATION AND A SECOND DEVELOPMENT POLICY OPERATION Public Disclosure Authorized AND A SECOND DEVELOPMENT POLICY OPERATION: SUPPLEMENTAL FINANCING December 20, 2017 Macroeconomics and Fiscal Management Global Practice Country Management Unit for Papua New Guinea and Pacific Islands East Asia and Pacific Region

2 CURRENCY EQUIVALENTS (Exchange Rate Effective December 20, 2017) Currency Unit US$ 1.00 = AUD 1.30 FISCAL YEAR [January 1 December 31] ABBREVIATIONS AND ACRONYMS AUD Australian Dollar PE Public Enterprise CAS Country Assistance Strategy PEFA Public Expenditure and Financial Accountability CIF Consolidated Investment Fund PFM Public Financial Management DFAT Department of Foreign Affairs and Trade PRM Policy Reform Matrix DPL Development Policy Loan QEA Quality at Entry DPO Development Policy Operation QSA Quality at Supervision FFA Fisheries Forum Agency RMS Revenue Management System FY Fiscal Year RPF Regional Partnership Framework GDP Gross Domestic Product SOE State Owned Enterprise GoT Government of Tuvalu SDE Special Development Expenditure GFC Global Financial Crisis SDR Special Drawing Rights ICR Implementation Completion and Results Report TK Te Kakeega IDA International Development Association TKII Te Kakeega II IFMIS Integrated Financial Management Information TKIII Te Kakeega III System IRD Internal Revenue Department TMTEF Tuvalu Medium Term Expenditure Framework LoCAL Local Climate Adaptive Living TMTI Tuvalu Maritime Training Institute M & E Monitoring and Evaluation TMTS Tuvalu Medical Treatment Scheme MFAT Ministry of Foreign Affairs and Trade TTF Tuvalu Trust Fund OAG Office of the Auditor General TWOG Tuvalu Whole of Government MFED Ministry of Finance and Economic Development USD United States Dollar PD Program Document VLH Vaiaku Lagi Hotel PDO Program Development Objectives Vice President: Country Director: Senior Practice Director Practice Manager: Task Team Leader: Lead ICR Author: Victoria Kwakwa Michel Kerf Carlos Felipe Jaramillo Ndiame Diop David Knight David Knight ii

3 A. BASIC INFORMATION Operation 1 Country Tuvalu Program Name: Tuvalu Development Policy Operation Program ID: P L/C/TF Number(s) IDA-H8890 ICR Date: 12/20/2017 ICR Type: Core ICR Financing Instrument: DPO Borrower MINISTRY OF FINANCE AND ECONOMIC DEVELOPMENT Original Total Commitment USD 3.00M Disbursed Amount USD 3.09M Implementing Agencies: Cofinanciers and Other External Partners: Operation 2 Country Tuvalu Program Name: Tuvalu Second Development Policy Operation Program ID: P L/C/TF Number(s) IDA-D0340 ICR Date: 12/20/2017 ICR Type: Core ICR Financing Instrument: DPO Borrower MINISTRY OF FINANCE AND ECONOMIC DEVELOPMENT Original Total Commitment USD 1.50M Disbursed Amount USD 4.62M Implementing Agencies: Cofinanciers and Other External Partners: Operation 3 Country Tuvalu Program Name: Tuvalu Second Development Policy Operation: Supplemental Financing Program ID: P L/C/TF Number(s) IDA-D0900 ICR Date: 12/20/2017 ICR Type: Core ICR Financing Instrument: DPO Borrower MINISTRY OF FINANCE AND ECONOMIC DEVELOPMENT Original Total Commitment USD 3.00M Disbursed Amount USD 3.07M Implementing Agencies: Cofinanciers and Other External Partners: B. KEY DATES Tuvalu Development Policy Operation P Process Date Process Original Date Revised / Actual Date(s) Concept Review: 05/08/2013 Effectiveness: 04/04/ /31/2014 iii

4 Appraisal: 09/30/2013 Restructuring(s): Approval: 11/22/2013 Mid-term Review: Closing: 06/30/ /30/2014 Tuvalu Second Development Policy Operation P Process Date Process Original Date Revised / Actual Date(s) Concept Review: 06/25/2014 Effectiveness: 05/18/ /06/2015 Appraisal: 01/16/2015 Restructuring(s): Approval: 03/26/2015 Mid-term Review: Closing: 06/30/ /31/2016 Tuvalu Second Development Policy Operation: Supplemental Financing P Process Date Process Original Date Revised / Actual Date(s) Concept Review: 07/21/2015 Effectiveness: 11/27/2015 Appraisal: 07/21/2015 Restructuring(s): Approval: 09/15/2015 Mid-term Review: Closing: 06/30/ /31/2016 C. RATINGS SUMMARY C.1 Performance Rating by ICR Overall Program Rating Outcomes Risk to Development Outcome Bank Performance Borrower Performance Moderately Unsatisfactory High Moderately Unsatisfactory Moderately Satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Overall Program Rating Bank Ratings Borrower Ratings Quality at Entry Moderately Unsatisfactory Government: Moderately Unsatisfactory Quality of Supervision: Moderately Unsatisfactory Implementing Agency/Agencies: Moderately Unsatisfactory Overall Bank Performance Moderately Unsatisfactory Overall Borrower Performance Moderately Satisfactory C.3 Quality at Entry and Implementation Performance Indicators Tuvalu Development Policy Operation P Implementation Indicators QAG Assessments (if any) Rating Performance Potential Problem Program No Quality at Entry (QEA) None at any time (Yes/No): Problem Program at any Quality of Supervision No None time (Yes/No): (QSA) DO rating before Satisfactory Closing/Inactive status Tuvalu Second Development Policy Operation P Implementation Indicators QAG Assessments (if any) Rating Performance iv

5 Potential Problem Program at any time (Yes/No): Problem Program at any time (Yes/No): DO rating before Closing/Inactive status No Quality at Entry (QEA) None No Quality of Supervision (QSA) None D. SECTOR AND THEME CODES Tuvalu Development Policy Operation P Original Actual Major Sector Public Administration Other Public Administration Sub-National Government Central Government (Central Agencies) Education Tertiary Education Health Health Major Theme/Theme/Sub Theme Economic Policy Fiscal Policy Tax policy Human Development and Gender Education Access to Education 5 5 Science and Technology 5 5 Standards, Curriculum and Textbooks 5 5 Teachers 5 5 Health Systems and Policies Health System Strengthening Public Sector Management Public Administration Transparency, Accountability and Good Governance Public Finance Management Domestic Revenue Administration Public Expenditure Management Tuvalu Second Development Policy Operation P Original Actual Major Sector Agriculture, Fishing and Forestry Livestock Fisheries Public Administration Other Public Administration Central Government (Central Agencies) v

6 Education Workforce Development and Vocational Education Health Health Major Theme/Theme/Sub Theme Economic Policy Fiscal Policy Fiscal sustainability Public Expenditure Policy 3 3 Tax policy Macro-financial policies 9 9 External Finance 3 3 Macroeconomic Resilience 3 3 Monetary and Credit Policies 3 3 Human Development and Gender Education Access to Education 5 5 Science and Technology 5 5 Standards, Curriculum and Textbooks 5 5 Teachers 5 5 Health Systems and Policies Health System Strengthening Public Sector Management Public Administration Transparency, Accountability and Good Governance Public Finance Management Debt Management Public Expenditure Management E. BANK STAFF Tuvalu Development Policy Operation P Positions At ICR At Approval Vice President: Victoria Kwakwa Axel van Trotsenburg Country Director: Michel Kerf Franz R. Drees-Gross Practice Manager/Manager: Ndiame Diop Vivek Suri Task Team Leader: Lucy Pan Shireen Mahdi ICR Team Leader: David Stephen Knight ICR Primary Author: David Stephen Knight Tuvalu Second Development Policy Operation P Positions At ICR At Approval Vice President: Victoria Kwakwa Axel van Trotsenburg Country Director: Michel Kerf Franz R. Drees-Gross Practice Manager/Manager: Ndiame Diop Mathew A. Verghis Senior Practice Director Carlos Felipe Jaramillo Marcelo Giugale Task Team Leader: Lucy Pan Lucy Pan ICR Team Leader: David Stephen Knight vi

7 ICR Primary Author: David Stephen Knight Tuvalu Second Development Policy Operation: Supplemental Financing P Positions At ICR At Approval Vice President: Victoria Kwakwa Axel van Trotsenburg Country Director: Michel Kerf Franz R. Drees-Gross Practice Manager/Manager: Ndiame Diop Mathew A. Verghis Senior Practice Director Carlos Felipe Jaramillo Marcelo Giugale Task Team Leader: Lucy Pan Lucy Pan ICR Team Leader: David Stephen Knight ICR Primary Author: David Stephen Knight F. RESULTS FRAMEWORK ANALYSIS Program Development Objectives (from Program Document) The objective of the proposed operation is to support reforms that strengthen public financial management and improve the delivery of social services, while providing critical financing to enable the rebuilding of fiscal buffers. Revised Program Development Objectives (as approved by original approving authority) The first pillar of the program focuses on strengthening public financial management with the specific development objectives of improved commitment control, enhanced oversight of fishing revenue, and streamlined reporting of local governments (Kaupules). The second pillar focuses on improved social service delivery with the specific objectives of enhancing the efficiency of the tertiary health and education schemes to enable strengthened focus on primary and preventative healthcare and primary and secondary education (including vocational training), with increased access for women. Indicator(s) Tuvalu Development Policy Operation P Indicator Baseline Value Original Target Values (from approval documents) vii Formally Revised Target Values Actual Value Achieved at Completion or Target Years Indicator 1 : Variance in composition of expenditure Value quantitative or Average variance > Average variance < Average variance = 17.4% Qualitative 10% 10% Date achieved 12/31/ /31/ /31/2015 Comments (incl. % Not met, although budget execution was affected by the impact of Tropical Cyclone Pam during the results period. achievement) Indicator 2 : Value quantitative or Qualitative Share of newly approved projects implmented in-line with the outer island plans for community development n/a 2015 => 50% No data available Date achieved 12/31/ /31/ /31/2016 Comments Cannot be assessed. No baseline or data source has been developed for this indicator.

8 (incl. % achievement) While some projects are in line with some aspects of the policy, no determination on adequacy has been made for either the baseline or results periods. Indicator 3 : Value quantitative or Qualitative Increased budget allocation to primary & secondary health through savings on TMTS AUD Increase AUD 4.6 million increase million on average by at on average of 15 percent least 5 percent per annum Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Met. Indicator 4 : Value quantitative or Qualitative Increased budget allocation to primary & secondary education (including vocational) through savings on tertiary scholarships AUD 2.8 million Increase on AUD 5.9 million - increase average by at on average of 28 percent least 5 percent per annum Date achieved 12/31/ /31/ /31/2015 Comments (incl. % Met. Even based on later years, and excluding investment-related expenditure, annual average growth in primary expenditure exceeds 5 percent. achievement) Tuvalu Second Development Policy Operation P Original Target Indicator Baseline Value Values (from approval documents) Indicator 1 : Stock and monitoring of arrears Value quantitative or Qualitative Stock of arrears = 8.9% No arrears monitoring Stock of arrears less than 2% of expenditure; Annual statement of arrears produced Formally Revised Target Values Actual Value Achieved at Completion or Target Years Stock of arrears = 1.7% Accounts receivable and accrued expenses reported in audited TWOG accounts Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Met. Named annual statement of arrears not produced, but accounts receivable and accrued expenses now reported annually in audited accounts. Despite ambiguity of the intended definition, this is a viable measure of arrears and met the target. Indicator 2 : Value quantitative or Qualitative Improvement management of revenue records Full reconciliation Up to date not available Treasury records are reconciled at least quarterly and do not deviate by more than 10% from Full reconciliation not available. viii

9 those of the revenue and fisheries management systems Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Indicator 3 : Value quantitative or Qualitative Not met. 2016, 2015 and 2014 audits of whole of government accounts states full reconciliations of Treasury records with revenues and fisheries management systems not available due to lack of comprehensive information in fisheries and revenue management Local government (Kaupule) financial accounts are compiled on a timely basis and made public Kaupule audited accounts are not compiled / published At least 3/4 of annual Kaupule accounts are compiled within the statutory time limit and published on the Auditor's website. No Kaupule audited accounts are published on the auditors website for any years after 2011, nor have they been approved by Parliament or published in another form. Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Not met. Kaupule accounts should be tabled in Parliament 9 months after FY end (1 Mar to 28 Feb). Kaupule accounts for FY2012 to FY2014 are reportedly all completed and awaiting final sign-off but not submitted to Parliament. FY2015 not yet completed. Indicator 4 : Value quantitative or Qualitative Increased rigor of the Tuvalu Medical Treatment Scheme referral process Number of referral Number of 146 referral cases, of which cases scrutinized referral cases consultations indicate that by a scrutinized by a all are scrutinized by multidisciplinary multidisciplinary referral committee although referral committee referral in the cases where treatment = 0 committee > 85% is not recommended their recommendation is not usually accepted. Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Partially met. The anticipated data source for assessment is not available, and the assessment is made partially based on discussion, rather than data, so the accuracy of the assessment is particularly uncertain. Indicator 5 : Value quantitative or Qualitative Increased budget allocation to primary & secondary education (including vocational) AUD 2.8 million Increase on AUD 5.9 million - increase average by at on average of 28 percent least 5 percent per annum Date achieved 12/31/ /31/ /31/2015 Comments (incl. % Even based on later years, and excluding investment-related expenditure, annual average growth in primary expenditure exceeds 5 percent. achievement) Indicator 6 : Availability of gender disaggregated data on beneficiaries of the broadened TMTI training program ix

10 Value quantitative or Qualitative None Data available No data available (because women no longer able to enroll in TMTI courses) Date achieved 12/31/ /31/ /31/2016 Comments (incl. % achievement) Not met. Mixed-sex course supported under the DPO stopped being offered, and the only course offered now is not open to women. Since the baseline no new gender-disaggregated data has been produced. G. RATINGS OF PROJECT PERFORMANCE IN ISRs Tuvalu Development Policy Operation P Date ISR Actual Disbursements No. DO IP Archived (USD millions) 1 04/20/2014 Satisfactory Satisfactory 3.09 H. RESTRUCTURING x

11 GOVERNMENT OF TUVALU First Development Policy Operation Second Development Policy Operation Second Development Policy Operation: Supplemental Financing CONTENTS DATA SHEET A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Program Performance in ISRs H. Restructuring 1. Program Context, Development Objectives and Design Key Factors Affecting Implementation and Outcomes Assessment of Outcomes Assessment of Risk to Development Outcome Assessment of Bank and Borrower Performance Lessons Learned Comments on Issues Raised by Borrower/Implementing Agencies/Partners Annex 1. Bank Lending and Implementation Support/Supervision Processes Annex 2. Borrower s Comments on Draft ICR Annex 3. Comments of Cofinanciers and Other Partners/Stakeholders Annex 4. List of Supporting Documents MAP xi

12 1. Program Context, Development Objectives and Design 1.1 Context at Appraisal 1. Tuvalu is a very small country located in the North Pacific Ocean, comprising nine islands with a total population of around 11,000 people and landmass of only 26km 2. Tuvalu is the smallest World Bank member state, and is remote from any other sizeable population centers, with the nearest population mass larger than one million people being New Zealand, 3,500 kilometers away. Tuvalu uses the Australian dollar as its sole official currency. Being extremely small and remote presents particular challenges, as well as opportunities, to Tuvalu. The set of feasible economic activities in Tuvalu are in large part determined by natural resources and economic geography. Tuvalu s GDP was around US$38.5m in 2012, and with onshore private sector value added limited, the public sector dominates GDP. Tuvalu is highly import dependent, even for basic items like staple foods. However, especially on a per capita basis, Tuvalu has rich natural maritime resources, and national income is supplemented by license fees from fishing activities in Tuvaluan waters, which contain valuable stocks, including tuna. Being a very small, undiversified economy, Tuvalu is highly vulnerable to economic shocks while its location and the nature of its low-lying coral atoll islands also make it vulnerable to natural disasters and adverse impacts of climate change. 2. With a dominant public sector and no monetary policy, fiscal policy is critical in determining livelihoods in Tuvalu. The annual government budget ranged between 75 and 100 percent of GDP between 2008 and A high ratio of government spending to GDP is almost inevitable in Tuvalu where the domestic private sector faces severe constraints and costs of public service provision are high. As a result, such levels of public spending do not necessarily reflect excessive spending, and indeed the quality of public services may not be at the level of larger countries, where economies of scale and lower transaction costs make it more cost-efficient to deliver public goods and services. The government budget had been financed by a combination of fishing license fees, donor financing and the earnings from the Tuvalu Trust Fund (TTF). 3. The Global Financial Crisis (GFC) severely affected Tuvalu, and led to both a deteriorating fiscal position and declining household income flows. The TTF, designed as a long-term investment fund, was established in 1987 and capitalized with contributions from donors and the Government of Tuvalu (GoT). The Consolidated Investment Fund (CIF) was subsequently also set up and is replenished by TTF payouts and can then be used for budget financing. The GFC led to a significant financial loss for the TTF, which according to its governance procedures would then not make payments to the CIF until it recovered the capital loss and met fund growth targets 1, which meant there may be no new funds for budget financing for approximately 10 years (which is the average time it has taken to reach the maintained value following previous shocks, stated in the Program Document for the first DPO). At the same time, remittances from seafarers, which had been a significant source of earning for Tuvaluan households, declined (from 17 percent of GDP in 2008 to 7 percent in 2012), as lower demand and industry changes mean that fewer seafarers were recruited from Tuvalu. Over a similar period, increases in commodity prices like oil and staple foodstuffs had also hit household budgets, leading to concerns about worsening poverty and hardship, especially in the outer islands of Tuvalu. 4. Fiscal consolidation was an important motivation for the program, along with the need to protect social service delivery and support reduced levels of hardship. In the years 2008 to 2011 the value of the CIF has been rapidly run down from AUD12m to only AUD1m, and 1 The TTF has a target maintained value which increases in line with the Australian Consumer Price Index. 1

13 due to losses incurred over the GFC, the level of budget financing from this source was expected to need to be significantly reduced. A focus of the program to achieve this was to limit costs from two large expenditure programs which involved sending people overseas for medical treatment and tertiary education, respectively. 5. The program supported one of the two priorities of a newly prepared Tuvalu Country Assistance Strategy (CAS): Building resilience against economic shocks. The focus of the program in supporting fiscal consolidation and improved management of public funds was in line with the CAS prepared November 4, To a lesser extent, some of the reforms, such as improvements in tertiary education outcomes, supported the other pillar by supporting regional and global integration (in this case, for labor markets). The CAS envisaged a DPO program to enable the WB to contribute to a strong policy engagement and coordinate with other development partners around a key set of reforms. The DPO program also marked the first operational engagement in Tuvalu, a relatively new World Bank member that joined the Bank in The DPO program served as an opportunity to build a closer partnership with the Government. 1.2 Original Program Development Objectives (PDO) and Key Indicators (as approved): 6. The original PDO was to support reforms that strengthen public financial management and improve the delivery of social services, while providing critical financing to enable the rebuilding of fiscal buffers. 7. The key outcome indicators were: i. Reduced variance in actual versus budgeted expenditure; ii. Improved management of revenue records; iii. Higher share of outer island projects implemented in-line with outer island plans for community development; iv. Increased expenditure allocation to primary and secondary health care; v. Increased expenditure allocations to primary, secondary, and vocational education. 1.3 Revised PDO and Key Indicators, and reasons/justification: 8. The high-level pillars of the PDO remained unchanged in the second operation, as strengthening public financial management and improved social service delivery, although the PDO no longer contains reference to providing critical financing to enable the rebuilding of fiscal buffers. The PDO was also elaborated in more detail with reference to specific objectives. The revised PDO in full is given below: 9. The first pillar of the program focuses on strengthening public financial management with the specific development objectives of improved commitment control, enhanced oversight of fishing revenue, and streamlined reporting of local governments (Kaupules). The second pillar focuses on improved social service delivery with the specific objectives of enhancing the efficiency of the tertiary health and education schemes to enable strengthened focus on primary and preventative healthcare and primary and secondary education (including vocational training), with increased access for women. 10. The key outcome indicators were: i. Stock and monitoring of arrears; ii. Improved management of revenue records; iii. Local government (Kaupule) accounts are compiled on a timely basis and made public; 2

14 iv. Increased rigor of the referral process; v. Increased budget allocation to primary & secondary education (including vocational); vi. Availability of gender disaggregated data on beneficiaries of the broadened Tuvalu Maritime Training Institute (TMTI) training program. 11. More detail on the changes to the PDO and key outcome indicators follows. 12. Strengthening public financial management: Two of the original three outcome indicators for this component ( reduced variance in actual versus budgeted expenditure and higher share of outer island projects implemented in-line with outer island plans for community development ) were removed and replaced with other indicators ( stock and monitoring of arrears and local government (Kaupule) accounts are compiled on a timely basis and made public ). No reason for the removal of the first outcome indicator and its replacement with one targeted on arrears is given in the documentation for DPO2, with the relevant text in the discussion of this policy area still referring to both, noting that supported actions are expected to lead to improvements in budget execution, reduce arrears and are likely to improve effectiveness in the use of public resources. However, from discussions, we understand that the indicator was changed to more accurately measure the impact of the prior action in DPO2, while the original outcome indicator was considered in retrospect to be too susceptible to other factors not related to the effectiveness of the actions. The outcome indicator on Kaupule project implementation was removed and replaced with one on Kaupule financial reporting. Again, no reason is provided in the documentation for DPO2 as to why the indicator was changed. However, it seems that the new outcome indicator measures the impact of the prior action in DPO2, while the dropped indicator measured the impact of the prior action in this policy area in DPO Improved social service delivery: Under the pillar, one outcome indicator was removed ( increased expenditure allocation to primary and secondary health care ) and replaced with another indicator in the same policy area ( increased rigor of the referral process ); a new outcome indicator was added ( availability of gender disaggregated data on beneficiaries of the broadened TMTI training program ) and the third original outcome indicator was kept, with changes to its wording that were not substantive. No reason for the removal of the fourth outcome indicator is given in the documentation for DPO2, nor is the inclusion of the new indicator of operation of the referrals board discussed. The relevant text of the PD for DPO2 continues to focus on intended cost-saving from the Tuvalu Medical Treatment Scheme (TMTS). The new outcome indicator more closely measures the implementation of the prior action in DPO2, while not measuring the impact of the supported reforms under this policy area in DPO1. The inclusion of a new outcome indicator that tracks the participation of women in TMTI training programs was added in order to measure the positive gender impact of some of the supported reforms, i.e. the prior action in the policy area supported in DPO Revisions to the outcome indicators (arising from the changes discussed above) are listed in the table below: Dropped indicators (numbering from section 1.2) 1. Reduced variance in actual versus budgeted expenditure 3. Higher share of outer island projects implemented in-line with outer island plans for No reason was provided in the PD for dropping this indicator, but the team indicated it was considered too subject to volatility outside of the scope of the reform to be an adequate outcome indicator for the policy area. No reason was provided in the PD for dropping this indicator. 3

15 community development 4. Increased expenditure allocation to primary No reason was provided for dropping this and secondary health care indicator. Revisions to original indicators (numbering from this section) 1. Stock and monitoring of arrears This indicator is discussed in the PD as measuring the intended outcomes of the commitment control action in DPO2, and acts as a replacement for the dropped indicator in this policy area. 2. Management of revenue records Minor, non-substantive changes in wording only. 3. Local government (Kaupule) accounts are compiled on a timely basis and made public This indicator is discussed in the PD as measuring the intended outcomes of the Kaupule financial reporting action in DPO2, and acts as a replacement for the dropped indicator in this policy area. 4. Increased rigor of the referral process This indicator is discussed in the PD as measuring the intended outcomes of the TMTM referrals board action in DPO2, and acts as a replacement for the dropped indicator in this policy area. 5. Increased budget allocation to primary & secondary education (including vocational) 6. Availability of gender disaggregated data on beneficiaries of the broadened TMTI training program Minor, non-substantive changes in wording only. This indicator is explained to capture gendersensitive results of the TMTI reforms supported in DPO Original Policy Areas Supported by the Program: 15. The development policy program was developed under the aegis of the government s national development strategy, Te Kakeega II National Strategy for Sustainable Development (TKII) and its mid-term review. This plan consisted of 8 strategic areas. The Government completed a mid-term review of TKII in 2011, recommending the following focus areas to support completion of TKII over : i. Good governance in both the capital and the outer islands; ii. Creating an attractive public sector to generate economic growth; iii. Maintenance of macroeconomic stability, through economic growth and stability, as well as private sector employment; iv. Providing quality health and education services that are equitable, balanced, cost effective, and that meet the needs of Tuvaluans; v. Sustainable development of Tuvalu, focusing on natural resources and infrastructure. 16. The two policy areas under the development policy program supported TKII, particularly the first, second and fourth focus areas, as well as the fifth to the extent that PFM reforms such as improved Kaupule project management support improved infrastructure development. Further detail on the policy areas is provided below. 4

16 1.4.1 Pillar I: Strengthening public financial management 17. The 2011 Public Expenditure and Financial Accountability (PEFA) assessment identified several weaknesses in PFM, including relatively weak treasury functions, bank reconciliation delays, and annual public account preparation delays. Among other areas, the PEFA also highlighted that greater public transparency of fiscal information would be supportive of greater public accountability and highlighted weaknesses in tax administration systems. Other analytical work available at the time supported these findings, including an AusAID Assessment of National PFM and Fiduciary Systems (ANS) and audit reports on the Tuvalu Whole of Government (TWOG) accounts. A performance audit on sustainable fisheries prepared by the Auditor General also highlighted weaknesses in the management of fisheries revenue. 18. Improving budget execution, transparency, and basic reporting: Given the importance of the public sector in a micro-state like Tuvalu, ensuring that public funds are well-managed to have the greatest possible impact is particularly important. Reforms in this area were intended to support good governance through a combination of more timely and transparent fiscal reporting and strengthened expenditure controls. Improvements in the timeliness of treasury reconciliations were planned to enable more frequent in-year budget monitoring, with the resulting information made public on a quarterly basis. The planned implementation of a simple commitment control system was expected to assist in basic cash planning over the fiscal year, and in particular to avoid the build-up of arrears due to overspending. 19. Improving revenue administration: In Tuvalu where the private sector is very small, a significant amount of domestic tax liabilities is generated by state-owned enterprises (SOEs). However, payment of tax dues by SOEs had been very weak, with limited documentation in both the government and the SOEs on the extent of tax payments that were unpaid, with ongoing tax audits of SOEs at the time of preparation of the PD identifying AU$1m in tax arrears. These problems have affected service provision due to the development of cross-liabilities. For example, the government had withheld payments such as scholarship fees to TMTI, leading to disruption of student intake, and had unpaid bills at the Vaiaku Lagi Hotel (VLH), affecting its ability to operate. In light of these issues, the program supported the settlement of cross-liabilities and sought to avoid the accumulation of tax arrears in the future. A second area of focus for improved revenue administration was timely recording of revenue in the Fisheries Department and reconciliation with the Treasury ACCPAC system. The lack of reconciled revenue information, particularly for major sources of revenue such as fishing licence fees, meant that the government had been unable to adequately verify if all fees due had been paid. 20. Outer island project and financial management: Around half of the population in Tuvalu live on outer islands, other than the capital. Local affairs and the implementation of local development projects on the outer islands are managed by local government councils, known as Kaupules. While a large proportion, reported as 40 percent of the capital budget, is dedicated to Kaupules, execution and reporting on this spending has been poor. Most Kaupule development projects are not clearly defined in a project proposal, and as a result many are under-budgeted and/or poorly designed. There is also a lack of project selection, project evaluation and financial reporting which reduces the likelihood that funds are being utilized as effectively as possible. This policy area supports government plans to improve Kaupule project implementation and timely financial reporting, of a kind that is consistent with limited capacity Pillar II: Improving the delivery of social services 5

17 21. The Tuvalu Medium Term Expenditure Framework Report (TMTEF) for Education and Health provided key recommendations on how to improve spending efficiency and equity, with a focus on the two large spending items for overseas service delivery in health and education, the TMTS and the overseas scholarship scheme, respectively. 22. Improving the management and efficiency of Tuvalu s overseas medical treatment scheme: While the TMTS is one of the most important health programs in Tuvalu, the program is spent on a relatively small number of patients. TMTS provides patients with tertiary care overseas which cannot be provided in Tuvalu. The TMTEF study found that increasing TMTS costs were attributable largely to non-dialysis cases and that a clearer entitlement framework and policy for utilization of TMTS should be put in place, to help ensure horizontal equity of access and maintain control of fiscal costs. This policy area aims to support a refocusing of healthcare spending to primary and preventative care, which is expected to be more effective in future in both supporting healthy lives and controlling fiscal costs. 23. Improving the management and efficiency of Tuvalu s overseas scholarship scheme and broadening other post-primary education opportunities: Given the very small size of the country, the PD notes that it is uneconomical to develop tertiary institutions. Therefore, the overseas scholarship scheme plays an important role in facilitating higher education for Tuvaluans. However, the program is costly and has a relatively high non-completion rate, reported in the first program document to be 32 percent and further 12 percent requesting extensions in order to complete. The reforms in this policy area support improved management of the program, while freeing up resources in the education budget to improve other levels of education, with a particular focus on secondary education and TVET noted in the PD. This policy area also encompasses a broadening of post-primary opportunities in the TMTS, with a focus on increasing opportunities for women. 1.5 Revised Policy Areas 24. The changes to the PDO had no substantial effect on the substance of the policy actions supported under the program in the second operation, with the prior actions in the second operation all remaining substantially the same as the indicative triggers in the first operation, and only undergoing minor changes in wording to increase their precision. 1.6 Other significant changes 25. Supplemental financing of US $3 million was provided after Tropical Cyclone Pam to help address immediate needs in the aftermath of the cyclone. There were also various changes in government and several by-elections held during the second operation. 2. Key Factors Affecting Implementation and Outcomes 2.1 Program Performance 26. The DPO program consisted of two operations, each with five prior actions. Both operations were approved broadly as scheduled, with only a two-month delay during preparation of each operation. The second operation included a supplemental financing which deployed IDA Crisis Response Window funds in the aftermath of the severe cyclone Pam, a category 5 tropical cyclone, which had multiple impacts on Tuvalu over the period of February and March The utilization of a supplemental financing enabled additional funds to be mobilized rapidly to support government-led recovery efforts following a World Bank led post-disaster needs assessment which 6

18 identified significant, urgent fiscal needs to support the recovery and reconstruction. All prior actions were completed Strengthening public financial management 27. The first policy area under this pillar targeted improved budget execution monitoring, transparency and controls and supported one prior action under each operation. The first was the approval of a new process to publish quarterly fiscal reports which assess key financial aspects of budget execution and provide explanation of the identified divergences. After the initial quarter, quarterly reports were no longer published but monthly reports were consistently prepared, disseminated and published. These monthly reports contained the same numerical information envisaged for the quarterly reports, but did not have any explanatory text. So while the reform has continued to be partially implemented through regular publication, it has also suffered from a less ambitious scope (of no descriptive analysis as in the first release reflected in the evidence for the prior action) which is likely to reduce its effectiveness in supporting public scrutiny of in-year public finance trends. Authorities note that the regular reporting is important for their internal use in monitoring the in-year cash and expense position, although there is little evidence of external scrutiny of the published reports. The prior action in DPO2 supported improved commitment control systems that would, inter alia, help the government to keep track of accrued, but not paid, expenses and avoid the accumulation of arrears. The system involved using the ACCPAC Financial Management Information System (FMIS) to generate electronic purchase orders which must precede any expenditure. While the system took 2014 and 2015 to be fully embedded, by 2016, the new procedures were sufficiently comprehensive to enable the Auditor General to remove their previous qualification to the TWOG accounts in this respect. While there are still pockets of problematic expenditure that are not in compliance notably TMTS expenditures and certain Cyclone Pam disaster response spending these expenditures are at least now being clearly identified and the reform is supporting the overall aim of progressively reducing the risk of arrears. However, unless these high-profile spending programs can be made consistent with the requirements, there is a risk that compliance with the system might unwind. 28. The second policy area under this pillar targeted improved revenue reconciliation and a lowering of revenue arrears, and included one policy area in each operation. Under the first operation, the government approved the settlement of cross-liabilities that had built up from two state-owned enterprises (SOEs) not paying taxes, and government in turn withholding fees due to be paid to the SOEs. These arrears were settled, although the cross-settlement did not fully settle one SOE s tax arrears (because they only settled an amount equal to the amount the government paid them in overdue payments, which was smaller in value). This SOE continues to have these tax arrears outstanding, and while the other SOE is now reportedly free of tax liabilities, a number of SOEs continue to have and accrue further revenue liabilities. In fact, by 2016, arrears due to the Inland Revenue Department (IRD) were higher than they were in So, while the crossliabilities were settled, the reform does not seem to have had much of an impact in broader improvements in SOE tax arrears. The second operation supported a full reconciliation of revenue data between Treasury and the Fisheries Department, to ensure that the two systems are updated and reflect the same revenue information. This process was carried out in late 2012 and certified by letter to be completed, along with a reconciliation table showing only a small discrepancy between the two sources of data. However, when the public accounts were published for 2013 and in all subsequent years, they continued to be qualified with a recommendation to reconcile the two sources of data, indicating that they were not reconciled to the satisfaction of the Auditor General. Therefore the evidence suggests that this prior action was not implemented as intended. As of November 2017, the government has just published a report on reconciled fisheries revenue, 7

19 although it is too early to confirm whether this has satisfactorily addressed the concerns of the Auditor General. 29. The third policy area under this pillar supported reforms aimed at improving the implementation and financial reporting of Outer Island (Kaupule) public expenditure. Under the first operation, a new policy was introduced that aimed to strengthen project management on the outer islands. The policy required improved reporting and due diligence at all levels of the project cycle. According to authorities, the policy has led to improvements in project proposals and project progress reporting, but finalization of the project, particularly providing financial acquittals and conducting evaluations, continues to be an issue. Financial acquittals are provided in some cases, albeit with delays. While the authorities report that project completion is generally good and completed as intended, they have not been able to carry out any formal post-completion evaluations due to a lack of budgeted resources for these activities. However, these additional requirements may be having a detrimental effect on budget execution, with only four Kaupule projects in total underway in 2017, which means that half of the Kaupules are not using any of their allocated grant resources in All this relates to the Special Development Expenditure (SDE) grant administered via the Ministry of Homes Affairs. However, there is a second grant for a similar purpose of nearly the same value that is administered directly by the Kaupule authority, and projects under this line reportedly do not adhere to the criterion set out in the policy. So while positive progress in some respects has been made, the policy is only being applied to part of the intended spending, and that is only partly compliant, indicating that there have been significant limitations in the effectiveness of implementation of this action. 30. The second operation supported the development of a new streamlined financial reporting framework for Kaupules and the completion of the financial accounts. While no accounts have been published yet, the new tools and guidance supported under the action have been implemented, with all Kaupules utilizing the key tools of an Excel-based cashbook and vote ledger. This has enabled the reconciliation and acquittal of monthly accounts. Most Kaupules are still struggling to compile their annual financial accounts and have received help from the Office of the Auditor General (OAG) to do so, but the improved availability of monthly revenue and expense data throughout the year has facilitated this process. As of November 2017, all Kaupule accounts for FY2012 to FY2014 had been completed and audited, but had not yet been submitted to Parliament, partly due to delays caused by the implementation of more rigorous audit oversight of the accounts. FY2015 and FY2-16 accounts are not yet completed. Overall, despite the clear capacity constraints, it seems that significant work has gone on to implement this reform and good progress has been made Improving the delivery of social services 31. The first policy area under this pillar targeted cost-efficiency gains in the Tuvalu overseas Medical Treatment Scheme (TMTS), and included one policy action in each operation the first of which instituted a series of measures aimed at improved oversight and cutting non-essential costs of the scheme, and the second to approve a new policy that reinstituted a Medical Board to oversee referrals to the TMTS, and also instituted a cost and length of treatment cap. Discussion with government officials indicates that cost-saving measures have been pursued, although not necessarily the ones supported in the prior action (for instance, the supported reform was primarily focused on addressing the scheme in Fiji, while increasingly patients are being sent to India). There is limited information on implementation of these cost measures, but a comparison of unit costs per overseas patient referral indicates that in 2016 the cost was significantly higher than in both 2011 and 2012, suggesting that cost-efficiency measure have had limited impact. Cost control remains an issue with an increasing number of patients and overseas locations, along with a lack of ex-ante cost controls meaning that expenditure often exceeds initial estimates. The Medical Board has reportedly 8

20 been meeting regularly, although minutes or other documentation has not been made available to the ICR team. The authorities indicate that while all referrals are scrutinized by the Medical Board, a large fraction (estimated by the authorities to be about 25 percent) of Medical Board recommendations are overruled by the Minister of Health. A total of 146 patients were treated under the overseas TMTS in 2016, compared to 116 in The second policy area under this pillar targeted cost-efficiency gains in the Tuvalu overseas scholarship schemes and an expansion of domestic vocational training opportunities, and included one policy action in each operation. The revised scholarship policies were aimed at reducing the cost involved in students repeating years of study, especially when they did not go on to successfully complete their study. These new rules, which included higher required pass rates to continue the scholarship have been maintained and have been effective in limiting unproductive time abroad and there has been a non-negligible number of students who have had their scholarships cancelled due to poor performance in line with the rules, so these reforms have been implemented as intended. The second operation supported the development of a new vocational training course to be delivered at the Tuvalu Maritime Training Institute (TMTI), aimed at preparing students for jobs on commercial fishing, particularly purse seine, vessels. While the course was launched as planned, it was discontinued shortly afterwards when it became apparent that graduates of the course were not finding the jobs that they had hoped to. The discontinuation of the course also meant that women were no longer able to enroll in any TMTI courses since the only other course is open to men only. 33. The current status of the prior actions under the DPOs is described in the table below: First Development Policy Operation Prior action, as stated in the Legal Agreement Status The Recipient, through its Cabinet, has Quarterly reports are not available, but monthly strengthened its budget execution and transparency reports have continued to be published to this day. through the preparation and publication of Monthly fiscal reporting is available with a varying quarterly fiscal reports, as evidenced through the time lag, but usually finalized within around a month Cabinet s Decision No.144/12 and 148/12 dated of month-end and is published on a government September 4, website and shared internally in government. These reports are numerical reports of expenditure and revenue to date, and do not include a descriptive analysis. The Recipient, through its Cabinet, has resolved to offset the tax related cross liabilities that have hindered operations of the Tuvalu Maritime Training Institute and the Vaiaku Lagi Hotel, as evidenced through the : (i) Cabinet Decision No. 24/12 entitled Offsetting Tax Payable by TMTI to Government with the amount receivable by TMTI from Government through the Ministry of Education dated September 6, 2012; and (ii) Cabinet Decision No. 61/13 entitled Offsetting of Government and Vaiaku Lagi Hotel Cross Liabilities dated October 1, The Recipient, through its Cabinet, has approved The audited statements of government accounts note that total revenue arrears in 2012 were AU$0.87m and increased in 2013 to AU$1.71m. Revenue arrears then fell back to AU$0.97m in 2015 and AU$1.09m in Consultations with the authorities indicate that while both TMTI and VLH cross-debts were settled, revenue arrears for VLH remained since they were higher than arrears for government services. As of 2017, these revenue arrears have still not been fully settled, and since the asset has now been sold to a different company, Funafuti Lagoon Hotel, it is unclear whether and from whom the debts are to be settled. A number of other SOEs continue to have revenue arrears, and efforts are ongoing to recover them, with a profitable SOE recently repaying close to AU$0.5m in revenue arrears. The new policy and associated templates entailed a 9

21 the Outer Islands Project Management Policy aimed at strengthening project selection, implementation and monitoring in its outer islands as evidenced through the Cabinet Decision No. 152/12, dated November 16, The Recipient, through its Cabinet, has: (a) on a pilot basis, implemented measures, designed to reduce health care costs and improve the quality of care provided through the Tuvalu Medical Treatment Scheme; and (b) submitted to its Parliament the report and recommendations from the TMTS pilot, as evidenced through: (i) the Recipient s Report entitled 2012 Tuvalu Medical Treatment Scheme Annual Report dated April 12, 2013; and (ii) the Cabinet s Decision dated June 14, 2013 endorsing said report and submitting it to the Recipient s Parliament. series of processes that Kaupules should go through during the project cycle, including preparing project proposals contained certain key information such as linkages to government priorities, providing project progress reports, and providing financial acquittals due from the project and project evaluations. From consultations with authorities, we understand that significant capacity-building has continued with Kaupules over the last few years to improve capacity to meet these steps. All Kaupules are completing project proposal and project progress reporting requirements for the Kaupule grant facility which is authorized by the Ministry of Home Affairs (currently AUD200,000 per Kaupule per year). However, Kaupules are not generally completing these steps for similarly-purposed direct grants (currently AUD160,000 per Kaupule per year), and for all projects, most Kaupules are not managing to provide sufficient financial acquittal reporting. Three Kaupules that have been part of a pilot project called LoCAL Local Climate Adaptive Living, which imposes minimum reporting requirements on Kaupules to receive financing have shown increased capacity to complete adequate financial project reporting. The recommendations from the TMTS pilot approved by Cabinet included: i) a capacity to appeal if a case is rejected; ii) allowing a caretaker to accompany the patient; iii) limit the time overseas by improving planning of timing of procedures, including a pre-assessment trip where necessary; iv) recruit a TMTS liaison officer in Suva; v) strengthen adherence to existing TMTS policies; vi) redirect funds from dialysis to both kidney transplant (where possible) and prevention services; vii) reduce accommodation costs by encouraging stays with family; use rental properties rather than hotels and have patients share rooms where feasible; viii) reduce consumables (like TVs, furniture, households items), repair and maintenance costs; ix) seek lowercost dialysis treatment in Nadi, instead of Suva; x) monitor that use of transport is for treatment purposes only. Discussions with the authorities indicate that while cost-saving measures have been pursued, it has been difficult to monitor costs at overseas locations. There are broader cost control issues, with no fixed set of treatment services agreed at the outset, and the scope of treatment can be expanded during the overseas stay. It is also noted that the TMTS is one of the few remaining government programs that does not comply with the purchase order requirements also implemented under this DPO program. While overall TMTS costs have spiraled upwards, from AUD2.3m in 2012 to 10

22 The Recipient, through its Cabinet, has approved the revised pre-service and in-service training scholarship policies in order to tighten the criteria for extension of scholarships as evidenced through: (i) the Cabinet Decision No. 40/13 dated June 11, 2013 approving the revised pre-service scholarship and training policy; and (ii) the Cabinet s Decision No. 3/12 dated June 4, 2012 approving the revised in-service training and scholarship policy. AUD4.6m in 2016 this includes both an expansion of the number of patients referred overseas and an increasing domestic referrals scheme. The unit cost of the overseas medical treatment scheme was AU$21,600 in 2011 and AU$15,100 in By 2016, the equivalent unit cost was AU$24,800, suggests that cost-efficiency measures may not have been effective. New policies were put in place to require a higher pass rate in order to qualify for an extension of the supported study period and continue to be monitored and maintained. At present, students who fail one (out of four) courses in a semester lose one quarter of their living allowance. If they fail two, or they fail another one the next semester, they lose all their living allowance until they have passed them again. If a student requests an extension, it is only granted if it can be completed in one semester in other words no more than four failed courses out of 24 (for a three year course). Data for the in-service scholarship program show that over the three years 2014 to 2016, 18 students had their scholarships cancelled due to failure (compared to 128 new awards), and for the pre-service scheme, there were 29 terminated compared to 110 new awards. Second Development Policy Operation Prior actions as stated in the Legal Agreement Status The Recipient, through its Ministry of Finance and The supported change introduced a new purchase Economic Development, has introduced centralized order system, managed centrally via the ACCPAC commitment control procedures to strengthen IFMIS. Purchase orders must be approved by a new treasury commitment and expenditure control, as Central Procurement Agency, who review and evidenced by the Financial Instructions 2014 Order approve based on a draft contract. The 2014 audit of No. GN 28 of 2014 made by the Minister of Finance and Economic Development on December 11, The Recipient, through its Ministry of Natural Resources, has: (i) updated all financial records in its fisheries management system; and (ii) concluded government accounts noted that some purchase orders were being issued only after commitments were made, and there was limited receipting. The 2015 audit repeated this issue, although noting an improvement in the management control environment due to the use of purchase orders. The 2015 audit report also noted concerns with control over the significant amounts of disaster relief expenditure, made from an account not part of the Treasury consolidated account or via ACCPAC and therefore not subject to this, and other, controls. However, in 2016 this qualification was removed from the TWG audit report, indicating that the purchase order system was operating comprehensively and to the satisfaction of the auditor general. Consultations with the authorities and OAG indicate that there are still areas of spending which are not fully compliant, such as the TMTS. The reconciliation for 2012 and 2013 was evidence for the completion for the prior action. However, the audit of TWOG government accounts for 2013 and 11

23 a full reconciliation of fishing revenues generated through its Fisheries Department for 2013 with the records of the Recipient s Treasury, to strengthen oversight of fishing revenues, as evidenced by the letter from the Minister of Natural Resources to the Association dated January 12, 2015 with a report of the fishing revenues reconciliation attached. The Recipient, through its Office of the Auditor General, has implemented a pilot streamlined Outer Island financial reporting framework to reduce administrative burden on the Kaupules and enable better tracking of funds, as evidenced by the letter from the Auditor General to the Association dated January 14, 2015 with completed piloted financial reports of selected Kaupules attached. The Recipient, through its Ministry of Health, has revised the Tuvalu Medical Treatment Scheme policy to strengthen the patient referral process and institute efficiency measures, as evidenced by: (i) the Cabinet Decision Memo of Special Meeting 17/14 held on October 23, 2014; (ii) the letter from the Attorney-General to the Association confirming the authority of the Minister of Health to revise the Tuvalu Medical Treatment Scheme policy; and (iii) the letter from the Minister of Health and Permanent Secretary for Health to the Association dated October 8, 2014 with the revised Tuvalu every subsequent year states that fisheries department revenue information and treasury revenue information is still not reconciled. Consultations with the authorities indicate that full reconciliation or both fisheries and IRD accounts receivable, on a commitment basis, has been problematic. However, all payments are made using the single government account through ACCPAC, so revenues are fully reconciled on a cash basis. Compounding difficulties for IRD is that their tax administration system, Revenue Management System (RMS), encountered technical problems in 2016 and as these could not be resolved, they have ceased using the system, and currently operate a manual, excel-based system. A new, excel-based cashbook and vote ledger was rolled out, along with new, clear guidance. The OAG has continued to provide capacity-building and support to the Kaupules, where capacity is limited in areas such as professional accountancy. Consultations with the authorities indicate that the monthly cashbook and vote ledger tools are now being used by all Kaupules to balance and acquit spending on a monthly basis, but most continue to struggle to compile their annual accounts. As a result, the OAG has supported them to prepare accounts and all accounts are completed and audited for FY2012 to FY2014. They are yet to be submitted to Parliament because some audited accounts are awaiting the formal approval of the Kaupule authority. The reforms have enabled a more detailed audit, and this round of accounts mark the first time the OAG has prepared and presented its audit opinion to each Kaupule authority. It is taking some time for some Kaupules to work through the issues and approve the audited accounts. No accounts have been submitted to Parliament or posted on the Auditor General s website since the approval of the operation. The last available accounts online are for FY2011. As of late 2017, audited accounts for years FY2012 to FY2014 are now reportedly ready but have not yet been submitted to Parliament. Through the TMTS policy revision, a medical board is empowered to advise the minister on referral cases. This actually reinstates a process that was previously in place but removed in The policy stipulates the board should meet weekly, be chaired by the Permanent Secretary of the Ministry of Health and include 6 other designated members. This process does not apply in the case of medical emergencies. The policy also places new limits on expenses incurred which include $50k total cost cap, maximum of 2 months support (unless an exemption is approved by Cabinet). The medical board is 12

24 Medical Treatment Scheme policy dated January 2014 attached. The Recipient, through its Ministry of Education and the Tuvalu Maritime Training Institute, has broadened vocational training programs to increase post-primary education opportunities, particularly for Tuvaluan women, as evidenced by: (i) the letter from the Minister of Education to the Association dated October 21, 2014; (ii) the minutes of the meeting of the Board of the Tuvalu Maritime Training Institute held on October 20, 2014, endorsing the introduction of a new curriculum for the training of seafarers for fishing vessels; and (iii) the student register of the Tuvalu Maritime Training Institute dated January 6, 2014 showing its opening to female enrolment. 2.2 Major Factors Affecting Implementation reportedly meeting regularly and all referrals are reviewed by the board, although a relatively large number of referrals are made against the advice of the board. The number of overseas referrals has increased from 116 in 2012 to 146 in The TMTI expanded its proposed offering of training to include a new course for purse seine fishing crews in 2014, with enrollment open for women for the first time for any course in TMTI. Three women enrolled, with one graduating. However, following the initial offering of this course, it was discontinued and women are no longer able to enroll in any TMTI courses. From consultations with TMTI and authorities, the graduates from the first course were not able to find work as hoped in the commercial fishing sector, while those enrolling in the general seafaring course continued to find other work, such as on container vessels. The sole woman graduate is a housewife and has never worked at sea since graduation. Due to a lack of interest in enrollments the course stopped being offered. TMTI management do not consider it appropriate for women to enroll in the general seafaring course, so they no longer have any course options open to them. 34. As it was noted in the program documents for both operations, Tuvalu is subject to significant volatility, and indeed a range of shocks of different kinds were observed over the course of the program. 35. The overall fiscal environment changed radically over the course of the program, driven by revenue growth, which is likely to have reduced focus on the fiscal consolidation elements of the program and led to some inconsistencies in commitment to aspects of the program. Total revenues more than doubled between 2012 and 2015, especially driven by a meteoric rise in fishing license fees as a regional cap and trade quota system led to sharp rise in the price of fishing licenses. This positive outcome for Tuvalu supported the high-level objectives set out in the program of improving service delivery quality and fiscal sustainability, although not via the mechanisms supported in the program. The first DPO program document highlighted that the Tuvalu Trust Fund (TTF) had suffered losses and would not recommence payments until it recovered lost capital and exceeded its maintained value. In fact, the TTF did recover its maintained value less than a year after the first DPO was approved and recommenced paying dividends, which have continued every year since in the order of AUD5 million to AUD9 million per year. These reversals in the revenue position for the country are likely to have, somewhat understandably, reduced the relative priority of expenditure consolidation and they led to an expansion of the fiscal envelope, with recurrent expenditure expanding by 60 percent between 2013 and 2015 and doubling over 2013 to The change in overall fiscal stance probably contributed to expansion in costs of many of the government programs that the DPO program targeted, like the TMTS and the overseas scholarship programs, and relaxed the focus on some aspects of the policy 2 Not including 15 cases of deceased persons brought back to Tuvalu in

25 actions which were framed as enhancing efficiency to protect service delivery in the face of tight cost control. Government of Tuvalu fiscal developments AUD million Total revenue Of which fish licences Of which TTF distribution Total expenditure Of which recurrent Budget balance Source: GoT budget documentation for 2011 to 2015, GoT staff correspondence for The burden of fully implementing the reform program was high given the very thin capacity of the government, especially in the context of a broader set of reforms supported by other budget support donors. Despite acknowledgement of and emphasis on thin capacity in the program documents, and a relatively small set of 5 prior actions each year, it still seems that some of the reforms imposed higher capacity requirements than were available. For instance, the implementation of regular fiscal reporting, new commitment controls, revenue reconciliation and outer island project and financial management all seem to suffer from a lack of capacity to fully implement. In some cases, anticipated capacity-strengthening did not occur (for instance the staff position responsible for fisheries revenue reconciliation anticipated in the program was not maintained). In others, there may have been a lack of full appreciation of the capacity requirements (for instance, it did not seem clear from the PDs that compliance with the new commitment control would also require compliance with new procurement procedures). Limited capacity was also stretched by the competing demands of different donors budget support programs which were only loosely coordinated under a Policy Reform Matrix (PRM) which contained many more required actions than those supported under the DPOs alone. The full PRM detailed in Annex 3 of the first DPO PD contains more than 30 actions targeted for completion over the course of around two years. 37. While the analytical underpinnings of the operation were sound, gaps in the evidence base and a lack of data availability meant there was a weak understanding of program implementation in some cases. Many of the supported actions were underpinned by sufficient analysis of both the problem they sought to address and the appropriateness of the intended reforms. However, certain areas seemed to have a weaker evidence base notably the Kaupule project and financial management reforms and the TMTI reform, which both sought to address clear problems, although it was less clear that sufficient feasibility analysis for the proposed solutions has been conducted. In these cases, the reforms proved to not be effective in supporting the intended objectives, and it is likely that with further work upfront, this might have been anticipated. In many areas, there was a lack of availability of data to monitor outcomes throughout the period, which impeded management of the implementation process. 38. Changes in government and the launching of a new development plan led to some delays and disruptions on implementation, and changes in focus. There were a series of changes of government over the course of the operation, which led to some delays and some refocusing of government priorities. Following the death of the Finance Minister, a by-election was held in June Further by-elections were held in September 2013, January 2014 and September General elections were held in March While successive governments supported the PRM 14

26 process, focus at times shifted away from implementation of specific actions due to other priorities, especially since the DPO supported actions form only a small subset of the total actions that are set out under the PRM, besides other priorities of government. 39. The impact of Tropical Cyclone Pam diverted some attention from the program, and affected the implementation of some reform programs. This major cyclone hit Tuvalu just following preparation of the second operation, so was not foreseen in the design of the program but had a significant bearing on government priorities during the supervision period. While the government remained engaged on the program, the required response to the cyclone added pressure to the thin capacity of government. It also presented risks to the implementation of those aspects of the program that were related to increased controls over government spending such as new commitment controls and increased Kaupules project management requirements which generally required more capacity and time to complete. In the context of an urgent need to execute spending, this meant that significant amounts of funding did not comply with these new requirements. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization: 40. Design: As presented in the first operation of the program, the results framework had a number of limitations in design which would have impeded effective monitoring and evaluation. For instance: The variance in composition of expenditure indictor was incompletely specified, because it noted that the measurement should exclude lumpy items but did not define how this would be measured. In retrospect, the indicator was considered by the team not to be a good outcome indicator for the targeted reforms, as it would have been subject to too many external factors not related to the success of failure of the prior actions. The indicator on revenue reconciliation targets both Department of Fisheries and the Inland Revenue Department (IRD). By including IRD, it encompasses both prior actions but has a significantly broader scope than the immediate outcome of the reforms supported under DPO1, which only targeted the settlement of cross-liabilities for two SOEs. This broad measure makes the targets in the area substantially more ambitious than the direct implementation of the supported reform alone. The indicator relating to outer island project management was incompletely specified, lacking a baseline value which suggested that a data source to measure the indicator had not been identified. Two results indicators which were based around increased budget allocations to certain aspects of health and education spending were only weakly related to the implementation of the prior actions. In both cases, the supported health and education actions were related to improved management of specific programs which were highlighted as having high costs. For the health sector reforms, the supported actions aimed to both increase effectiveness and reduce costs of the TMTS, yet the indicator targeted an increase in primary healthcare spending, which does not track improved effectiveness of the TMTS at all, and is only weakly related to reduced costs of the scheme. In the case of education, the supported reforms included both cost-cutting measures for expenditure programs not part of the outcome indicator, and also expansion of programs that were part of it, so the outcome indicator seems to better reflect the outcomes of the supported reforms. 41. The results framework was substantially redesigned in the second operation, with three (of five original) outcome indicators dropped, and four new outcome indicators added. While the revised results framework presented design improvements, there remained problems of design even at the outset of DPO2: 15

27 The added indicator on expenditure arrears was not clearly defined, with no definition of the concept of arrears given or referred to, and the intended data source for monitoring was not available at the time. 42. In general, a relatively large number of the outcome indicators were based on information that was not readily or publicly available with the information for five of the six indicators in the revised results framework based on information sources that were not based on pre-existing data sources (arrears, revenue reconciliation, TMTS referrals, scholarships, TMTI gender-disaggregated data), and in the eventuality almost all of these data sources were not made available as anticipated. 43. Implementation: The Ministry of Finance and Economic Development were identified as the responsible agency for collecting and reporting on outcome indicators. The PD indicates that they will provide regular reports on the PRM to the World Bank and other development partners against agreed timelines. Regular reporting and interaction carried on over the course of the program, aided by the use of the PRM and a continuation of the DPO engagement in a new program following the completion of this one. The PRM group convened on a relatively regular basis - at least twice a year, if not quarterly at times. While the PRM arrangement was a useful tool for discussion of reforms, there was relatively less focus on the identified results indicators, with data for many of them not seeming to be prepared or monitored. One reason for this was that as the annual PRM cycle progressed, reporting and results indicators changed in line with the latest round of reform, and did not consistently report on the results and implementation of previously supported reforms. 44. Overall, the quality and reliability of the data presented in this ICR is subject to some uncertainty, in particular: The concept of arrears is not clearly defined in the PDs, and was not subsequently defined by the government as seems to have been anticipated at the time of preparation. Therefore, while data are available for accounts payable and accrued expenses, it is not clear whether these accords with the intention of the results indicator. TWOG accounts have a disclaimer on the revenue data noting that it cannot be adequately reconciled due to a lack of information from fisheries and IRD. As a result, it is not possible to assess the deviation in reported revenue data as specified in the outcome indicator. The expected source of TMTS referral data a TMTS annual report was never repeated after the first year, and information on the number of referrals reviewed by the Medical Board is based on ad-hoc data and informed estimates from Ministry of Health officials. 45. Utilization: It is not clear whether the outcome indicator data were used to inform decisionmaking and resource allocation. An ISR was carried out for DPO1 and approved on April 8, This ISR did not contain any data on the outcome indicators other than what was presented in the PD. 46. Reports prepared as part of a development partner assessment of the PRM framework indicated that while some monitoring data has been of value to the government and development partners, such as more regularly monthly fiscal reporting, other indicators may not have been sufficiently well-designed to provide a robust basis for policy-making. For instance, it was noted that the increase non-salary expenditure on primary education of the 2012 outturn by at least 5%, does not reflect improvements made to primary education by the increased number of teachers engaged through the Fiji voluntary teachers program. 2.4 Expected Next Phase/Follow-up Operation: 16

28 47. Following completion of this DPO series, a new programmatic series of two DPO operations was commenced, with the first operation of the new series (called Third DPO ) approved by the Board in December 2016, and the second Fourth DPO under preparation. This program continues to support policy reforms under a pillar of improving social service delivery, and has a second pillar entitled improving macroeconomic stability which generally includes a broader range of policy areas than the improving public financial management under this program, although continues to have policy reforms that support increased control of expenditure items, focused on wage bill management. 48. While a follow-up program is already underway, it is clear from the assessment of this DPO program that a number of reforms are still to be fully implemented. Subsequent engagement should focus on continuing to monitor these areas and try to address issues in implementation where possible. Given the pace of reform is slower in thin capacity countries, and development assistance needs are higher, the preparation of future DPOs or TA programs should consider whether there are appropriate areas of support that would facilitate the continued implementation of these reforms, or complementary measures that would strengthen their impact are a first priority before moving on to tackle other areas. 3. Assessment of Outcomes 3.1 Relevance of Objectives, Design and Implementation Relevance of Objectives: Substantial 49. Relevance of objectives is rated as substantial. The overall objectives of improving public financial management, social service delivery and rebuilding fiscal buffers were relevant at the time of design, and in general they continue to be relevant. Since the beginning of the DPO program, the World Bank Group has launched a new Regional Partnership Framework (RPF) for FY2017 to FY2021 cover Tuvalu and eight other small Pacific Island countries. The RPF has four focus areas: i) Fully exploiting the available economic opportunities; ii) enhancing access to employment opportunities; iii) protecting incomes and livelihoods; and iv) strengthening the enablers of growth opportunities. Against this framework, the DPO program remains relevant, by focusing on supporting improved employment opportunities through education, supporting health outcomes which help to protect incomes and livelihoods and strengthening enablers of growth such as improved fiscal management. The GoT has launched a new National Development Plan, Te Kakeega III (TKIII) which covers the period 2016 to The plan contains 12 strategic focus areas, including good governance, health and social development, Kaupule development, education and employment. The DPO program remains very relevant to these priority areas. 50. Relevance of design and implementation is rated as moderate. The design of the original program was highly relevant and reflected well the high-level PDO, although the result framework had shortcomings. There were no substantive changes to the indicative triggers in the second operation which, while indicating the strength of the original project design, also suggests some lack of flexibility in adjusting to relatively rapidly changing circumstances, or to implementation challenges of the reforms in the first program. In retrospect, some policy reforms such as in the areas of revenue and the TMTS program could have warranted a refocusing on actions that were either more closely related to the follow-up of reforms in the first year, which still required considerable action to yield the intended results, or a reformulation. While not clearly stated in the two pillars, a strong cross-cutting theme of the program, as explicitly stated in the formulation in the PDO of the first operation, was fiscal consolidation, and the framework of most of the prior actions were framed to serve the dual purpose of reducing costs while protecting service delivery. The 17

29 dramatic change in fortunes (for the better) for Tuvalu over the period with large increases in fiscal revenue meant that these careful cost controls were probably less important to the government, who rapidly increased the fiscal envelope. There was no real change in the program in the second operation to account for this, other than the reframing of the PDO to remove the third component on rebuilding fiscal buffers, which may have been a missed opportunity to maintain relevance. 3.2 Achievement of Program Development Objectives Overall rating: Moderately Unsatisfactory 51. Only two out of six outcome indicators are assessed to be met, one partially met, one unassessable and two not met. This indicates relatively weak performance against the results indicators, with a lack of results in those areas arising from a combination of insufficient capacity to follow-through subsequent steps to achieve results, mixed or limited stakeholder ownership for the reforms, and weaknesses in planning of expected results. The assessment of results has been affected by a weak results framework, with a lot of data sources not available as envisaged, despite considerable revisions in the second operation. These problems suggest relatively weak monitoring and evaluation at all stages design, implementation and utilization, where the authorities may have had less, or inconsistent, ownership than would be desired. Objective 1 Strengthening public financial management Rating: Unsatisfactory 52. Under this pillar, one outcome indicator was met. One of the two outcome indicators that were removed after the first operation was partially met, on Kaupule project management, while the other, on expenditure variance, was not met, although it was affected by the impact of Tropical Cyclone Pam. Indicator 1: Stock and monitoring of arrears. Met. 53. This indicator targets the production of a statement of arrears and for arrears to be below 2 percent of total expenditure. The definition of arrears is not clearly specified in the PD, with the baseline value for arrears comprising a discrete estimate of arrears of vessel maintenance plus the value of whole of government accounts payable, although on the latter the PD notes it is unknown what proportion of this is in arrears, the team has included the full amount but will confirm. This clarification was not made to our knowledge and there is no other known definition of arrears in Tuvalu 3. The results indicator targeted a new arrears statement be produced to define and measure arrears. In the 2014 TWOG audit report, the note to the accounts on Current Liabilities was expanded to include Accounts Payable, and further expanded to include Accrued Expenses in the 2016 audit report. While in 2014 and 2015 the audit opinion was qualified based on concerns around the accuracy of Accounts Payable, that qualification was removed in the 2016 audit. As a result, the current liabilities presented in the 2016 audited TWOG accounts are considered to be an adequate basis for measurement of potential arrears, or an upper limit as to what the true value of arrears is likely to be (given that some items, such as accrued expenditure may not be formally considered arrears, or other accounts payable may not have passed an agreed payment due date). The TWOG audit for 2016 reports total accounts payable and accrued expenses of AUD 1,164,976, 3 While arrears are referred to in the GoT Treasury Instructions, which indicates that they should be reported annually in the annual budget statement, no standard definition of the term arrears is given, nor is there reporting of arrears in any recent budget statement. 18

30 which is 1.7 percent of total expenditure, and no longer raises an issue of lack of reconciliation of purchase orders 4. Based on the availability of data supported by an unqualified audit opinion in the report, and numerically meeting the target, this indicator is assessed to be met. Indicator 2: Management of revenue records. Cannot be assessed. 54. This indicator targets the compilation of treasury records at least quarterly, and that they do not deviate by more than 10 percent from those of the revenue and fisheries management systems, with the baseline being that a full reconciliation is not available. The audited TWOG accounts continue to state that reconciliation between the different sources is not being done, and indeed in 2016, the audit report contains a disclaimer to that effect. The Auditor General states in successive years up to the latest the Account Receivable balance is understated in the Financial Statements. This is due to outstanding Taxation, Customs, Fishing Licenses and Wharfage revenues not being recognized as Receivables in the Financial Statements. Indeed, they state that in 2014 and 2015, the Revenue Management System (RMS) was not updated for revenue transactions at all, meaning that no reconciliation was possible and in 2016 that the lack of information constituted a significant limitation to the scope of the audit. The lack of availability of information from IRD and Fisheries means that reconciliation as set out in the results indicator cannot be done, and this indicator cannot meaningfully be assessed. Indicator 3: Timely publication of local government (Kaupule) accounts. Not met. 55. This indicator targets the preparation and publication of finalized Kaupule reports, based on a new streamlined financial reporting framework that was supported under DPO2. There are eight outer island Kaupules, and the target for the indicator is that ¾ of annual Kaupule accounts are compiled within the statutory time limit and published on the Auditor General s website. This means that at least six Kaupules should have posted their FY2012 (end 31 March 2012) financial accounts by 31 December 2012; FY2013 by 31 December 2013; FY2014 by 31 December 2014 and FY2015 by 31 December At the time of assessment in November 2017, no Kaupule accounts for FY2012-FY2015 had been posted on the website, meaning that this indicator has not been met. Objective 2 Improving the delivery of social services Rating: Moderately unsatisfactory 56. Under this pillar, one of three outcome indicators is assessed as met, one as un-assessable and one is not met. The indicator that was removed, targeting spending on primary and secondary health, was met. Indicator 4: Increased rigor of the Tuvalu Overseas Medical Treatment Scheme referral process. Partially met 57. The indicator targeted an increase from zero of referral cases scrutinized by a multidisciplinary Medical Board to at least 85 percent. The evidence base for this indicator is particularly poor, with no published or prepared data available. Following the first year, an envisaged TMTS Annual Report which was to be the basis of tracking this indicator was never replicated, and while minutes of the meetings may be produced, they have not been made available 4 The numerical target was also met in 2015 with accounts payable of AUD 960,496 while a further AUD 6,147 was unaccounted aged purchase orders, together equivalent to 1.6 percent of total expenditure in that year, although the accounts in this year were still qualified in this respect. 19

31 to the ICR team. Based on consultations with the authorities and figures provided by them, it has been possible to make an assessment, although it is subject to uncertainty. Data are not available for 2015, so the assessment is based on 2016 instead. The number of people treated under the overseas component of the TMTS was 116 in 2012, and increased to 146 in 2016, with a further 15 cases of deceased persons requiring transportation back to Tuvalu in The Medical Board has continued to meet regularly over the period, and consultations indicate that all referrals are reviewed by them, although the team did receive contradictory advice on this by different counterparts consulted. An estimated 25 percent of referrals are rejected by the Medical Board, of which consultations indicate that most of these decisions are subsequently overruled by the Minister of Health and these patients do take part in the scheme. Due to the lack of the anticipated data sources to track this measure and the subsequent uncertainty about the assessment, much of which is based on verbal consultation, and the fact that the Medical Board seems to be playing a weaker-thananticipated role in supporting selectivity based on need, this indicator is assessed to be partially met. Indicator 5: Increased budget allocation to primary and secondary education (including vocational) through savings on tertiary scholarships. Met 58. This indicator targeted an increase in the spending on primary and secondary education from a baseline of AUD 2.1m in 2012 by at least 5 percent on average per year. The definition of primary and secondary education was as determined in the Tuvalu Medium Term Expenditure Framework estimates report in 2012 and is based on classification of expenditure at the program level. In assessing the indicator, the baseline was found to be somewhat different, at AUD 2.8 million. Expenditure increased to AUD5.9 million in 2015 (including AUD2.2 million special development expenditure, or SDE, which is generally for investment) which corresponds to an annual average increase of 28 percent. Even when excluding the SDE expenditure in 2015, the annual average increase in 5 percent. Although actual data for 2016 is not yet available, using budget appropriations as estimates for outturn, the annual average growth of expenditure is still above 5 percent. Therefore, this indicator is found to have been met. Indicator 6: Availability of gender-disaggregated data on beneficiaries of the broadened TMTI training program. Not met. 59. The indicator targeted both beneficiaries of the new commercial purse seine fishing course supported under DPO2, and the admission of women into the course. It is not made clear in the PD how beneficiaries is to be interpreted whether it is enrollees or graduates, but since the DPO2 prior action identifies a data source of enrollments only, it is interpreted as that. Following the first year of operation, the course was discontinued due to a lack of interest from students. In the first year, there were three women and 59 men enrolled in TMTI courses. Of the three women, one graduated in December From 2015 onwards, since the purse seine course was no longer offered, there were no further female enrollments or graduates in 2015 or The interpretation of the indictor is that gender-disaggregated data should be available for the assessment year 2015 and for this year and later years there is no gender disaggregated data. While it could be argued that since it is known that there are no women allowed on the course, gender-disaggregated information is available (100 percent men) since this is trivially true, and the only genderdisaggregated data was provided as evidence for completion of the prior action, it is not considered to even partially meet the target. 3.3 Justification of Overall Outcome Rating (combining relevance, achievement of PDOs): Rating: Moderately unsatisfactory 20

32 60. The overall moderately unsatisfactory rating is driven by poor performance against the measured outcome indicators, despite the program being moderately relevant. There are major shortcomings in the achievement of the PDO, as described by the outcome indicators. Only two of six indicators are assessed to have been met with a third partially met, which makes it very difficult to argue that the operation has been effective in supporting either component of the PDO. However, despite that poor performance, the program was generally relevant in its stated PDO, with the highlevel objectives of strengthening PFM and improving social service delivery still important for Tuvalu. It is clear that some supported reforms have been implemented and are considered to be important to the stakeholders involved in Tuvalu, such as regular fiscal reporting, commitment controls and new scholarship policies. There has been substantial progress in implementation in other areas, but they have been affected by weak capacity relative to what is required for the reforms, with capacity-building and implementation still ongoing for Kaupule reforms for instance. In other areas, it is less clear whether the policy reforms enjoy continued support and are having the intended effects, such as the revenue, TMTI and TMTS reforms. Overall, many reform areas suffered setbacks, which are a combination of over-ambition of the program given the limited capacity that was well-known at the time of design, at times a seemingly weak understanding of what would be required to achieve the intended results, and changes over time particularly the sharp upswing in revenue, a natural disaster hitting and a series of political events. While some of these factors could not have been anticipated at the outset, quite a number could have been mitigated at the design stage. 3.4 Overarching Themes, Other Outcomes and Impacts a) Poverty Impacts, Gender Aspects, and Social Development 61. The second pillar of Strengthening Public Service Delivery was generally intended to reduce inequality, although the mechanism by which this would happen was not very clearly set out. DPO1 set out a rationale for a rebalancing of limited expenditure away from very expensive overseas tertiary health and education schemes that only benefitted very few, towards primary and secondary in-country services which could support improved human development outcomes for the whole population. However, the link was less clear because the prior actions relating to the overseas TMTS and scholarship scheme were also claimed to not be focused on reducing the utilization of those schemes, but that savings would be made by efficiency improvements, making it less clear how significant a redistributive effect the reforms would actually have. In the eventuality, the imperative for maintaining a fixed, binding fiscal envelope was removed as large increases in revenue flowed in, meaning that all levels of education and health saw increases in spending. It is hard to argue that the supported measures had a significant impact in their own right in supporting broad-based improvements in human development outcomes. The reforms aimed at expanding incountry tertiary vocational training, especially for women, would have had a strong, positive impact but unfortunately it was short-lived and only one woman graduated, and she and her male fellow graduates were unable to find related employment following completion of the course. 62. Public financial management reforms may be contributing indirectly to improve social development through better public sector effectiveness. While it is difficult to say with certainty because of the indirect nature of the link, improved fiscal transparency and commitment control may support more effective use of public funds through greater scrutiny and control, and in doing so, increase the impact of social spending in reducing poverty and inequality. In a slightly more direct way, the reform to improve Kaupule project management could be expected to lead to better results on the ground in the outer islands, although the evidence is not available to assess this. b) Institutional Change/Strengthening 21

33 63. A number of reforms in the program supported relatively ambitious institutional development for Tuvalu. If these reforms can be sustained given limited capacity, they would mark a strengthening in longer-term institutional development, particularly in relation to public financial and investment management. In particular, commitment control changes and revenue arrears settlement supported in DPO2 and the associated aim to better tackle both expenditure and revenue arrears would be important to better manage the public finances, and the Outer Island project management policy would improve the oversight and effective of public investment. However, a number of these reforms have been slower to successfully implement than may have been initially envisaged and may require continued focus at a high, policy level to bed down as long-term institutional improvements. c) Other Unintended Outcomes and Impacts 3.5 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops 64. N/A 4. Assessment of Risk to Development Outcome Rating: High 65. The three risks highlighted in the DPO2 PD continue to encapsulate well the high risk to the development outcomes for this program: Limited capacity; Government ownership; and external and domestic shocks. In a small country environment, all these risks interact with each other and contribute to a heightened risk that the supported reforms will not have a lasting impact for any or all of these reasons. 66. While capacity constraints were noted to be considered in the design of the program, continuation of certain reforms and their full implementation still remains very ambitious. In fact, most of the actions entailed additional capacity requirements from a range of different actors, which raises the question as to whether DPOs in such a constrained environment should be targeting increases in administrative burdens without a detailed risk assessment. The government has already struggled to continue a number of reforms following the initial completion of a prior action, such as adequate revenue reconciliation, arrears monitoring, Kaupule project management and TMTS reporting. It remains unclear whether institutional improvements for some of these will become embedded and the required capacity maintained. It should also be borne in mind that capacity is particularly stretched due to the annual nature of the PRM reform program and the much larger number of reforms supported by all donors. Despite their good intent, this may simply be too much for GoT to handle in a short space of time, and by constantly trying to keep up with new reforms, certain completed ones lag in implementation. 67. Government ownership is critical to achieving outcomes, and multi-operation development policy programs should exercise in-built flexibility to maintain their relevance in the face of a changing external and political environment. It is clear that the Government has supported the preparation of the DPO operations and the completion of the policy framework in each year. For some reforms, such as the expenditure reporting and controls, led by MFED, there has been strong ownership. Other reforms, such as revenue reconciliation and Kaupule reporting have clearly been led by the center of government, but ownership across government, including in key implementing bodies is less clear. TMTI, TMTS and revenue reforms seem to have suffered from weak commitment to the policy reform during implementation. Other than stroke of the pen approvals, much of the work involved in policy and institutional development has, perhaps 22

34 necessarily in light of capacity constraints discussed above, been done by external advisors. As a result, such reforms will generally carry a heightened risk that they do not have bona fide government ownership, and this may lead to a lack of focus on follow-up and implementation of reforms, especially when there are many new priorities that also need to be addressed. 68. External and domestic shocks have affected implementation. The relatively volatile nature of politics, as well as a high vulnerability to economic shocks makes it more likely that attention will be diverted when priorities change. Ironically, over the period so far, it may have been positive shocks (such as massively increased revenue receipts) that have presented the biggest risks to the reforms, particularly related to ensuring cost-efficiency and expenditure controls. A variety of other shocks have presented themselves as discussed above, including a change of government and a major natural disaster. In such a small, undiversified country, shocks of any nature will tend to have a disproportionately large impact, and so even though systems to boost resilience are in place such as the TTF and CIF the risk of further shocks adversely impacting the achievement of the PDO remains high. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance a) Bank Performance in Ensuring Quality at Entry Rating: Moderately unsatisfactory 69. While the program development objective and supported policy areas of the program were generally relevant, there was some ambiguity in the overall fiscal strategy which led to difficulty in clearly elaborating the likely impacts and outcomes of prior actions and the general narrative. In the design of the first operation, it was somewhat unclear whether the operation was supporting fiscal consolidation or not. At some points the PD emphasized the severe impact of the GFC, the precarious nature of the public finances and that the TTF has not recovered lost value. However, the speed at which the fiscal outlook reversed, a matter of months after the approval of the first operation, raises some questions as to whether this really could not have been foreseen in the preparation of the operation, and whether the macro-fiscal outlook could not have been more accurately presented. 70. The identification of the policy framework was generally good, followed many good practice principles for DPOs, and was mostly based on sound analytical underpinnings. However, there were notable gaps, and as discussed above some supported reforms seem to be largely rationalized on the basis of fixing a problem, but with little evidence presented on the feasibility of the proposed solution in the context, especially whether they would work in the political and capacity environment. These include the Kaupule project management reform, the TMTS reforms, and revenue administration reforms. Policy reforms which subsequently failed, such as the TMTI reform, seem to be based on limited evidence of whether the reform was likely to have the intended effect, which at least should have been highlighted as a high risk at the outset. This points to patchy and in places weak analytical underpinnings, with insufficient evidence to support the feasibility of proposed reforms on TMTI, revenue administration and, to an extent, Kaupule reforms. 71. There were also clear weaknesses in the monitoring and evaluation framework from the outset, including numerous problems with the results indicators and poor recording of 23

35 documentation, with some of the evidence of prior action completion unable to be located by the ICR team 5. While the results framework was comprehensively overhauled in the design of the second operation, in retrospect it continued to exhibit problems in design, with some ambiguities and some data sources that were not available. b) Quality of Supervision (including M&E arrangements): Rating: Unsatisfactory 72. The Bank team carried out relatively frequent missions, jointly supervising DPOs and preparing new ones, although there was less oversight during the supervision period of DPO2. The DPO team conducted in-country consultations on six different missions in 2013, four missions in 2014 and two in The triggers for DPO2 were converted to prior actions without any substantive changes, which suggests that they were well prepared in advance. However, with the results framework in DPO2 comprehensively overhauled and significant changes in the government s fiscal position as described in section 2.2, it is a little surprising that there were no changes to the triggers that might have been justified to maintain program relevance in the context. The rationale for the changes to the results framework were not explained in the program documentation, and in the eventuality, the revised results framework proved to also have weaknesses in design. An ISR prepared for DPO1 hints at key weaknesses in supervision, as it does not contain any updated data on the results indicators. There seems to have been relatively little focus on ensuring that the impacts of the reforms were identified and measurable, with much of the data proposed for results indicators ill-defined or not available from the source (if a source was cited). c) Justification of Rating for Overall Bank Performance: Rating: Moderately unsatisfactory 73. With quality at entry rated as moderately satisfactory, and quality of supervision is rated as unsatisfactory, overall bank performance is rated as moderately unsatisfactory. 5.2 Borrower Performance Rating: Moderately satisfactory 74. The government showed strong support for preparation of the DPOs. Through leadership of the PRM process and good engagement with the Bank team during a series of missions, established the program in each year. This preparation involved a range of different actors across government, led by the Ministry of Finance and Economic Development, but with involvement of representatives from the Prime Minister s Office, the Ministry of Natural Resource, the Ministry of Health, the Ministry of Education, the Ministry of Home Affairs and the Office of the Auditor General, as well as the Tuvalu Trust Fund Secretariat. While in many cases, there was sustained commitment to continue and implement the supported reforms, even where there were challenges given the limited capacity, in some cases there seems to have been less consistent implementation over the period, and some reforms have likely suffered from a lack of focus following the completion of a prior action. Commitments made regarding the preparation of reports and data seem not to have been followed through, and the weaknesses in the assessment of results of this program reflect a lack of 5 Although the team was able to retrieve duplicate documentation of the evidence from the client. 24

36 coordination of monitoring and evaluation on the part of the government as well as relatively weak supervision on the part of the Bank team. 6. Lessons Learned 75. Focus on fewer actions or less capacity-intensive actions to have a greater impact overall. In a micro-state with well-known severe limits on capacity, the reform agenda has proved far too challenging, undermining, or threatening to undermine, many of the results. Focusing on a smaller number of actions and seeing them through would improve the impact of DPO programming. While the number of DPO prior actions, at five per year, is not necessarily the problem, their implementation required in many cases further actions (e.g. defining and preparing an arrears report; annual publications on the TMTS). In addition to that, the DPO sits under a coordinated reform matrix which contains many more actions, all supported by different development partners budget support operations. Taken together, the extent of this reform program is evidently entirely unmanageable. A more focused program would also help to ensure that sufficient preparation and due diligence is carried out for each reform area, and enable more oversight of supervision from both the Bank and client side. 76. Stronger analytical underpinnings and understanding of the implementation process. While all policy areas are supported by analytical underpinnings of some form, their detail and targeting vary substantially, and in a number of cases, proposed solutions have proved to not be appropriate in the context, which with better planning, could have at least in some cases been foreseen. The TMTI training course, while laudable, seems to be based on weak analysis of the constraints to employment for Tuvaluan seafarers and the viability of the proposed solution, and there is no evidence presented that it was part of a broader program to support new destinations in commercial seafaring. As a result, it comes across as a high-risk, opportunistic reform which might have been better planned. 77. Improve development partner coordination. The DPO program was part of a coordinated budget support framework which also included the Government of Tuvalu, Asian Development Bank, Australia DFAT and New Zealand MFAT. The intent of this coordination framework is valuable, yet in practice, at least from the perspective of the World Bank program, effective coordination has been weak. Donors have generally not supported the same reform actions, meaning that the total number of actions that budget support is dependent on is much larger than the set presented in the DPO program alone. While there are regular forums for coordination including roundtable meetings, audioconferences and PRM missions, in practice much of the substantive work on analytical preparation of actions and discussion with the government is done bilaterally or in smaller groups, meaning that it is difficult to reach a convergence of views. Greater coordinated dialogue and programming would support a collective agreement on a small, critical set of reforms that all donors would be able to support. Formal appraisal is also done separately or in sub-groups, further weakening the jointness of the budget support program. Overall, the lack of close coordination leads to severe administrative burdens on the country and creates confusion, with the government needing to simultaneously deal with a handful of different donor systems and conventions rather than there being a solid attempt to simplify the process through coordination. 78. Even in the context of an ongoing DPO engagement, emphasis needs to be placed on supervision of previous operations. This ICR has highlighted the importance of supervision activities, on both the part of the Bank and the Borrower, to support continued commitment to achieving results. In this case, even though the commitment to a development policy engagement remained on both the Bank and Borrower side, a lack of visibility of previously-supported reforms and their associated outcome indicators and data contributed to poor overall performance. Especially for clients such as Tuvalu that have only begun engaging with the Bank on operational 25

37 activities, including DPOs, recently, and are thinly spread across various ongoing reform programs, maintaining supervision is important, including some form of reporting, be it ISRs, Aide-Memoires, or incorporation in government-led coordination exercises like the PRM. 79. As far as possible, the lesson from this first DPO program should be factored into future DPO engagements in Tuvalu and elsewhere. This DPO program marked the World Bank s first operational engagement in Tuvalu, its smallest member state. Small size and remoteness pose special challenges, including those that bear on how to most effectively support development policy reform. This ICR suggests that this was a valuable learning exercise to could help to improve the effectiveness of future operations. A new programmatic DPO series is already underway in Tuvalu, and while the first operation was approved prior to preparation of this ICR, as far as possible, its supervision, as well as the design of the second operation, should factor in these lessons. 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partners (a) Borrower/Implementing Agencies: The ICR team greatly appreciates the support of the Government of Tuvalu in preparing this ICR, in supporting data gathering and detailed consultations, as well as providing comments on, and discussions of, the preliminary findings, and then providing a written assessment. We appreciate the continued commitment of the Government to these reforms programs, and the lessons learnt are supported by the findings of the ICR. (b) Cofinanciers: The ICR team wishes to thank all the coordinated budget support development partners who reviewed drafts, providing information and comments and met with the ICR team. The team acknowledges and appreciates the comments provided by DFAT and notes that many of the recommendations therein are supportive of the findings of the ICR. 26

38 Annex 1. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members P Tuvalu Development Policy Operation Name Title Unit Responsibility/ Specialty Lending Lucy Pan Economist OPSCE Task Team Leader Stephen Hartung Financial Management Specialist GGO20 Financial Management Marjorie Mpundu Senior Counsel LEGAM Country Lawyer Patricia De la Fuente Senior Financial Management OPSPF Loan Officer Hoyes Specialist David Knight Senior Economist GMF10 Team Member Samantha Evans Program Assistant EACNF Team Member Supervision Lucy Pan Economist OPSCE Task Team Leader P Tuvalu Second Development Policy Operation Name Title Unit Responsibility/ Specialty Lending Lucy Pan Economist OPSCE Task Team Leader Stephen Hartung Financial Management Specialist GGO20 Financial Management Marjorie Mpundu Senior Counsel LEGAM Country Lawyer Patricia De la Fuente Senior Financial Management OPSPF Loan Officer Hoyes Specialist Jinan Shi Senior Procurement Specialist GGO06 Procurement David Knight Senior Economist GMF10 Team Member Samantha Evans Program Assistant EACNF Team Member Supervision Lucy Pan Economist OPSCE Task Team Leader P Tuvalu Second Development Policy Operation: Supplemental Financing Name Title Unit Responsibility/ Specialty Lending Lucy Pan Economist OPSCE Task Team Leader Stephen Hartung Financial Management Specialist GGO20 Financial Management Marjorie Mpundu Senior Counsel LEGAM Country Lawyer Patricia De la Fuente Senior Financial Management OPSPF Loan Officer Hoyes Specialist Jinan Shi Senior Procurement Specialist GGO06 Procurement David Knight Senior Economist GMF10 Team Member Samantha Evans Program Assistant EACNF Team Member Supervision Lucy Pan Economist OPSCE Task Team Leader 27

39 (b) Staff Time and Cost Stage Staff Time and Cost No. Staff Weeks Cost (US$) including travel and consultant costs Lending First Development Policy Operation 0.7 2,787 Second Development Policy Operation ,454 Total Supervision/ICR First Development Policy Operation Second Development Policy Operation ,410 Total 28

40 Annex 2. Borrower s Comments on Draft ICR The World Bank s development policy operation series marked the start of a new partnership with both the Bank and other budget support partners to improve the efficiency and effectiveness of government, put in place new policies, supported by high quality technical assistance and providing needed budget financing. The DPOs were part of a government-led Policy Reform Matrix (PRM) program and making sure the government is in the driving seat, but with close coordination of other donors, has been a strength of this arrangement. At a high level, the PRM reforms over this period helped to improve fiscal discipline and drive reform progress in some key service delivery areas. Comments on implementation of the program We think the reforms supported in this operation have been very important in improving fiscal oversight and transparency, and the DPO engagement, and the broader PRM mechanism, has been critical in helping us to achieve some major successes. Monthly financial reporting which was supported by DPO1 in this program has become central to the Ministry of Finance and Economic Development s (MFED) in-year cash and budget monitoring and provides a very useful information source to discuss and analyze emerging trends in revenue and execution. The monthly reporting in a standardized format has supported MFED to maintain a medium-term fiscal framework which is updated on a rolling basis. The reform supported in DPO2 was the roll-out of electronic payment vouchers in the governments IFMIS system, ACCPAC, which replaces a manual and less efficient previous system. This change, accompanied with strengthening of procurement oversight, meant that the government gained a comprehensive and real-time picture of its current commitment levels. While the new process took some time to bed down across all parts of government, we are happy to say that it is now working effectively and is enabling us to maintain good fiscal discipline and due process for all expenditure. GoT is using a Modified Accounting System and to date only does one year-end adjustments at the end of the financial year. In 2017 this has been flagged with a push to maintain these reconciliations of stock accounts and reduce and reconcile arrears and accrue accordingly at year end. These are also being factored into reform plans and Treasury are to control these functions. The government led efforts to try to improve the comprehensive monitoring of revenue or accounts receivable and have been working to reconcile Treasury cash records with Fisheries and IRD records, as well as settling cross-debts with SOEs. There reforms have been difficult, due to a limited ability in the line ministries to fully record accrued liabilities, and indeed there have been problems with the revenue administration system (RMS). However, we have recently completed a full reconciliation of historical fisheries revenue. We will need more efforts, supported by TA, to rectify issues with the revenue administration system and ensure full reconciliation is being done. The 2016 Audited Financial Statements of the Govt has flagged unreconciled revenue data from Tax only. This will be flagged in the PFM matrix and monitored and reported henceforth. Fisheries, whilst trying to use the ACCPAC systems to record and manage fishing revenue, continues to have systems hiccups. To date in 2017 Fisheries has established a Fisheries Forum Agency (FFA) supported Fisheries Database and Information Management System. The challenge now exists where data will be integrated and updated into the ACCPAC FMIS from December 2017 and future years. PFM matrix will flag this and monitor for corrective measures to be implemented. Over these years, the government implemented new policies aimed at strengthening the management of Kaupule project management and financial management. Both the outer island project management policy and the new Kaupule financial reporting framework were designed in close consultation with Kaupule authorities and were developed mindful of the small size and limited capacity of their bureaucracy. The government has continued to roll-out these measures and build 29

41 capacity in the Kaupules. Given the small staff and high turnover, this has been challenging but the new frameworks have provided a good basis to supported improved capacity and processes, with Kaupules now providing regular in-year financial reporting and project reporting. The government has been constrained in how much support they can provide the Kaupules, and the reforms have been particularly successful where development partners have supported selected Kaupules (such as under the LOCAL program). The DPOs also supported reforms to both overseas education and health programs, aiming to improve their cost efficiency to free up funds for high-priority spending in-country. Our scholarship schemes and the overseas medical treatment schemes have been essential for development in Tuvalu, given the lack of facilities in-country. The reforms to tighten the cost control of the TMTS program have been implemented, with the government being more selective in identify cost-efficient locations for treatment. But there continues to be expansion in the cost of the program because of an increase in patient numbers, and this is exacerbated by the weak capacity in country to proactively tackle NCD risk factors or adequately identifying and treat conditions at an early stage. We are now planning to focus on building sustainable capacity in parts of the health service in country to carry out procedures that would previously have had to be sourced overseas, which is expected to lead to considerable savings as well as improved health outcomes. In late 2017 a new PS of Health with finance background was appointed who has commenced rigorous work to manage and control TMTS with MFED support and will continue. Already addressed in 2017 are outstanding accounts and communications to date, and we plan to review the Acts and possibly determine appropriate regulations to support these measures. The DPO program also supported the development of a new commercial fishing training course at TMTI, which was open to both men and women. Unfortunately, despite the first batch of students graduating from this course, they were unable to identify jobs in the sector, which mean there was a lack of take-up from students in future years. We think we need to think carefully about how we can make this into a successful venture, and what other supporting measures have to be taken. Some questions to evaluate include: What outcomes are expected here; can women get jobs in men s fields; can courses and trainings be tailored to women s targeted areas and, if yes, can they easily be employed? Resources can be allocated into areas that women could benefit, e.g. pairing or attachments with international organizations remains other alternatives to reconsider. Also were TMTI/Fisheries staff trained to maintain the database? If not rescope and implement to meet targeted outcomes. Lessons learned We continue to believe that coordinated budget support is an effective means of development assistance, and would like to see this mechanism strengthened even further. The PRM has been the government-led coordination mechanism that forms the basis of budget support programs from five different development partners. The PRM process has evolved over time, and by now is a well-tested detailed process which includes regular reporting, discussion and monitoring of TA, implementation and results. The DPO reflects well how we aim to use the PRM for the budget support dialogue. Yet the DPO is only one of a number of different budget support operations and there continue to be differences both in actions supported, processes and timings between the donors. We would like to see closer coordination of partners, including the World Bank, to increase the focus on a critical reform program and reduce the risk that ministries will be overstretched with too many reform programs and bureaucratic requirements for budget support programs. Some of the reforms supported were quite optimistic given the thin capacity in Tuvalu, and despite the best efforts of government, they have been slow to implement. Having a good evidence base and clear understanding of how the whole reform program is expected to play out should be a prerequisite before reforms are committed to. We appreciate the evidence presented in the DPO PDs to show why the policy reforms supported are important and how they will be monitored. But in some cases, there 30

42 was a huge amount of work to do to make sure these reforms were implemented, and there was a lack of support from donors in terms of TA in some areas to do this. It would be good to make sure that the resource needs to deliver a reform are well thought-through and discussed between donors and the government, and we identify how these requirements will be resourced. Similarly, in many areas, Tuvalu looks to partners to provide high-quality analysis on how to delivery on government objectives. In some areas, while reforms have seemed to be good, like the TMTI course, it turns out that there were a number of problems which hadn t been thought through. Through our partnership with development partners, we are keen to ensure that we work through such issues in advance. The DPO program contains a lot of results indicators, many of which were based on data that was not available. Ensuring that data sources are developed and maintained is a reform in its own right, and in many areas we have been stretched to try to develop new monitoring systems, while there has been relatively little TA support in these areas. Every reform area needs to be focused very carefully to make sure that it is fully embedded in the government structure. With so few staff in functions, there is a constant changeover in staffing so it is not adequate to rely on experience of one or two people. So it is good that the DPO program focuses on high-level reforms to policies and laws, but there should also be a strong focus on making sure that these reforms are fully reflected in the ongoing work of the government. In some cases, reforms can be driven by a TA to produce a report for example, and when the TA leaves, there is not enough capacity for the report to be replicated. This is a summary of the key objectives the we have for the PRM and budget support processes: 1. Government must take the driving seat and take ownership of reforms working closely with DPs. 2. The PRM is a 'living' document and therefore there should be room to modify reforms given the capacity to implement and other issues such as staff turn-over. 3. Reforms are encouraged to be designed in a progressive manner through phases. 4. Use of reports from Tuvalu Trust Fund Advisory Committee and other reports that are available to verify actions and results. This can avoid duplication of work. 5. Given our capacity to implement reforms there should be a limit to the number of reforms donors tie their budget support to. 31

43 Annex 3. Comments of Cofinanciers and Other Partners/Stakeholders Comments received from Lily-Anne Homasi, Senior Program Manager Tuvalu, Development Cooperation, Australia Department of Foreign Affairs and Trade The assessment on individual reforms for DPO 1 and 2 is fair in my view, noting the lack of information in some areas. It would be good to outline the methodology used to conduct the ICR, including any assumptions guiding/influencing the assessment. For instance, it might be useful to state in the beginning that to improve public financial management and delivery of social services it takes many years for a smaller island state government to deliver/achieve reforms. In an ideal world, some of the assumptions to progress and sustain reforms could include whether the client has the political appetite and ownership of such reforms, high capacity (quantity and quality) to implement reforms and whether it is complemented by timely technical support from donors and development partners. There are also other factors to consider (e.g. high staff turnover and low institutional knowledge on reforms, poor record management, limited targeted TA provided to support proposed reforms) beyond the reforms agreed to ensure they have the desired impact. We need to acknowledge as well that achievement of reform actions and triggers will not necessary lead to improvements in development outcomes as we need a few rounds of PRMs to effect positive outcomes. It would be useful to include feedback on any unintended consequences (if any) on changes in reforms and progressing revised reforms. As appropriate, it would be useful to reference sources of information e.g PRM review, 2011 and 2015 PEFAs, 2011 and 2015 ANS, Tuvalu MTEF and what about TTFAC reports, MFED? I agree (in principle) with the lessons learned section, it mirrors the 2016 PRM review lessons learned and recommendations. In terms of donor coordination, there is room for improvement to strengthen our working relationship. Some of these areas could include: WB to commit and provide targeted TA to support reforms that they are requesting government to take forward, particularly in areas where there is limited capacity. We can t be complacent in leaving GoT to procure and manage TAs as this is an area that they are struggling given the limited number of staff multi-tasking. On an annual basis, we have about 5-7 advisers in Tuvalu supporting these reforms and their own individual assignments objectives as agreed with GoT. Sharing of analytical work to inform reforms is critical. We have shared reviews we ve done with the Bank in a timely manner but I think the Bank could do better to share analytical work it carries out relevant to the PRM with development partners working together under that framework in a timely manner to inform discussions. Embedding reforms in government processes/established mechanisms, e.g. on monitoring and reporting on budget support. Is there a possibility for the WB to do its monitoring/assessment of prior actions with the TTFAC (an established mechanism consisting of economists/development and PFM advisors) to reduce the burden on government consultations? For DFAT and I believe some other PRM partners, we rely on TTFAC reporting to verify the achievement of PRM actions. However, there are rare occasions that we may have to verify an action outside of this standard reporting, if the government officially requests to receive the budget support earlier. Otherwise, 32

44 disbursements really depend on the government achieving actions and that TTFAC verifies it being achieved. Brief comments were also received on the ICR from representatives of the Asian Development Bank, New Zealand Ministry of Foreign Affairs and Trade, and the European Union. 33

45 Annex 4. List of Supporting Documents 1. Program Document, First Development Policy Operation (P145488), October 21, Letter of Development Policy, September 10, Program Document, Second Development Policy Operation (P150194), February 27, Letter of Development Policy, January 9,

46 MAP 35