## Section I. Detailed Assessment and Recommendations

## Source details

**Canonical URL:** [Section I. Detailed Assessment and Recommendations](https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025027-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/tar/2025/english/tarea2025027-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/tar/2025/english/tarea2025027-print-pdf.pdf.json)

---

### Overall Assessment
- Assessment against the IMF’s Data Quality Assessment Framework (DQAF) for Public Sector Debt Statistics (PSDS) 2024.
- Mission funded by the Government of Japan through the Japan Administered Account for Selected IMF Activities (JSA).
- Mission participants: Mr. Rasiel Vellos (World Bank) and Ms. Joanne Allin (Commonwealth Secretariat).
- Based on information provided prior to and during a mission from September 23–27, 2024 and publicly available information.
- Agencies compiling assessed datasets: Ministry of Finance (MOF), Bank of Tanzania (BOT), National Bureau of Statistics (NBS) Tanzania.
- IMF STA capacity development focused on GFSM 2014 and PSDSG 2013; progress made but significant gaps remain for East African Monetary Union (EAMU) data needs.
- Key chronology and system developments:
  - Implemented e-GDDS when launched in May 2015.
  - Implemented National Summary Data Page (NSDP) in October 2016.
  - Latest available public debt data on NSDP is gross central government debt for Q4 2015.
  - DMD modernized by implementing CS Meridian in financial year (FY) 2022/23; first country to interface CS Meridian with IFMIS.
- Article IV 2023: public sector debt reporting covers central government debt, central government-guaranteed debt (excluding non-guaranteed debt of SOEs and local governments), and central bank debt.
- Systemic gaps identified:
  - Lack of granularity on local government debt and public corporations’ debt.
  - Persistent high levels of expenditure arrears and weak capacity to monitor them.
- Expenditure arrears defined: unpaid claims outstanding for over 30 days for delivery of goods and services and for over 90 days for construction work.
- Mission objective: identify key weaknesses in PSDS compilation and dissemination and priority actions to enhance debt data transparency.
- Summary conclusion: substantial improvements achieved (CS Meridian interface, data validation, security, reconciliation), but remaining gaps in coverage, valuation, metadata, dissemination, and revision practices.

### A. Prerequisites of Quality
- Legal environment:
  - Primary responsibility clearly established.
  - CAP 134, Government Loans, Grants and Guarantees, Act 1974 (as amended) provides guidance; regulations (Government loans, grants, and guarantees regulations 2003 (as amended)) specify debt reporting arrangements.
  - DMD in MOF is the custodian of public debt data.
- Institutional environment:
  - Data sharing and coordination arrangements among data-producing agencies are adequate.
  - National Debt Management Committee (NDMC) advises Minister of Finance; receives technical advice from Technical Debt Management Committee (TDMC).
  - TDMC chaired by Commissioner for DMD; DMD serves as Secretariat and is sufficiently resourced for Resource Mobilization, Risk Analysis, and Data Management.
  - Recommendation: publicize recent achievements in data management and seek feedback from data users outside government (see Appendix II).

### B. Assurances of Integrity
- Professionalism, transparency, and ethical standards:
  - Debt data producing agencies demonstrate professionalism, transparency, and promote high ethical standards.
  - Choice of source data, methodologies, and statistical techniques based on measurement objectives and data requirements.
  - Clear ethical standards set out in Government Code of Ethics and Staff Regulations, and the Civil Service Law.
  - Recommendation: enhance transparency by widely publishing the latest version of CAP 134 Act together with updated legislation.

### C. Methodological Soundness
- General compliance:
  - Compilation broadly follows international standards.
  - Reported PSDS cover all loans and debt securities of the budgetary central government (BCG) with separate information on government guarantees and borrowing of other public sector units.
  - Domestic vs external debt identified based on creditor residency; individual loans aligned with residency concept.
  - MOF (Accountant General’s Division) produces audited consolidated financial statements covering nearly 600 public entities.
- Valuation and recording:
  - Accounting basis broadly follows statistical standards but valuation used is face value rather than preferred nominal value.
  - Time of recording aligns with statistical standards; some information available by original and remaining maturity.
- Identified methodological improvements:
  - Include liabilities for accounts payable, currency and deposits, and insurance, pension, and standardized guarantee schemes.
  - Use CS Meridian capabilities to change reported valuation from face value to nominal value.
  - Provide metadata when the conversion rate used differs from the recommended mid-point rate.
  - Align taxonomy within PSDS reports for debt instruments and creditors to international standards.
- Specific table recommendations (from Table 1):
  - Use CS Meridian to produce and disseminate reports presenting PSDS at nominal valuation.
  - Expand sector coverage of core PSDS reports to include local governments, social security funds, extrabudgetary units, and public corporations.
  - Include other accounts payable, including expenditure arrears, and deposit liabilities.
  - Use CS Meridian reporting to present PSDS according to international debt instrument classifications by original maturity and remaining maturity.
  - Compile foreign currency denominated debt using the mid-point exchange rate between the selling rate and the buying rate.

### D. Accuracy and Reliability
- System integration and validation:
  - Tanzania interfaced CS Meridian with IFMIS, becoming the first country to do so.
  - New system enables migration and validation of debt data against contractual documents and reconciliation with creditor holdings.
  - Security measures, access control, and audit features ensure secure data and managed updates.
- Remaining accuracy tasks:
  - Republish historical series where data have been corrected and revised.
  - Socialize a draft revision policy.
  - Leverage web-based features of CS Meridian to expand government perimeter to local authorities and state-owned enterprises.

### E. Serviceability
- Periodicity, timeliness, consistency, and revisions:
  - PSDS are consistent across balance of payments, monetary accounts, and government finance statistics.
  - PSDS meet or exceed periodicity and timeliness of e-GDDS requirements (noting NSDP updates stopped in 2016).
- Identified serviceability improvements:
  - Commence submission of debt data to the World Bank Quarterly Public Sector Debt (QPSD) statistics database.
  - Update the NSDP on the IMF’s Dissemination Standards Bulletin Board.
  - Initiate validation of historical series of migrated data jointly with the World Bank.
  - Align public debt presentation to international standards.
  - User Survey echoed need for improvements (see Appendix II).
  - With CS Meridian and DMD commitment, improvements are deliverable.

### F. Accessibility
- Current dissemination:
  - BOT and DMD publish many debt tables in monthly, quarterly, and annual publications.
- User needs and recommendations:
  - Modern users prefer raw data in easily downloadable formats.
  - Low-hanging improvements:
    - Link users to raw data through the NSDP.
    - Publish quarterly debt statistics regularly through the QPSD.
  - Enhance serviceability by publishing and maintaining key documentation: concordances, sources and methods, and a general dissemination policy across websites.

### Table 1 — Selected Highlights of PSDS DQAF Assessment
- 0. Prerequisites of quality
  - 0.1 Legal and institutional environment — O: Primary responsibility clearly established.
  - 0.2 Resources — O: Staff, facilities, computing resources, and financing commensurate with debt management programs.
  - 0.3 Relevance — LO: Lack of feedback from data users outside Government.
    - Recommendation: Introduce structured and periodic outreach to data users outside Government.
  - 0.4 Other quality management — O: Recognition that quality builds trust.

- 1. Assurances of integrity
  - 1.1 Professionalism — O: Terms/conditions in accordance with professional independence.
  - 1.2 Transparency — LO: DMD not a specialized statistical unit; public debt data freely available but DMD should publish statistical products as primary source.
  - 1.3 Ethical standards — O: Management are role models.

- 2. Methodological soundness
  - 2.1 Concepts and definitions — LO: In line with GFSM 2014, only disbursed outstanding debt recognized as debt; contingent liabilities recorded separately; recommendation to use CS Meridian to present PSDS at nominal valuation.
  - 2.2 Scope — LNO: Sector coverage limited to budgetary central government; recommendation to expand sector and instrument coverage (include local governments, social security funds, extrabudgetary units, public corporations, other accounts payable, expenditure arrears, deposit liabilities).
  - 2.3 Classification / sectorization — LNO: Main debt bulletins do not present data by maturity or use GFSM 2014 nomenclature; recommendation to use CS Meridian reporting for instrument and maturity presentation.
  - 2.4 Time of debt recording, valuation, and consolidation — LO: Data disseminated at face value though CS Meridian can present nominal valuation; interest arrears included; PPPs and financial leases not recorded, nor is the debt consolidated.
    - Recommendation: Compile foreign currency denominated debt using the mid-point exchange rate between the selling rate and the buying rate.

### 3. Accuracy and reliability — Source data and validation
- 3.1 Source data
  - LO: Source data are audited and of high quality for the debt securities and loans of BCG.
  - Good quality data are available for compiling PSDS of other sectors and instruments, but they are not yet used.
- 3.2 Assessment of source data
  - LO: Source data in CS Meridian are subject to validation, such as internal consistency and cross checks with creditor data.
  - Contingent liabilities data not subject to same robust checks.
  - Recommendation: Review and, where required, improve the data collection and reporting processes for contingent liabilities and approved borrowing of other public sector entities, automating where possible.
- 3.3 Statistical techniques
  - LNO: Data are on a cash basis and no statistical techniques are employed to include additional liabilities (such as finance leases) or refine existing ones (such as on assumptions of residency).
  - Note: While coverage remains at budgetary central government this is not a priority area for improvements.
- 3.4 Assessment and Validation of Statistical Outputs
  - LO: Introduction of CS Meridian has introduced debt revisions as a result of strengthened validation checks.
  - Required actions:
    - Publish a historical timeseries comparing the new debt data, as validated and included in CS Meridian, with that previously published to inform users of any revisions.
    - Publish a reconciliation of the change in debt stocks with the transactions and other flows driving the changes.
  - Additional guidance: Compare new data with previously published debt data and check and explain revisions.
- 3.5 Revision studies
  - NO: No revision analysis undertaken, and those revisions that are observed are not explained.
  - Recommendation: When disseminating PSDS for the latest period revise (as required) debt reported for previous periods and provide summary metadata for users to understand the reasons for the revisions.

### 4. Serviceability — Periodicity, consistency, revisions
- 4.1 Periodicity and timeliness
  - O: The PSDS are reported with a high degree of periodicity and timeliness.
- 4.2 Consistency
  - O: The data show good levels of consistency both between PSDS publications and with publications of other macroeconomic statistics.
- 4.3 Revision policy and practice
  - NO: MOF and BOT have not developed and published a PSDS revision policy hence revision studies cannot be undertaken.
  - Recommendation: Develop and publish a PSDS revision policy to guide users and compilers on the process for reporting and analyzing revisions.
  - Note: The GFS and PSDS technical committee has a workplan to develop and publish a revision policy by the end of October 2025.

### 5. Accessibility — Data, metadata, assistance
- 5.1 Data accessibility
  - LO: Published debt data by both MOF and BOT on the national websites are accessible and include commentaries and charts/tables that can facilitate user analysis and understanding.
  - Recommendation: Resume publication of BCG gross debt in the NSDP and subscribe to the QPSD.
- 5.2 Metadata accessibility
  - LNO: Metadata for published reports is limited and not available to users in a structured format.
  - Recommendation: Metadata on PSDS should be prepared in a structured format, published, and sustained.
- 5.3 Assistance to users
  - LNO: Published PSDS reports provide some contact information (emails, postal addresses, phone numbers).
  - Recommendation: Proactively engage with PSDS users, survey users on needs, and raise awareness on where users can get assistance.

Legend: O = Practice Observed; LO = Practice Largely Observed; LNO = Practice Largely Not Observed; NO = Practice Not Observed

### Data migration, validation, and governance (CS Meridian rollout)
- External and domestic debt data migrated from the previous debt database to CS Meridian during 2022/23.
- Debt data from 2000 onwards brought into the new system.
- Extensive validation against original contractual documents and reconciliation with creditors undertaken.
- Structural changes in MOF in FY2021/22 to allow DMD to be subject to annual financial audits.
- Annual financial audits used as a governance process to ensure accuracy and reliability of debt data managed by DMD.

### CS Meridian functionality, access, and linkages
- CS Meridian is a web-based system allowing staff in MOF, BOT and President’s Office – Finance and Planning (POFP) of the Government of Zanzibar to access and report data.
- Security measures, access control, and audit features ensure debt data is held securely and updates are appropriately managed.
- Tanzania connected CS Meridian to other government data systems:
  - Government information system, MUSE (“Mfumo wa Ulipaji Serikalini”): debt servicing information shared and payment confirmation received.
  - Government Securities System (Bank of Tanzania): auction data related to Treasury bonds and bills shared.
  - POFP (Government of Zanzibar): Domestic Debt of the Government of Zanzibar — none connected.

### Synchronization with World Bank Debtor Reporting System (DRS)
- Transition from CS-DRMS to CS Meridian led to advancements in management and validation of external debt data.
- Historical central government external debt data (loan-by-loan) cleaned, validated, and reconciled with creditors within CS Meridian.
- DMD and the World Bank agreed to undertake a comprehensive reconciliation between CS Meridian and the DRS.
  - This includes updating historical data previously reported to the DRS to reflect cleaned and validated figures.
  - Assistance will be provided for tracking and updating general government and public corporation (including SOEs) debt in the DRS under the World Bank Policy on External Debt Reporting and Financial Statements.
- An output will be a timeseries of revisions between validated CS Meridian data and previously published data.
  - Recommendation: publish the new timeseries for as many years and quarters as possible with commentary explaining causes of revisions.
  - Note: current MOF policy is not to restate the debt for previous quarters in their quarterly debt report; any debt revisions should be explained to users.

### Contingent liabilities and approved borrowing
- MOF publishes detailed contingent liabilities of BCG in annexes to their quarterly debt report, including individual guarantees, other contingent liabilities, and outstanding stock of approved borrowing by public entities.
- Comparison June 2024 vs December 2023 highlighted large movements:
  - Footnote: Over the six months there was a drop in contingent liabilities (excluding the guarantee provided to the National Insurance Corporation) of 1.1 trillion TZS from 3.4 trillion TZS at the end of December 2023, and a drop in the outstanding stock of approved borrowing of 1.0 trillion TZS from 2.3 trillion TZS at the end of December 2023.
- Some contingent liabilities data collected through spreadsheets and not CS Meridian, introducing potential for errors.
- Recommendation: review data and compilation processes for contingent liabilities to ensure robustness comparable to loans and debt securities; automate where possible.

### Local government, public corporations, and other public sector units
- Comprehensive data source exists for local government debt inclusion:
  - 184 local government entities (including city councils, municipal councils, town councils and district councils) all use MUSE and a common chart of accounts.
  - End-of-year local council data review with the President’s Office for Regional Administration and Local Government prior to finalizing financial statements, which are audited.
  - Local government data accessible through MUSE; viable to source annual, quarterly, and monthly data.
- Financial statements available for:
  - 4 social security funds,
  - 214 (approx.) extrabudgetary units,
  - 86 (approx.) public corporations.
- These public sector units are not currently included in PSDS, but audited timely financial statements provide a strong foundation for future annual PSDS for the entire public sector.
- MOF has a Consolidation Unit experienced in consolidation, which should facilitate expanding PSDS beyond BCG while eliminating double counting.

### Strengths identified
- CS Meridian enforces checks and validations on granular debt data.
- Re-structuring of Vote 001 (Public Debt) to allow audit provides governance framework for DMD data accuracy.
- Government Loans, Guarantees and Grants Act 1974 (as amended) ensures MOF has a comprehensive view of borrowing activities.
- Timely audited annual financial statements for public sector entities provide rich data for instrument and sector expansion.
- PSDS reported with high periodicity and timeliness; BOT publishes monthly, quarterly, and annual data:
  - Monthly and quarterly data published within a month of the end of the reporting period.
  - Annual data published within six months of the end of the year.
- DMD published a quarterly debt report significantly expanded and upgraded in 2024; revamped quarterly bulletin (≈40 pages) from July 2024 (three such reports published at time of mission: two in July 2024 and one in September 2024).
- Published PSDS by BOT and DMD are internally consistent and largely consistent with other macroeconomic datasets (e.g., BOP, monetary statistics).

### Recommendations for improvements (consolidated)
- Review and improve data collection and reporting processes for contingent liabilities and approved borrowing of other public sector entities; automate where possible.
- Publish a historical timeseries comparing CS Meridian validated data with previously published data.
- When disseminating PSDS for the latest period, revise (as required) previous period debt and provide summary metadata explaining revisions.
- Publish a reconciliation of changes in debt stocks with the transactions and other flows driving the changes (see Annex III Table A2 for an example).
- Develop and publish a PSDS revision policy (workplan target: publish by the end of October 2025).
- Publish explicit reconciliation tables linking changes in debt stocks with transactions and other flows.
- Resume publication of BCG gross debt in the NSDP and subscribe to the QPSD.
- Prepare, publish, and sustain structured metadata on PSDS (use NSDP or QPSD metadata templates).
- Proactively engage with PSDS users (survey users, raise awareness, provide assistance information).

### Accessibility and dissemination issues
- MOF and BOT national websites publish debt reports in .pdf format:
  - BOT reports are editable allowing table/chart extraction.
  - DMD reports are published in non-editable format, making extraction harder.
- No PSDS currently made available on national websites in machine-readable formats such as SDMX, Excel or .csv (except Tanzania PSDS on the EAC website can be extracted in these formats).
- NSDP established October 2016, but latest debt data relate to December 2015 and metadata not updated since 2014.
- Tanzania has never disseminated data via the joint IMF-World Bank QPSD.
- Commentary in DMD and BOT publications adds value, but additional metadata should be included:
  - Valuation of debt stocks,
  - Sector coverage,
  - Exchange rate used for converting foreign currency debt,
  - Approach to identifying residency.
- Advanced release calendar managed by National Bureau of Statistics was out-of-date, having ended in June 2024.
- Contact information provided in publications, but no evidence of assistance provided to data users or materials to raise awareness of PSDS.
- Dissemination recommendations reiterated:
  - Resume BCG gross debt in NSDP with consistent periodicity and timeliness.
  - Subscribe to QPSD for wider dissemination and cross-country comparisons.
  - Publish structured metadata and proactively engage with users.

---

### APPENDIX II. USERS’ SURVEY RESULTS — Key findings
- Overall user feedback: generally positive; most users found PSDS "very useful" or "useful".
- Primary uses: Investment decision-making; Compilation of other macroeconomic statistics; Academic research; Policymaking.
- Where users obtain PSDS (counts):
  - Official national debt bulletins and press releases: 12
  - IMF or World Bank databases and/or publications: 12
  - Other international organization databases and/or publications: 0
  - Media reports: 2
  - Data published on the National Summary Data Page (NSDP): 4
  - Private sector summaries and analyses: 1
  - Other sources: 2
- Respondent consent to quote responses: Yes = 13, No = 4
- Responders by institution:
  - International organizations: 4
  - Financial institutes: 2
  - Foreign governments: 2
  - Research institutes: 1
  - Audit agencies: 4
  - Public entities: 4
- Users’ stated main interest in PSDS data (counts):
  - Policymaking: 3
  - Compilation of other macroeconomic statistics: 4
  - Investment decisions: 6
  - Other analytical work: 1
  - Academia or research: 3
  - No response: 0

- Overall assessment (survey responses):
  - Analytic usefulness:
    - Very useful: 10
    - Useful: 6
    - Somewhat useful: 1
    - Not useful: 0
    - Undecided: 0
  - Agreement with "The disseminated PSDS are unbiased and accurate":
    - Strongly agree: 2
    - Agree: 3
    - Neither agree nor disagree: 10
    - Disagree: 1
    - Strongly disagree: 1
  - Satisfaction with timeliness and level of detail:
    - Very satisfied: 1
    - Satisfied: 14
    - Neither satisfied nor dissatisfied: 1
    - Dissatisfied: 1
    - Very dissatisfied: 0

- Coverage and concepts:
  - Satisfaction with sector coverage:
    - Very satisfied: 2
    - Satisfied: 12
    - Neither satisfied nor dissatisfied: 0
    - Dissatisfied: 3
    - Very dissatisfied: 0
  - Satisfaction with coverage of debt instruments:
    - Very satisfied: 1
    - Satisfied: 12
    - Neither satisfied nor dissatisfied: 1
    - Dissatisfied: 3
    - Very dissatisfied: 0
  - Satisfaction with coverage of public nonfinancial corporations (SOEs):
    - Very satisfied: 1
    - Satisfied: 8
    - Neither satisfied nor dissatisfied: 3
    - Dissatisfied: 4
    - Very dissatisfied: 1
  - Instruments inadequately covered (responses):
    - None: 3
    - Accounts payables (arrears): 6
    - Loans: 7
    - Debt securities: 3
    - Pension liabilities: 2
    - Other: 1
  - Sub-sectors where more coverage desired (responses):
    - None: 1
    - Local governments / State governments: 15
    - Public corporations: 12
    - Budgetary central government: 6
    - Extrabudgetary units: 6
    - Other: 1
  - Valuation appropriateness of outstanding stock positions:
    - Yes: 15
    - No: 1

- Periodicity and timeliness:
  - Satisfaction with periodicity:
    - Very satisfied: 2
    - Satisfied: 11
    - Neither satisfied nor dissatisfied: 1
    - Dissatisfied: 1
    - Very dissatisfied: 2
  - Are PSDS disseminated timely?
    - Yes: 11
    - No: 3
    - Not sure: 3

- Dissemination practices:
  - Is there a published release schedule/calendar and are PSDS released punctually?
    - Yes, and the data is released on schedule: 5
    - Yes, and the data is sometimes released on schedule: 5
    - Yes, and the data is never released on schedule: 2
    - No: 1
    - Not sure: 4
  - Is analysis published to explain revisions, breaks, adjustments, discontinuities, or unusual changes?
    - Yes: 8
    - No: 3
    - Not sure: 5
    - No response: 1
  - At dissemination, are users informed whether data are preliminary, revised or updated?
    - Always informed: 7
    - Sometimes informed: 5
    - Never informed: 2
    - Not aware: 3
  - Do users receive advance notice of major changes?
    - Always: 2
    - Sometimes: 6
    - Never: 4
    - Not sure: 5

- Accessibility and usefulness:
  - Agreement that PSDS are presented to facilitate interpretation and comparisons:
    - Strongly agree: 2
    - Agree: 14
    - Neither agree nor disagree: 0
    - Disagree: 1
    - Strongly disagree: 0
  - Availability of comprehensive published source and methods document or metadata:
    - Yes, available and accessible: 6
    - Partial information available: 5
    - No information available: 2
    - Not sure: 4
  - Consistency and reconcilability over time with other macroeconomic datasets:
    - Fully consistent: 6
    - Partly consistent: 7
    - Not consistent: 1
    - Not sure: 3

*Source: IMF Technical Report – United Republic of Tanzania Data Quality Assessment for PSDS (Section I, Appendix II).*

### Section I. Detailed Assessment and Recommendations .....................................................................

### Section I. Detailed Assessment and Recommendations

### Overall Assessment
- Assessment against the IMF’s Data Quality Assessment Framework (DQAF) for Public Sector Debt Statistics (PSDS) 2024.
- Mission undertaken under a project funded by the Government of Japan through the Japan Administered Account for Selected IMF Activities (JSA).
- Mission participants included Mr. Rasiel Vellos (World Bank) and Ms. Joanne Allin (Commonwealth Secretariat).
- Based on information provided prior to and during a mission from September 23–27, 2024 and publicly available information.
- Agencies compiling assessed datasets: Ministry of Finance (MOF), Bank of Tanzania (BOT), National Bureau of Statistics (NBS) Tanzania.
- IMF STA capacity development focused on GFSM 2014 and PSDSG 2013; progress made but significant gaps remain for East African Monetary Union (EAMU) data needs.
- Key chronology and system developments:
  - Implemented e-GDDS when launched in May 2015.
  - Implemented National Summary Data Page (NSDP) in October 2016.
  - Latest available public debt data on NSDP is gross central government debt for Q4 2015.
  - DMD modernized by implementing CS Meridian in financial year (FY) 2022/23; first country to interface CS Meridian with IFMIS.
- Article IV 2023: public sector debt reporting covers central government debt, central government-guaranteed debt (excluding non-guaranteed debt of SOEs and local governments), and central bank debt.
- Identified systemic gaps: lack of granularity on local government debt and public corporations’ debt; persistent high levels of expenditure arrears and weak capacity to monitor them.
- Expenditure arrears defined: unpaid claims outstanding for over 30 days for delivery of goods and services and for over 90 days for construction work.
- Mission objective: identify key weaknesses in PSDS compilation and dissemination and priority actions to enhance debt data transparency.
- Summary conclusion: substantial improvements achieved (CS Meridian interface, data validation, security, reconciliation), but remaining gaps in coverage, valuation, metadata, dissemination, and revision practices.

### A. Prerequisites of Quality
- Legal environment:
  - Primary responsibility clearly established.
  - CAP 134, Government Loans, Grants and Guarantees, Act 1974 (as amended) provides guidance; regulations (Government loans, grants, and guarantees regulations 2003 (as amended)) specify debt reporting arrangements.
  - DMD in MOF is the custodian of public debt data.
- Institutional environment:
  - Data sharing and coordination arrangements among data-producing agencies are adequate.
  - National Debt Management Committee (NDMC) advises Minister of Finance; receives technical advice from Technical Debt Management Committee (TDMC).
  - TDMC chaired by Commissioner for DMD; DMD serves as Secretariat and is sufficiently resourced for Resource Mobilization, Risk Analysis, and Data Management.
  - Recommendation: publicize recent achievements in data management and seek feedback from data users outside government (see Appendix II).

### B. Assurances of Integrity
- Professionalism, transparency, and ethical standards:
  - Debt data producing agencies demonstrate professionalism, transparency, and promote high ethical standards.
  - Choice of source data, methodologies, and statistical techniques based on measurement objectives and data requirements.
  - Clear ethical standards set out in Government Code of Ethics and Staff Regulations, and the Civil Service Law.
  - Recommendation: enhance transparency by widely publishing the latest version of CAP 134 Act together with updated legislation.

### C. Methodological Soundness
- General compliance:
  - Compilation broadly follows international standards.
  - Reported PSDS cover all loans and debt securities of the budgetary central government (BCG) with separate information on government guarantees and borrowing of other public sector units.
  - Domestic vs external debt identified based on creditor residency; individual loans aligned with residency concept.
  - MOF (Accountant General’s Division) produces audited consolidated financial statements covering nearly 600 public entities.
- Valuation and recording:
  - Accounting basis broadly follows statistical standards but valuation used is face value rather than preferred nominal value.
  - Time of recording aligns with statistical standards; some information available by original and remaining maturity.
- Identified methodological improvements:
  - Include liabilities for accounts payable, currency and deposits, and insurance, pension, and standardized guarantee schemes.
  - Use CS Meridian capabilities to change reported valuation from face value to nominal value.
  - Provide metadata when the conversion rate used differs from the recommended mid-point rate.
  - Align taxonomy within PSDS reports for debt instruments and creditors to international standards.
- Specific table recommendations (from Table 1):
  - Use CS Meridian to produce and disseminate reports presenting PSDS at nominal valuation.
  - Expand sector coverage of core PSDS reports to include local governments, social security funds, extrabudgetary units, and public corporations.
  - Include other accounts payable, including expenditure arrears, and deposit liabilities.
  - Use CS Meridian reporting to present PSDS according to international debt instrument classifications by original maturity and remaining maturity.
  - Compile foreign currency denominated debt using the mid-point exchange rate between the selling rate and the buying rate.

### D. Accuracy and Reliability
- System integration and validation:
  - Tanzania interfaced CS Meridian with IFMIS, becoming the first country to do so.
  - New system enables migration and validation of debt data against contractual documents and reconciliation with creditor holdings.
  - Security measures, access control, and audit features ensure secure data and managed updates.
- Remaining accuracy tasks:
  - Republish historical series where data have been corrected and revised.
  - Socialize a draft revision policy.
  - Leverage web-based features of CS Meridian to expand government perimeter to local authorities and state-owned enterprises.

### E. Serviceability
- Periodicity, timeliness, consistency, and revisions:
  - PSDS are consistent across balance of payments, monetary accounts, and government finance statistics.
  - PSDS meet or exceed periodicity and timeliness of e-GDDS requirements (noting NSDP updates stopped in 2016).
- Identified serviceability improvements:
  - Commence submission of debt data to the World Bank Quarterly Public Sector Debt (QPSD) statistics database.
  - Update the NSDP on the IMF’s Dissemination Standards Bulletin Board.
  - Initiate validation of historical series of migrated data jointly with the World Bank.
  - Align public debt presentation to international standards.
  - User Survey echoed need for improvements (see Appendix II).
  - With CS Meridian and DMD commitment, improvements are deliverable.

### F. Accessibility
- Current dissemination:
  - BOT and DMD publish many debt tables in monthly, quarterly, and annual publications.
- User needs and recommendations:
  - Modern users prefer raw data in easily downloadable formats.
  - Low-hanging improvements:
    - Link users to raw data through the NSDP.
    - Publish quarterly debt statistics regularly through the QPSD.
  - Enhance serviceability by publishing and maintaining key documentation: concordances, sources and methods, and a general dissemination policy across websites.

### Table 1 — Summary Results of PSDS DQAF Assessment (selected highlights)
- 0. Prerequisites of quality
  - 0.1 Legal and institutional environment — O: Primary responsibility clearly established.
  - 0.2 Resources — O: Staff, facilities, computing resources, and financing commensurate with debt management programs.
  - 0.3 Relevance — LO: Lack of feedback from data users outside Government.
    - Recommendation: Introduce structured and periodic outreach to data users outside Government.
  - 0.4 Other quality management — O: Recognition that quality builds trust.

- 1. Assurances of integrity
  - 1.1 Professionalism — O: Terms/conditions in accordance with professional independence.
  - 1.2 Transparency — LO: DMD not a specialized statistical unit; public debt data freely available but DMD should publish statistical products as primary source.
  - 1.3 Ethical standards — O: Management are role models.

- 2. Methodological soundness
  - 2.1 Concepts and definitions — LO: In line with GFSM 2014, only disbursed outstanding debt recognized as debt; contingent liabilities recorded separately; recommendation to use CS Meridian to present PSDS at nominal valuation.
  - 2.2 Scope — LNO: Sector coverage limited to budgetary central government; recommendation to expand sector and instrument coverage (include local governments, social security funds, extrabudgetary units, public corporations, other accounts payable, expenditure arrears, deposit liabilities).
  - 2.3 Classification / sectorization — LNO: Main debt bulletins do not present data by maturity or use GFSM 2014 nomenclature; recommendation to use CS Meridian reporting for instrument and maturity presentation.
  - 2.4 Time of debt recording, valuation, and consolidation — LO: Data disseminated at face value though CS Meridian can present nominal valuation; interest arrears included; PPPs and financial leases not recorded, nor is the debt consolidated.
    - Recommendation: Compile foreign currency denominated debt using the mid-point exchange rate between the selling rate and the buying rate.

*Source: IMF Technical Report – United Republic of Tanzania Data Quality Assessment for PSDS (Section I).*

### 3. Accuracy and reliability

### 3. Accuracy and reliability

### 3.1 Source data
- LO: Source data are audited and of high quality for the debt securities and loans of BCG.
- Good quality data are available for compiling PSDS of other sectors and instruments, but they are not yet used.
- [see 2.2]

### 3.2 Assessment of source data
- LO: Source data in CS Meridian are subject to validation, such as internal consistency and cross checks with creditor data.
- Data on contingent liabilities (and similar) is not subject to the same robust data quality checks.
- Recommendation:
  - Review and, where required, improve the data collection and reporting processes for contingent liabilities and approved borrowing of other public sector entities, automating where possible.

### 3.3 Statistical techniques
- LNO: Data are on a cash basis and no statistical techniques are employed to include additional liabilities (such as finance leases) or refine existing ones (such as on assumptions of residency).
- Note: [While the coverage remains at budgetary central government this is not a priority area for improvements.]

### 3.4 Assessment and Validation of Statistical Outputs
- LO: Introduction of CS Meridian has introduced debt revisions as a result of the strengthened validation checks.
- Actions required to finalize the process:
  - Publish a historical timeseries comparing the new debt data, as validated and included in CS Meridian, with that previously published to inform users of any revisions.
  - Publish a reconciliation of the change in debt stocks with the transactions and other flows driving the changes.
- Additional guidance:
  - The new data should be compared with the previously published debt data and the revisions checked and explained.

### 3.5 Revision studies
- NO: No revision analysis undertaken, and those revisions that are observed are not explained.
- Recommendations:
  - When disseminating PSDS for the latest period revise (as required) debt reported for previous periods and provide summary metadata for users to understand the reasons for the revisions.

### 4. Serviceability

#### 4.1 Periodicity and timeliness
- O: The PSDS are reported with a high degree of periodicity and timeliness.

#### 4.2 Consistency
- O: The data show good levels of consistency both between PSDS publications and with publications of other macroeconomic statistics.

#### 4.3 Revision policy and practice
- NO: MOF and BOT have not developed and published a PSDS revision policy hence revision studies cannot be undertaken.
- Recommendation:
  - Develop and publish a PSDS revision policy to guide users and compilers on the process for reporting and analyzing revisions.

### 5. Accessibility

#### 5.1 Data accessibility
- LO: Published debt data by both MOF and BOT on the national websites are accessible and include commentaries and charts/tables that can facilitate user analysis and understanding.
- Recommendation:
  - Resume publication of BCG gross debt in the NSDP and subscribe to the QPSD.

#### 5.2 Metadata accessibility
- LNO: Metadata for published reports is limited and not available to users in a structured format.
- Recommendation:
  - Metadata on PSDS should be prepared in a structured format, published, and sustained.

#### 5.3 Assistance to users
- LNO: The published PSDS reports provide users with some contact information, such as emails, postal addresses, and phone numbers of compiling agencies.
- Recommendation:
  - Proactively engage with PSDS users, by surveying users on the degree to which the published statistics are meeting their needs and raising awareness on where users can get assistance and more information.

- Legend: O = Practice Observed; LO = Practice Largely Observed; LNO = Practice Largely Not Observed; NO = Practice Not Observed

*IMF | Technical Report – United Republic of Tanzania Data Quality Assessment for PSDS*

### 31. The new debt management and reporting system provides a platform for accurate,

### 31. The new debt management and reporting system provides a platform for accurate,

### Data migration, validation, and governance
- External and domestic debt data were migrated from the previous debt database system to CS Meridian during 2022/23.
- Under this data migration debt data from 2000 onwards has been brought into the new system.
- Extensive effort was undertaken to validate the data against original contractual documents and reconcile information with creditors.
- Structural changes were made in the MOF in FY2021/22 to allow for the DMD to be subject to annual financial audits.
- Annual financial audits are used as a governance process to ensure the accuracy and reliability of the data on debt managed by the DMD.

### CS Meridian functionality, access, and linkages
- CS Meridian is a web-based system allowing staff in the MOF, BOT and President’s Office – Finance and Planning (POFP) of the Government of Zanzibar to access and report data.
- Security measures, access control, and audit features ensure debt data is held securely and updates are appropriately managed.
- Tanzania has connected CS Meridian to other government data systems to increase efficiency and reduce human error in data entry.
  - Notably linked systems:
    - Government information system, MUSE (“Mfumo wa Ulipaji Serikalini”): Allowing debt servicing information to be shared with CS Meridian and payment confirmation received.
    - Government Securities System (Bank of Tanzania): Allowing auction data related to Treasury bonds and bills to be shared with CS Meridian.
  - POFP (Government of Zanzibar): Domestic Debt of the Government of Zanzibar — none connected.

### Synchronization with World Bank Debtor Reporting System (DRS)
- Transition from CS-DRMS to CS Meridian led to significant advancements in management and validation of external debt data.
- Historical central government external debt data (loan-by-loan) was cleaned, validated, and reconciled with creditors within CS Meridian.
- The external debt data reported to the World Bank’s DRS needs synchronization with the updated data now housed in CS Meridian.
- The DMD and the World Bank have jointly agreed to undertake a comprehensive reconciliation process between the two databases.
  - This includes updating historical data previously reported to the DRS to reflect the cleaned and validated figures in CS Meridian.
  - Assistance will be provided for tracking and updating general government and public corporation (including SOEs) debt in the DRS in accordance with the signed agreement on the World Bank Policy on External Debt Reporting and Financial Statements.
- An output will be a timeseries of revisions between validated CS Meridian data and previously published data.
  - Recommendation: publish the new timeseries for as many years and quarters as possible with commentary explaining causes of revisions.
  - Note: current MOF policy is not to restate the debt for previous quarters in their quarterly debt report; any debt revisions should be explained to users.

### Contingent liabilities and approved borrowing
- MOF publishes detailed data on contingent liabilities of BCG in annexes to their quarterly debt report, including individual guarantees, other contingent liabilities, and outstanding stock of approved borrowing by public entities.
- Comparison of data for June 2024 with December 2023 highlighted some large movements not explained in report metadata/commentary.
  - Footnote (as presented): Over the six months there was a drop in contingent liabilities (excluding the guarantee provided to the National Insurance Corporation) of 1.1 trillion TZS from 3.4 trillion TZS at the end of December 2023, and a drop in the outstanding stock of approved borrowing of 1.0 trillion TZS from 2.3 trillion TZS at the end of December 2023.
- Some contingent liabilities data are collected through spreadsheets and not CS Meridian, introducing potential for errors.
- Recommendation: review data and compilation processes for contingent liabilities to ensure robustness comparable to loans and debt securities; automate where possible.

### Local government, public corporations, and other public sector units
- Comprehensive data source exists for inclusion of local government debt within PSDS as direct local government liabilities.
  - There are 184 local government entities (including city councils, municipal councils, town councils and district councils) and all use a common IT system (MUSE) and chart of accounts.
  - End-of-year local council data review with the President’s Office for Regional Administration and Local Government prior to finalizing financial statements, which are then subject to audit.
  - Local government data is accessible through MUSE; viable to source annual, quarterly, and even monthly data.
- Financial statements are available for:
  - 4 social security funds,
  - 214 (approx.) extrabudgetary units,
  - 86 (approx.) public corporations.
- These public sector units are not currently included in PSDS, but audited timely financial statements provide a strong foundation for future annual PSDS for the entire public sector.
- MOF has a Consolidation Unit experienced in conducting consolidation, which should facilitate expanding PSDS beyond BCG while eliminating double counting.

### Strengths (as identified)
- CS Meridian enforces a variety of checks and validations on granular debt data, highlighting potential issues or inconsistencies to system users.
- Re-structuring of Vote 001 (Public Debt) to allow it to be audited provides a strong governance framework for data accuracy and quality produced by the DMD.
- Government Loans, Guarantees and Grants Act 1974 (as amended) requirements ensure MOF has a comprehensive view of all borrowing activities in the public sector.
- Production of timely audited annual financial statements for all public sector entities provides rich data for instrument and sector expansion of PSDS.
- PSDS are reported with a high degree of periodicity and timeliness; BOT publishes monthly, quarterly, and annual data.
  - Monthly and quarterly data are published within a month of the end of the reporting period and annual data within six months of the end of the year.
- DMD publishes a quarterly debt report significantly expanded and upgraded in 2024; revamped quarterly bulletin (≈40 pages) from July 2024 (three such reports published at time of mission: two in July 2024 and one in September 2024).
- Published PSDS by BOT and DMD are internally consistent and largely consistent with other macroeconomic datasets (e.g., BOP, monetary statistics).

### Recommendations for improvements (as listed)
- Review and, where required, improve the data collection and reporting processes for contingent liabilities and approved borrowing of other public sector entities, automating where possible.
- Publish a historical timeseries comparing the new debt data, as validated and included in CS Meridian, with that previously published to inform users of any revisions.
- When disseminating PSDS for the latest period revise (as required) debt reported for previous periods and provide summary metadata for users to understand reasons for revisions.
- Publish a reconciliation of the change in debt stocks with the transactions and other flows driving the changes. (See Annex III Table A2 for an example).
- Develop and publish a PSDS revision policy to guide users and compilers on the process for reporting and analyzing revisions. (The GFS and PSDS technical committee has a workplan to develop and publish a revision policy by the end of October 2025.)

### Serviceability and reconciliation issues
- Monthly and quarterly BOT publications were fully consistent internally; differences in debt stock values between BOT and DMD publications reflect revisions due to updated source data and differing exchange rates used by BOT and DMD.
- Quarterly BCG debt reports include information on debt inflows and outflows, but insufficient detail to fully reconcile drivers for movements in debt stocks.
  - International statistical standards recommend publishing a table explicitly showing changes in debt and related transactions and other flows (example in Appendix III Table A.2).
- Recommendation: publish explicit reconciliation tables linking changes in debt stocks with transactions and other flows.

### Accessibility and dissemination
- Debt data on MOF and BOT national websites are accessible; reports are downloadable in .pdf format.
  - BOT reports are editable allowing table/chart extraction; DMD reports are published in non-editable format, making extraction harder.
  - No PSDS currently made available on national websites in machine-readable formats such as SDMX, Excel or .csv (except Tanzania PSDS on the EAC website can be extracted in these formats).
- Wider international dissemination gaps:
  - Tanzania National Summary Data Page (NSDP) established October 2016, but latest debt data on portal relate to December 2015 and metadata not updated since 2014.
  - Tanzania has never disseminated data via the joint IMF-World Bank QPSD.
- Commentary in DMD and BOT publications adds value, but additional metadata should be included:
  - Valuation of debt stocks,
  - Sector coverage,
  - Exchange rate used for converting foreign currency debt,
  - Approach to identifying residency.
- Advanced release calendar: National Bureau of Statistics manages one, but it was out-of-date, having ended in June 2024.
- Contact information is provided in publications (e-mails, postal addresses, phone numbers), but mission found no evidence of assistance provided to data users or materials to raise awareness of PSDS.
- Recommendations for dissemination improvements:
  - Resume publication of BCG gross debt in the NSDP and maintain it consistently with expected periodicity and timeliness.
  - Subscribe to the QPSD to disseminate debt data to a wider audience and enable cross-country comparisons.
  - Prepare, publish, and sustain structured metadata on PSDS (use existing NSDP or QPSD metadata templates).
  - Proactively engage with PSDS users (survey users, raise awareness, provide assistance information).

*IMF | Technical Report – United Republic of Tanzania Data Quality Assessment for PSDS*

### APPENDIX II. USERS’ SURVEY RESULTS

### APPENDIX II. USERS’ SURVEY RESULTS

### Overall user feedback
- The feedback from users was generally positive.  
- Most users found the disseminated PSDS to be analytically "very useful" or "useful" in their work.
- A majority either "agreed" or "strongly agreed" that the PSDS was unbiased and accurate.
- Users gave similar assessments regarding the timeliness and level of detail of the disseminated PSDS.

### Primary uses and data sources
- Primary interests in the disseminated PSDS:
  - Investment decision-making
  - Compilation of other macroeconomic statistics
  - Academic research
  - Policymaking
- From where users obtain PSDS and related data (responses):
  - Official national debt bulletins and press releases: 12
  - IMF or World Bank databases and/or publications: 12
  - Other international organization databases and/or publications: 0
  - Media reports: 2
  - Data published on the National Summary Data Page (NSDP): 4
  - Private sector summaries and analyses: 1
  - Other sources: 2

### Respondent profile and consent
- Consent to quote responses: Yes = 13, No = 4
- Responders by institution:
  - International organizations: 4
  - Financial institutes: 2
  - Foreign governments: 2
  - Research institutes: 1
  - Audit agencies: 4
  - Public entities: 4
- Users’ stated main interest in PSDS data (counts):
  - Policymaking: 3
  - Compilation of other macroeconomic statistics: 4
  - Investment decisions: 6
  - Other analytical work: 1
  - Academia or research: 3
  - No response: 0

### Overall assessment (survey responses)
- Analytic usefulness of disseminated PSDS:
  - Very useful: 10
  - Useful: 6
  - Somewhat useful: 1
  - Not useful: 0
  - Undecided: 0
- Agreement with "The disseminated PSDS are unbiased and accurate":
  - Strongly agree: 2
  - Agree: 3
  - Neither agree nor disagree: 10
  - Disagree: 1
  - Strongly disagree: 1
- Satisfaction with timeliness and level of detail:
  - Very satisfied: 1
  - Satisfied: 14
  - Neither satisfied nor dissatisfied: 1
  - Dissatisfied: 1
  - Very dissatisfied: 0

### Coverage and concepts
- Satisfaction with sector coverage (inclusion/exclusion of budgetary unit, extrabudgetary entities, subnational governments, public corporations):
  - Very satisfied: 2
  - Satisfied: 12
  - Neither satisfied nor dissatisfied: 0
  - Dissatisfied: 3
  - Very dissatisfied: 0
- Satisfaction with coverage of debt instruments (loans, bonds, arrears, accounts payable, pension liabilities):
  - Very satisfied: 1
  - Satisfied: 12
  - Neither satisfied nor dissatisfied: 1
  - Dissatisfied: 3
  - Very dissatisfied: 0
- Satisfaction with coverage of public nonfinancial corporations (state-owned enterprises):
  - Very satisfied: 1
  - Satisfied: 8
  - Neither satisfied nor dissatisfied: 3
  - Dissatisfied: 4
  - Very dissatisfied: 1
- Which debt instruments are inadequately covered (responses):
  - None – coverage is adequate for all debt instruments: 3
  - Accounts payables (arrears): 6
  - Loans: 7
  - Debt securities: 3
  - Pension liabilities: 2
  - Other: 1
- For which public-sector sub-sectors more coverage is desired (responses):
  - None – coverage is adequate for all sub-sectors of the public sector: 1
  - Local governments / State governments: 15
  - Public corporations: 12
  - Budgetary central government: 6
  - Extrabudgetary units: 6
  - Other: 1
- Valuation appropriateness of outstanding stock positions (e.g., market, nominal, face, principal outstanding):
  - Yes: 15
  - No: 1

### Periodicity and timeliness
- Satisfaction with periodicity of disseminated PSDS (monthly, quarterly, annual):
  - Very satisfied: 2
  - Satisfied: 11
  - Neither satisfied nor dissatisfied: 1
  - Dissatisfied: 1
  - Very dissatisfied: 2
- Are PSDS disseminated in a sufficiently timely manner (time lag after the period to which the PSDS pertain)?
  - Yes: 11
  - No: 3
  - Not sure: 3

### Dissemination practices
- Is there a published release schedule/calendar and are PSDS released punctually?
  - Yes, and the data is released on schedule: 5
  - Yes, and the data is sometimes released on schedule: 5
  - Yes, and the data is never released on schedule: 2
  - No: 1
  - Not sure: 4
- Is analysis published to identify and explain causes of revisions, breaks, adjustments, discontinuities, or unusual changes?
  - Yes: 8
  - No: 3
  - Not sure: 5
  - No response: 1
- At time of PSDS dissemination, are users informed whether data are preliminary, revised or updated?
  - Always informed: 7
  - Sometimes informed: 5
  - Never informed: 2
  - Not aware: 3
- Do users receive advance notice of major changes (coverage, methods, data sources)?
  - Always: 2
  - Sometimes: 6
  - Never: 4
  - Not sure: 5

### Accessibility and usefulness
- Agreement that disseminated PSDS are presented to facilitate proper interpretation and meaningful comparisons:
  - Strongly agree: 2
  - Agree: 14
  - Neither agree nor disagree: 0
  - Disagree: 1
  - Strongly disagree: 0
- Availability of comprehensive published source and methods document or metadata:
  - Yes, available and accessible: 6
  - Partial information available: 5
  - No information available: 2
  - Not sure: 4
- Consistency and reconcilability over time with other macroeconomic data sets (government finance statistics, national accounts, international investment position, external debt statistics, monetary and financial):
  - Fully consistent: 6
  - Partly consistent: 7
  - Not consistent: 1
  - Not sure: 3

*Source: APPENDIX II. USERS’ SURVEY RESULTS (IMF Technical Report – United Republic of Tanzania Data Quality Assessment for PSDS).*

---


_Source: https://www.imf.org/-/media/files/publications/tar/2025/english/tarea2025027-print-pdf.pdf_