Report on the Observance of Standards and Codes-Data Module, Response by the Authorities, and Detailed Assessments Using Data Quality Assessment Framework

The data module of this Report on the Observance of Standards and Codes (ROSC) of Tanzania provides a summary assessment of practices on the coverage, periodicity, and timeliness of the data categories against the IMF’s General Data Dissemination System (GDDS). It is complemented by an assessment of the quality of national accounts, consumer and producer price indices, and government finance, monetary, and balance-of-payments statistics, using the Data Quality Assessment Framework (DQAF) developed by the IMF’s Statistics Department (STA).


The data module of this Report on the Observance of Standards and Codes (ROSC) of Tanzania provides a summary assessment of practices on the coverage, periodicity, and timeliness of the data categories against the IMF’s General Data Dissemination System (GDDS). It is complemented by an assessment of the quality of national accounts, consumer and producer price indices, and government finance, monetary, and balance-of-payments statistics, using the Data Quality Assessment Framework (DQAF) developed by the IMF’s Statistics Department (STA).

I. Introduction

1. The data module of this Report on the Observance of Standards and Codes (ROSC) provides a summary assessment of practices on the coverage, periodicity, and timeliness of the data categories against the IMF’s General Data Dissemination System (GDDS).1 It is complemented by an assessment of the quality of national accounts, consumer and producer price indices, and government finance, monetary, and balance of payments statistics, using the Data Quality Assessment Framework (DQAF) developed by the IMF’s Statistics Department (STA). This report is based on public information, as well as information provided during a staff mission during October 8–23, 2002.2

2. Section II provides an overview of the GDDS and an assessment of Tanzania’s data dissemination practices against the GDDS. Section III presents a summary assessment of the quality of the principal macroeconomic datasets, following the dataset-specific assessment frameworks. Finally, Section IV sets out recommendations for further improvements in Tanzania’s macroeconomic statistics.

II. Data Dissemination Practices and the GDDS: Current Dissemination Practices

3. Tanzania’s data dissemination practices are assessed against the GDDS.3 Tanzania’s participation in the GDDS formally started with the posting of its metadata on the DSBB on July 6, 2001.

4. Macroeconomic statistics in Tanzania are mainly produced by three institutions: (i) the National Bureau of Statistics (NBS) is responsible for the national accounts, prices, other economic statistics, and socio-demographic statistics; (ii) the Bank of Tanzania (BoT) is responsible for monetary and balance of payments statistics; and (iii) the Ministry of Finance (MOF) is responsible for statistics on central government operations.

5. Access to macroeconomic data is provided through several publications and at the following Internet websites:

Data dimension: coverage, periodicity, and timeliness

6. Tanzania meets the GDDS recommendations for the coverage for all comprehensive frameworks and for many of the core indicators—the notable exceptions being (i) the manufacturing or industrial production index, (ii) the primary commodity, agricultural or other indices, (iii) labor market statistics, and (iv) and a maturity breakdown and debt schedule for the public and publicly guaranteed external debt outstanding. Tanzania does not meet the coverage recommendations for either of the two encouraged comprehensive frameworks, namely, general government operations and the international investment position. However, it meets the GDDS recommendations for the coverage of the encouraged core indicators, with the exception of (i) the producer prices index, (ii) government guaranteed debt, and (iii) reserve-related liabilities.

7. Regarding the periodicity of the data, Tanzania meets the GDDS recommendations for all comprehensive frameworks, and exceeds the GDDS periodicity in the case of the national accounts, central government operations, and central government debt. Tanzania also meets or exceeds the GDDS recommendations for periodicity for those core indicators and encouraged core indicators for which data are disseminated.

8. The GDDS recommendations for timeliness are met, and indeed exceeded, for all comprehensive frameworks, except the depository corporations survey, where the timeliness can be up to five months due to delays in the publication of the BoT’s Monthly Economic Review in which the data are disseminated. The timeliness recommendations of the GDDS are also met or exceeded for the core indicators, with the exception of (i) broad money and credit aggregates, (ii) central bank aggregates, (iii) gross official reserves, and (iv) total exports and imports. Again, delays in the publication of the Monthly Economic Review are the reason for the timeliness recommendations not being met for all these core indicators, except the gross official reserves. Tanzania meets or exceeds the GDDS recommendations for timeliness for those encouraged core indicators for which data are disseminated. Table 1 gives an overview of current practices regarding the coverage, periodicity, and timeliness of the data for Tanzania compared with the GDDS recommendations.

Table 1.

Overview of Current Practices Regarding Coverage, Periodicity, and Timeliness of Data Compared to the General Data Dissemination System

article image
article image
Italics indicate encouraged categories

Quality dimension

9. The quality dimension of the GDDS is particularly important, given that the primary focus is on improvements in data quality over time. The assessment of quality is based on the dissemination of (i) documentation on methodology and data sources, and (ii) component detail and reconciliation with related data. Although summary documentation on the methodology for all macroeconomic statistics in Tanzania is available on the DSBB, detailed documentation on concepts and methods is not available from national publications, nor is the DSBB publicized in the websites of the compiling institutions.

Integrity dimension

10. The GDDS recommends the disclosure of the legal framework for the collection, compilation and dissemination of data, including the confidentiality of the collected data. The terms and conditions under which most official statistics are compiled and disseminated in Tanzania provide a legal framework that supports the integrity of the statistical system. Procedures on the government’s internal access to the data prior to public release are, by and large, not disseminated on the DSBB or in national publications.

Access dimension

11. Dissemination of official statistics is an essential feature of statistics as a public good. Ready and equal access are principal requirements for the public, including market participants. The access dimension is based on two practices that facilitate ready and equal access to data: (i) dissemination of advance release calendars; and (ii) simultaneous release to all parties. Advance release calendars are not provided to the users of Tanzania’s statistics although, at least in some cases, internal fixed-term release schedules are followed. Dissemination processes, whereby select users obtain access to statistics prior to other users, are problematic.

Plans for improvement

12. Plans for improving data quality are an integral part of the GDDS. Updates to the GDDS plans for improvement are shown in Tables 1 and 2.

Table 2.

Data Quality Assessment Framework: Summary Presentation of Results

article image
article image

Comparison with the SDDS requirements

13. The IMF’s Special Data Dissemination Standard (SDDS) is a “best practice” disclosure standard that focuses on encouraging the authorities to provide information to users, including information that will enable them to access data. A review of Tanzania’s data dissemination practices against SDDS requirements shows that Tanzania currently fails to meet a large number of these requirements. For example, national accounts are compiled with a semi annual periodicity as against the quarterly periodicity required under the SDDS; the present compilation of the IIP is only partial; analytical accounts of the central bank and the commercial banks are disseminated with a lag of two to five months against the SDDS requirement of two weeks for the accounts of the central bank and one month for the accounts of the banking sector; and the balance of payments are compiled with an annual periodicity as against the quarterly periodicity required under the SDDS. The latter also requires compilation and dissemination of the IMF’s reserve template, and further development of general government statistics. Compliance with the SDDS requirements will therefore pose a significant challenge to Tanzania’s capacity for mobilizing statistical effort. However, by rigorously implementing an action plan, Tanzania should be in a position to subscribe to the SDDS within five years.

III. Summary Assessment of Data Quality

14. Since mid-2001, the IMF complements the GDDS and SDDS elements of the ROSC data module with an assessment of data quality based on the IMF’s DQAF. The DQAF comprises a generic framework, and a set of dataset-specific frameworks. The frameworks cover a set of prerequisites and five dimensions of data quality—integrity, methodological soundness, accuracy and reliability, serviceability, and accessibility.4

15. An assessment of six macroeconomic datasets (national accounts, consumer price index (CPI), producer price index (PPI), and government finance, monetary, and the balance of payments statistics) was conducted using the frames of reference provided by the dataset-specific DQAFs. Conclusions resulting from the application of this framework to Tanzania’s statistical system are presented below. These are also presented in the form of a summary table in which the assessment of data practices is made on a qualitative basis, using a four part scale (Table 2).

16. Tanzania’s macroeconomic statistics and statistical base are in general adequate for effective surveillance. Nevertheless, IMF staff identified shortcomings in some statistical practices that have the potential for detracting from the accurate and timely analysis of economic and financial developments and the formulation of appropriate policies.

Prerequisites of quality

This category in the DQAF identifies conditions within the agency in charge of producing statistics that have an impact on data quality. The elements within the category refer to the legal and institutional environment, resources, and quality awareness.

17. The main statistics producing agencies in Tanzania—the NBS, the MOF, and the BoT—have an adequate legal and institutional framework that supports data quality. The Statistics Act, 2002 and the Bank of Tanzania Act, 1995 (BoT Act) respectively specify the responsibility of the NBS for the compilation of national accounts and price statistics, and that of the BoT for monetary statistics, and contain legal powers to enforce their mandates. The Public Finance Act, 2001 assigns the Minister of Finance responsibility for presenting a full set of government accounts to the National Assembly, although there is no mandate to produce comprehensive government finance statistics to international standards. The BoT Act also in effect provides the BoT with authority to collect the balance of payments statistics, although this authority is somewhat undermined by the absence of legal powers to collect information from outside the financial sector. Data sharing and coordination among the various data sharing agencies are adequate, although formal arrangements have been developed only as regards the government finance statistics. Confidentiality of respondents’ data is protected by law except in a number of well defined cases.

18. Recent recognition by the Tanzanian authorities of the importance of statistics as a critical element of a country’s economic infrastructure has led to a significant improvement in the resources available to the NBS. Nevertheless, the NBS continues to need to find outside financing even for such basic needs as the maintenance of its buildings. Moreover, lack of resources is likely to constrain the envisaged necessary development program in national accounts, and appears to be a factor underlying the absence of wide dissemination of, for instance, the PPI. Resources devoted to statistics remain limited at the MOF, constraining the authorities’ ability to develop fully a comprehensive set of government finance statistics beyond the budget execution data, and indeed leading the Minister of Finance to warn that shortages of professional staff were threatening to jeopardize the government’s reform of financial management. At the BoT, resources seem broadly adequate for the compilation of the monetary and balance of payments statistics. At the NBS and BoT, measures are in place to ensure the efficient use of resources for the compilation of data. At the MOF, the costs of compiling government statistics are not yet measured; moreover, if the consultants in the MOF are not used to pass on their expertise to MOF staff, the sustainability of recent statistical development may be jeopardized after their departure.

19. In past years, all three agencies have made significant progress in implementing work programs that are supportive of data compilation in accordance with the international standards. Quality awareness is indicated in the NBS by a mission statement focusing on quality, by recent exercises of self-assessment, and by commissioning external reviews. At the MOF quality awareness is fostered by internal audits as well as the need to have the data audited by the Controller and Auditor-General and to submit them to the National Assembly. The BoT conducts data reviews and cross checks, and seeks to educate suppliers in the importance of high-quality data. Some recent problems in the balance of payments data, however, suggest that these processes may not always have been fully effective.


Integrity identifies features that support firm adherence to objectivity in the collection, compilation, and dissemination of statistics so as to maintain users’ confidence. Elements refer to the professionalism and ethical standards that should guide policies and practices, which should be reinforced by their transparency.

20. The laws covering the statistical activities of the NBS, MOF, and BoT include provisions to encourage professionalism. The status of the NBS as a government executive agency, and provisions for the independence of the BoT, provide statistical compilers in these agencies with protection from undue influence. Compilers at the MOF are protected by the provisions for public accountability; the Public Finance Act, 2001 stipulates in turn that employees of the national audit office should perform their functions without outside interference. Professionalism is also fostered in all compiling agencies by recruitment and promotion of staff on a competitive basis and on the basis of professional qualifications, and by encouraging participation in outside training. Choices of sources and statistical techniques in all compiling agencies are guided solely by statistical considerations.

21. As to transparency, the laws and codes governing the collection, compilation, and dissemination of statistics as applicable to each of the compiling agencies are widely available to the public. However, it is not disclosed to the public that national accounts, and price and balance of payments statistics are provided to the government before they are released to the public. There is only a partial reference to this effect on the IMF’s DSBB. Monthly reports on government finance are compiled for use within the government sector, but are made available to the general public on request, although this availability is not widely disclosed. The PPI is not disseminated beyond major users. MOF participation in the BoT’s Monetary Policy Committee provides access to monetary statistics before publication, which is not explicitly disclosed. Users are generally notified of major changes in methodology, source data, and statistical techniques when they occur; however, no advance notice of such changes is formally provided. In most cases the products of the statistical agencies are clearly identified as such. Ethical standards for staff behavior in the NBS and MOF are set out in the Civil Service Department Standing Orders, 1994. The NBS has also developed a Code of Ethics that is attached to the NBS Staff Regulations. BoT staff are required to abide by regulations published in its Staff Regulations, 2002. These documents are all well known to the staff concerned.

Methodological soundness

Methodological soundness refers to the application of international standards, guidelines, and agreed practices. Application of such standards, which are specific to the dataset, is indicative of the soundness of the data and fosters international comparability. Elements refer to the basic building blocks of concepts and definitions, scope, classification and sectorization, and basis for recording.

22. The national accounts follow the concepts and definitions of System of National Accounts 1968 (1968 SNA), as well as some elements of the System of National Accounts 1993 (1993 SNA). The minimum requirements for the scope of the national accounts tables, established by the Intersecretariat Working Group on National Accounts, are met. The internationally reported tables cover Tanzania Mainland only, and exclude Zanzibar for which only provisional data exist. Quarterly national accounts and annual supply and use tables are not yet compiled. The production and assets boundaries are not fully in accordance with the 1993 SNA recommendations—for instance, production of intangible assets is not included. The classification/sectorization is broadly consistent with internationally accepted standards, with the exception that household final consumption expenditures are not classified by purpose. The basis for recording is partially in accordance with the 1993 SNA. The exceptions are market output, which is valued at factor costs instead of at basic or producer’s prices, and government and some external transactions, which are recorded on a cash basis, rather than on an accrual basis. Grossing and netting procedures are adhered to in so far as the source data permit.

23. The overall structures in terms of concepts and definitions of both the CPI and the PPI are in broad accordance with international good practice. However, in contrast to international guidelines, their scope excludes Zanzibar. The PPI is further restricted to measuring changes in the output prices of large manufacturing establishments. The functional classification of goods and services in the CPI is in accordance with the 1993 SNA, but the PPI still uses the outdated International Standard Industrial Classification of All Economic Activities, Rev. 2. As regards the basis for recording, the CPI has the disadvantage of only measuring prices that consumers have to pay in the first half of the month. In particular, the prices collected for fresh food items are not representative for the full month. The PPI suffers from lack of specificity in the definition of the type of prices collected.

24. The concepts and definitions used by the MOF to compile government finance statistics are prescribed by the budget guidelines, which are not consistent with A Manual of Government Finance Statistics 1986 (GFSM 1986). At the aggregated level, the overall structure makes a distinction between recurrent revenue and grants (foreign donor funding) and between recurrent expenditure and development expenditure without identifying lending minus repayment. Net financing is presented, but includes items such as privatization and payments in arrears. The scope of data covered in government finance statistics includes the transactions of the central government of Tanzania Mainland to the extent that all receipts of budgetary and extrabudgetary units of government are included in revenue. Expenditure includes all budgetary expenditure and total transfers to Zanzibar, extrabudgetary units and local governments. However, detailed information of the nature of all revenue and expenditure is not presented according to the full scope of tables as recommended by the international standard. Local government transfers are not separately identifiable in the expenditure data, and payments such as salaries of local government employees are included in central government expenditure. Although data on the Zanzibar government finances are separately disseminated, it is not possible to consolidate the data with those of the mainland. The scope of the data is not yet extended to the general government, since data on local government own revenue and corresponding expenditure are not available. The classification/sectorization differs in significant ways from the international standards. The data on nontax revenue include repayments of loans extended for policy purposes, capital revenue, and gross revenue of departmental enterprises. The data on expenditure do not separately identify current and capital expenditure, and include all wages and salaries paid from the centralized payroll system, as well as loans extended for policy purposes. Net financing includes privatization proceeds and payments for arrears. Debt data include debt guarantees of government. The basis for recording of transactions is checks received and issued at cash values, in line with the GFSM 1986 recommendations. However, cash recording is not strictly applied as grants received include grants received in kind, which contributes to inconsistencies. The grossing/netting procedures outlined in the GFSM 1986 are broadly adhered to, with the exception of lending minus repayment and the transactions of departmental enterprises.

25. Although the analytical framework adopted by the BoT for monetary statistics reflects concepts and definitions based on the Monetary and Financial Statistics Manual (MFSM), the transition to MFSM methodology has not been completed. The scope of depository corporations survey is currently limited to commercial banks, and excludes rural banks and nonbank deposit-taking institutions. Sectorization of domestic credit is not adequate, and accrued interest on assets and liabilities is currently treated in a highly aggregated manner. The basis for recording flows and stocks is largely consistent with MFSM and, although the general valuation of asset and liability positions is at current market-determined prices and exchange rates, the valuation of foreign securities reflects face values rather than market prices.

26. The concepts and definitions used in compiling the balance of payments statistics conform to the fifth edition of the Balance of Payments Manual (BPM5) with only minor exceptions. In principle, the scope of the data covers all transactions between the residents of the United Republic of Tanzania, including Zanzibar and the export processing zones, and the rest of the world. However, there are some coverage gaps, such as unrecorded trade. The classification/sectorization of transactions is largely in accordance with BPM5, with the exceptions of the sectorization of debt transactions, and the classification of technical assistance, income items, and current transfer debits. The basis for recording is cash. Market prices are used to value flows and stocks, and grossing/netting procedures are consistent with BPM5, with minor exceptions.

Accuracy and reliability

Accuracy and reliability identifies features that contribute to the goal that data portray reality. Elements refer to identified features of the source data, statistical techniques, and supporting assessments and validation.

27. Except for the agriculture and public agencies, the availability of source data for the national accounts is rather limited, partly owing to the inadequate business register and lack of periodic surveys. The lack of sources is most severe for the informal sector, production of services, and household consumption expenditure. The available source data are generally consistent with the definitions, scope, and classifications needed for national accounts purposes, and are available on a timely basis. Tanzania had a broadly comprehensive range of economic statistics around 1992, the year upon which the present GDP estimates are based. Since that baseline the statistical techniques have been largely based on extrapolation. Intermediate consumption is in most cases compiled as a fixed ratio of output, and the single indicator method is used to estimate GDP at constant prices. The deflation techniques for general government expenditures and subsidies are not in line with good practice. Estimates for informal and hidden activities are made. Although the assessment and validation of the source data and intermediate data guide statistical improvements, no supply and use framework exists for investigating discrepancies in the statistical output. Revision studies are not routinely undertaken.

28. The source data for the CPI and the PPI are not always based on precise definitions for the items for which prices are collected. Furthermore, the basic data for CPI items do not necessarily measure prices at regular monthly intervals. The basic data for the PPI are collected quarterly rather than monthly. Regarding statistical techniques, both indices use an appropriate formula to aggregate elementary indices. However, both have outdated base years and lack exhaustive documentation. The technique to compile the CPI does not distinguish rigorously between quality changes and price changes. Although on one particular occasion in the past new products were introduced in the PPI, this has not been done since. If basic price information for the PPI is missing, the last quoted prices are carried forward for overly extended periods. The assessment and validation of basic data mostly take place at the NBS headquarters, which causes problems for correcting misreported prices, particularly for the CPI. The importance of revisions is limited for the CPI and the PPI. The analysis of revisions does not feed into the process of imputing missing prices for the PPI.

29. Source data for the government finance statistics are comprehensive for data on recurrent revenue and expenditure of budgetary accounts, and broadly adhere to the international standards. However, comprehensive source data on extrabudgetary and development expenditure are not available. Source data based on the administrative records of government are timely, and do not rely on statistical techniques. The assessment and validation of source data are routinely undertaken, and the data are reconciled with other sources. The assessment and validation of intermediate data and statistical outputs are undertaken, and have led to improvements in these data. Authorities investigate discrepancies and are aware of potential problems in statistical outputs. Efforts such as improved systems, capacity building, periodic internal audits, and external audits were introduced in order to improve the output data. Revisions of data are analyzed and incorporated whenever improved information becomes available. Revision studies take the form of memorandums for files.

30. The source data for compiling monetary statistics are derived from the accounting records of the BoT and commercial banks. While the balance sheet data of the BoT are provided by its Domestic Accounts Department, those of the commercial banks are derived from the monthly report form that is completed and returned to the Banks Supervision Department by all commercial banks in Tanzania. The source data, however, do not capture the full range of economic sectors as defined under MFSM methodology. The system allows timely compilation of monetary statistics. No statistical techniques are used. Assessment and validation of source data are carried out by the Banks Supervision Department. Data revisions, which are usually small, do not affect the important statistical aggregates, with the result that no revision studies are undertaken.

31. The source data for the balance of payments statistics are timely and relatively comprehensive, except for the sources for transfers and direct and portfolio investment. Although the statistical techniques employed for data compilation are, in principle, sound, the data review process is not rigorously followed, and the disseminated data contain numerous errors. The statistical techniques used for the adjustment from cost, insurance, and freight (c.i.f.) to free on board (f.o.b.) imports data are reasonably adequate, but no adjustments are made for unrecorded trade, or for the removal of double-counting of current transfer debits. Assessment and validation of source data are undertaken and have led to improvements in the source data, such as a survey of foreign capital flows, and travel surveys, both of which are about to replace the cash-based International Transactions Recording System (ITRS) or investment approvals as data sources for those items. The assessment and validation of intermediate data and statistical outputs are not comprehensive. Main intermediate data, such as exports, are validated against other applicable information, such as crop harvests and commodity prices. However, the behavior of data series is not routinely cross-checked and investigated, and long-standing inconsistencies, such as those in debt arrears, have not yet been addressed in the disseminated data. In addition, statistical discrepancies in the compiled data are not always monitored and investigated, with the result that significant problems with the “migration” of historical data to a new compilation database were not noticed for some time. No formal revision studies are conducted.


Serviceability focuses on practical aspects of how well a dataset meet users’ needs. Elements refer to the extent to which data are relevant, produced, and disseminated in a timely fashion with appropriate periodicity, are consistent internally and with other datasets, and follow a predictable revisions policy.

32. There are several channels by which the NBS is able to consult users about the relevance, and ensure that the statistical series, including national accounts, address issues of concern for the users. For instance, the NBS has instituted several technical commissions or user/producer workshops, in which donors, government agencies, other users, and providers of data are represented. The periodicity and timeliness of the GDP estimates meet the GDDS requirements. Consistent time series are available from 1987 onwards. The national accounts statistics are reportedly consistent with the balance of payments and government finance statistics, but discrepancies between the national accounts data and the balance of payments data do exist, apparently, at least in part, because of differences in revisions policies. The major revisions of national accounts do not follow a predetermined cycle, and advance notice of revisions is not given. Although major changes in methodology and sources are explained to the users, the national accounts annual publication does not identify provisional data, in order to alert users that these data are subject to revision.

33. No periodic monitoring of user needs takes place to assess the continued relevance of the CPI and PPI. The CPI meets both the encouraged timeliness and periodicity of GDDS, and even meets the requirements of the SDDS. However, the PPI is not published at all and, according to existing plans, would only be disseminated one month in each quarter. Both indices are internally fully consistent. The PPI data that are made available to certain users do not identify preliminary data. Furthermore, no studies are made at the time of major revisions of the PPI.

34. The relevance of government finance statistics is ensured by the fact that the budget documentation uses the same analytical framework. However, the data have limited relevance for international comparisons, and are not monitored against the needs of users other than those in the government. The timeliness and periodicity of government finance statistics meet the GDDS requirements. Although there is consistency within the dataset, alternative presentations in various publications use different analytical frameworks. Data are consistent over a reasonable period of time, and publications clearly identify breaks in the time series. Reconciliation of the government finance statistics with other datasets is hampered by the differences in coverage. The revision policy and practice for final annual data are well established and transparent. Annual data are revised when the audited results for all government accounts are available, and the audit reports are presented to the National Assembly and widely discussed. However, revisions are performed during the year whenever changes occur, and do not follow a revision schedule. Revised data are not clearly distinguished from other data, and the status of data is not clearly identified as being preliminary or audited.

35. The relevance of monetary statistics in meeting users’ needs is not monitored. Monthly central bank data, as well as monthly commercial bank data, are disseminated with a timeliness of between two and five months after the reference period, which is not fully consistent with the GDDS recommendations on timeliness and periodicity. Monetary statistics show consistency within the dataset. They are also consistent with the government finance and balance of payments statistics, as well as over time. Revisions to monetary statistics are carried out, but since these are usually largely insignificant, users are not informed on the BoT’s revision policy.

36. There is no formal user advisory group to monitor the relevance of the existing balance of payments statistics in meeting users’ needs. The timeliness and periodicity exceed the GDDS recommendation in the case of the former, and meet the GDDS recommendations in the case of the latter. The data are reasonably consistent within the dataset and consistency in the data series is maintained for a reasonable period of time. However, there are inconsistencies between the balance of payments estimates of transfers and the MOF data on these transactions, primarily because of coverage gaps in the MOF data. There are also inconsistencies with the balance of payments data published by the President’s Office—Planning and Privatization (POPP), and between the balance of payments data on transactions related to the debt arrears and the Debt Management Department stock data on debt arrears. The balance of payments data are, however, consistent with the external sector of the national accounts, for which they are the primary data source, as well as with the monetary statistics. In the case of revision policy and practice, the policy, which is not disseminated, is to incorporate revisions whenever changes in the data are received, and to publish any revisions to the annual data without further delay in the BoT’s Monthly Economic Review. In practice, revisions, which are frequent, are not identified in the publications, and the public are not informed of the reasons for revision. No studies of revisions are made.


Accessibility deals with the availability of information to users. Elements refers to the extent to which data and metadata are clear and easily available and to which assistance to the user is adequate to help them find and use the data.

37. Regarding data accessibility, the national accounts publication presents the statistics in a largely clear manner. However, the publication does not include any analysis of recent developments. The national accounts data are disseminated simultaneously to all users through the Minister’s Budget Speech. Nonpublished, but nonconfidential, subaggregates of the statistics are made available to users upon request. There is at present no access to data via the Internet. Extensive documentation on concepts and methods is published whenever the base year is changed, and the metadata are made available upon request. The NBS’s assistance to users suffers from lack of guidance on how to obtain additional information and from the absence of a catalog of products.

38. While the presentation in CPI releases is clear, the data accessibility is not optimal because preannounced release schedules are lacking. In addition, the NBS serves main users better than the general public. This is particularly true for the PPI, where dissemination is nearly exclusively restricted to a small number of preferred users. The metadata accessibility could be improved significantly. In the case of both the CPI and the PPI, the availability of documentation is not publicized, and no metadata are available that are directed at non-specialist users. While the NBS has sufficiently qualified staff to answer queries from users, no specific contact is publicized with the data, and the assistance to users further suffers from the lack of a catalog of publications.

39. Data accessibility for the government finance statistics is mainly achieved through the dissemination of electronic reports on government operations and through the BoT’s quarterly Economic Bulletin tables on debt and government operations. Data on central government operations are presented in a way that facilitates proper interpretation and comparison with budget documentation. However, the layout does not facilitate international comparisons. The dissemination media and format for central government operations are adequate. However, the format in which data on central government debt is presented is not comprehensive and does not disclose total liabilities of central government. The central government metadata accessibility is limited to the DSBB of the IMF. Assistance to users is provided through the contact details disseminated on the metadata pages of the DSBB. The official website of the Tanzanian government provides a general link for inquiries.

40. Data accessibility for the monetary statistics is primarily achieved through the BoT’s publications, in particular the Monthly Economic Review, which is also available on the BoT’s website. Data are not simultaneously released to the public, as draft copies of the Monthly Economic Review are distributed to the major newspapers, as well as to selected embassies. Although some methodological notes are disseminated in the BoT publications, no comprehensive metadata are published in national publications. Consequently, metadata accessibility is limited to the GDDS metadata posted on the DSBB. Assistance to users is achieved through the provision of information on a contact person in the quarterly Economic Bulletin, and the DSBB. A catalog of BoT’s publications is available on the BoT website. All BoT publications are made available free of charge.

41. Data accessibility is largely insufficient for the balance of payments statistics. Although the statistics are made available to all users at the same time through the release of the BoT’s quarterly Economic Bulletin, the presentation of the data tables has numerous problems that hinder proper interpretation and meaningful comparison. The dissemination media present only highly aggregated data that do not enable users to assess the accuracy of the data. Although it is intended to publish an expanded version of the balance of payments in the next edition of the quarterly Economic Bulletin, no consideration has been given to disseminating on the BoT website the full range of detailed data that are available. Nonpublished, nonconfidential, subaggregates are not made available on request. There is also no advance release calendar, and the timeliness of the publication in which the data are disseminated varies. Metadata accessibility is largely achieved, as documentation on concepts, scope, classifications, data sources, etc., is disseminated in the GDDS metadata on the DSBB, and a shorter version is included in the IMF Balance of Payments Statistics Yearbook. However, national publications do not include metadata, and the public are not informed of the metadata disseminated on the DSBB or in the Balance of Payments Statistics Yearbook. The levels of detail in the metadata are not adapted to the needs of the different audiences. Assistance to users is achieved through the provision of information on a contact person in the quarterly Economic Bulletin, and on the DSBB. A catalog of BoT’s publications is available on the BoT website. All BoT publications are made available free of charge.

IV. Staff Recommendations

42. Based on the results of the data quality assessment, discussions with the Tanzanian authorities in the respective statistics-compiling institutions, and responses from data users, the measures below are proposed to further enhance Tanzania’s adherence to international statistical standards.

General Recommendations

High Priority

  • Publicly identify internal governmental access to data prior to release.

  • Disseminate advance release calendars. Release statistics simultaneously to all users.

  • Make the revision policies of all institutions known to the public.

  • Conduct surveys among users to receive feedback on the quality of statistics.

  • Identify preliminary data on a consistent basis in all publications.

  • Regularly update and enhance the GDDS metadata posted on the DSBB for all sectors.

  • Commission a plan of action for subscription to SDDS in the medium-term.

Other Key Recommendations

  • Identify the compiling agency as a source of statistics on a consistent basis in all publications.

  • Inform users in advance of major changes in methodology and source data.

  • Identify the contact person for statistics in all publications.

National Accounts

High Priority

  • Disseminate a “flash report” of the most important GDP tables at the same time as the provisional GDP estimates are provided to the President’s Office-Planning and Privatization (POPP).

  • Implement 1993 SNA as soon as possible.

  • Implement new benchmarks and base year as soon as possible.

  • Ensure that the business register is regularly maintained after the updating of the business census 2003.

  • Include Zanzibar in the national accounts.

  • Establish quarterly national accounts in the medium-term.

  • Develop new short-term statistics (indicators), especially related to service production and household final consumption expenditure.

Other Key Recommendations

  • Increase the detail of goods and services in the compilation of intermediate consumption.

  • Develop short term statistics on changes in inventories.

Consumer and Producer Price Indices

High Priority

  • Define more precisely the items selected for price collection.

  • Introduce procedures to identify and correct for quality changes.

  • Strengthen current validation procedures, particularly for the CPI.

  • After the results of the Business Census 2003 have become available, rebase the PPI and make it representative of all establishments in mining, manufacturing, and public utilities.

  • In cooperation with the Zanzibar authorities, start the compilation and dissemination of price statistics for the full Tanzanian territory.

Other Key Recommendations

  • Disseminate CPI and PPI methodologies.

Government Finance Statistics

High Priority

  • Disseminate grants data excluding data on transactions in kind, and disclose the latter as a memorandum item.

  • Improve the dissemination of revisions made during the course of the year, identifying revised data and indicating the reasons for revision.

  • Correct the classification of lending and repayments transactions, and present these on a net basis.

  • Disseminate comprehensive debt data and disclose the data on debt guarantees as a memorandum item.

  • Disseminate data on central government that clearly identify transactions between different subsectors of central government

  • Standardize the analytical framework in which central government data are presented to that used in the GFSM 1986.

  • Improve the classification of the source data on extrabudgetary and development expenditure.

  • Disseminate data on central government that adheres to the economic classification of all transactions, as recommended by the GFSM 1986.

  • Compile data on local government operations consistent with the international standard.

  • Extend the coverage to the general government, and disseminate data on consolidated general government, and identify transactions between the different subsectors of general government.

  • Investigate data inconsistencies with the national accounts, monetary and financial statistics, and balance of payments statistics, and implement procedures to reconcile data on an ongoing basis.

  • Compile government finance statistics for Zanzibar in a form that can be consolidated with those of the mainland.

Other Key Recommendations

  • Publish methodological notes on compilation practices adapted to user needs.

  • Develop a migration path to GFSM 2001.

Monetary and Financial Statistics

High Priority

  • Include all depository corporations in the coverage of the depository corporations survey.

  • Provide adequate information on the sectoral distribution of domestic credit, as well as on the accrued interest, in monetary statistics.

  • Revise the bank report form to collect adequate information on the sectoral distribution of assets and liabilities.

Other Key Recommendations

  • Change the valuation basis of foreign securities to market valuation.

Balance of Payments Statistics

High Priority

  • Investigate how to address deficiencies in the BoT’s legal mandate for compiling balance of payments statistics.

  • Bring the methodology for debt arrears and reserve assets into accordance with BPM5.

  • Address deficiencies in the classification/sectorization of debt transactions, and current transfer debits in particular.

  • Change the data sources for travel, investment income, and financial flows from the ITRS records or adjusted investment approvals data to the survey-based data.

  • Improve the accuracy and reliability of the data by (i) correcting the problem with the “migration” of the historical data to the new data compilation program, (ii) rigorously following the procedures for data compilation and review prior to dissemination, (iii) preparing documentation on the data compilation and review process.

  • Improve the consistency of the data by identifying and explaining the reasons for revisions in the disseminated data.

  • Improve the accessibility of the data by (i) addressing the problems with the layout of the present data tables that hinder proper interpretation by users, and (ii) disseminating on the BoT website the detailed data compiled for the standard balance of payments components.

  • Revise historical data consistent with the changes recommended above.

Other Key Recommendations

  • Continue to develop alternative data sources for perceived gaps in the ITRS data, such as road transport receipts.

  • Continue to work with the MOF on the development of comprehensive and timely data on transfers.

  • Introduce a system of regular reconciliation between the debt arrears transactions in the balance of payments statistics and the debt arrears stocks in the Debt Management Department data.


A detailed description of the GDDS may be found on the IMF’s Data Standards Bulletin Board (DSBB) on the Internet at http://www.dsbb.imf.org.


The mission team was headed by Mr. Ulhas Gunjal and included Mr. Cornelis Gorter, Ms. Marie Montanjees (all STA), Mmes. Ann Lisbet Brathaug and Sagé de Clerck (External Consultants), and Ms. Claudia Fagbemi (STA–Staff Assistant). Mr. Charles Enoch joined the mission in the second week.


Socio-demographic statistics were not covered in the assessment.


Information on data quality can be found on the IMF website at the “Data Quality Reference Site” (http://dsbb.imf.org/dqrsindex.htm).