Turkey: Report on the Observance of Standards and Codes (ROSC)—Data Module Volume III

This report examines the Observance of Standards and Codes on the Data Module for Turkey. Progress in data quality reflects the authorities’ commitment to set statistical development on a high and sustainable path, and to address remaining challenges. A cornerstone of this effort is the more robust framework established by the new Statistics Law of Turkey of November 2005. This law assigned compilation responsibilities for a broad range of official data, including macroeconomic statistics.

Abstract

This report examines the Observance of Standards and Codes on the Data Module for Turkey. Progress in data quality reflects the authorities’ commitment to set statistical development on a high and sustainable path, and to address remaining challenges. A cornerstone of this effort is the more robust framework established by the new Statistics Law of Turkey of November 2005. This law assigned compilation responsibilities for a broad range of official data, including macroeconomic statistics.

Detailed Assessment Using the Data Quality Assessment Framework (DQAF)

The following detailed information on indicators of statistical practices in the areas of the national accounts, prices, government finance, balance of payments, and monetary statistics was gathered from publicly available documents and information provided by Turkish officials. This information, which is organized along the lines of the generic DQAF (see Appendix II), was used to prepare the summary assessment of data quality elements, based on a four-part scale of observance, shown in Turkey’ s Report on the Observance of Standards and Codes (ROSC)—Data Module.

I. National Accounts

0. Prerequisites of quality

0.1 Legal and institutional environment

0.1.1 The responsibility for collecting, processing, and disseminating the statistics is clearly specified.

The existence of an official statistical agency dates back to 1926 in Turkey. A centralized statistical system was established in 1930. The former statistical agency was reorganized to become the State Institute of Statistics (SIS) under the Law on Statistics, Law No. 53, of June 1962. That Law was supplemented by Decrees No’s 219, 357 and 403 in 1984, 1989, and 1990, respectively. A revised statistical law was prepared, with a view to making Turkey’s statistical provisions fully consistent with Eurostat guidelines in support of Turkey’s candidacy for European Union (EU) membership. The new law, the Statistics Law of Turkey No. 5429, was enacted on November 10, 2005. This new law totally replaces all previous legislation relating to the production of statistics in Turkey.

The new law establishes the basic principles and standards concerning the production and organization of official statistics. It provides for the creation of the Turkish Statistical Institute (Turkstat) and specifies its duties and authorities which are to collect, assess and disseminate all kinds of statistics relating to the economic, social, and cultural activities of the country, and to ensure coordination among other institutions and organizations that are involved in the statistical process prescribed in the Official Statistics Program (Article 1).

The Official Statistics Program (OSP) establishes the framework for official statistics that should be produced on subjects required at both national and international levels. Censuses and surveys are carried out within the framework of this Program. Taking into account existing resources, respondent burden, and cost-benefit analyses, the Program specifies the statistics needed to determine and monitor the situation of the country in the fields of the economy, social issues, demography, culture, environment, science, technology and any other required areas. The OSP is a detailed rolling five year plan prepared by Turkstat with the benefit of advice from the Statistical Council (a high level advisory body that is technically part of Turkstat). The OSP needs to be approved by the Council of Ministers and its implementation is the responsibility of Turkstat. Turkstat produces an annual report on the implementation of the program (Article 3).

The technical independence of Turkstat is clearly enunciated in Article 17 of the Statistics Law of Turkey. No person or organization outside Turkstat can give instructions to the staff of Turkstat in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; and observance of statistical confidentiality.

0.1.2 Data sharing and coordination among data-producing agencies are adequate.

Turkstat has a central role in coordinating the whole statistical system in Turkey. The duties and authorities of Turkstat are proscribed in Article 18 of the Statistics Law of Turkey. Turkstat is responsible for preparing the Official Statistics Program and for monitoring its implementation. Turkstat is responsible for determining the statistical methods, definitions, classifications and standards to be used in the production of official statistics (both within Turkstat and other institutions) in line with national and international norms. Each year Turkstat issues an advance release calendar for all official statistics.

0.1.3 Individual reporters’ data are to be kept confidential and used for statistical purposes only.

Article 13 of the Statistics Law of Turkey states that individual reporters’ data should not be provided to any administrative, judiciary or military authority, or person. Under Article 53 of the Statistics Law of Turkey staff who violate these provisions are liable to imprisonment and a (nominal) fine according to Article 258 of the Turkish Penal Code no. 5237.

Special aggregation rules have been developed to ensure that indirect disclosure of individual data does not occur when aggregations of data are presented. Staff review all data prepared for dissemination for possible indirect disclosure. Access to individual data is restricted to staff who require the information in the performance of their duties.

Article 14 of the Statistics Law of Turkey allows unit record data to be provided for research purposes provided that any information that could be used to identify individual respondents has been removed from the data file.

Steps are taken to secure the premises of the data-producing agency, and computer systems are password protected to prevent unauthorized access to data, which is limited according to need.

0.1.4 Statistical reporting is ensured through legal mandate and/or measures to encourage response.

According to Articles 7 and 9 of the Statistics Law of Turkey, Turkstat is authorized to request the information it deems necessary from all public institutions and organizations, and real and legal entities. All public institutions and organizations, and real and legal entities must furnish the requested information accurately and in the format and time period determined by Turkstat. Turkstat is authorized to investigate the accuracy of the information furnished, and to request additional information and documents from those concerned. The Law also proscribes that noncompliance can result in an administrative fine (Article 54). In practice, this provision is seldom applied, as serious efforts are made to create goodwill among data providers. In the case of households, penalties are never imposed. On the other hand, businesses that fail to provide the required information after repeated requests may be, and often are, fined.

To assist respondents to the annual business surveys to correctly complete their questionnaires, commonly used commercial accounting codes for relevant items have been included for each data item in the survey questionnaire. This practice was introduced for the survey in respect of reference year 2005 and continued in subsequent surveys.

0.2 Resources

0.2.1 Staff, facilities, computing resources, and financing are commensurate with statistical programs.

Management considers that staff resources at present are commensurate with operational needs. Staff turnover is not a problem for the NAG with many staff having more than 15 years experience in national accounts.

Computer resources are adequate as all staff have their own PCs. National accounts data files are backed up both locally and in the data processing centre. However, national accounts data are not stored in a time series database.

0.2.2 Measures to ensure efficient use of resources are implemented.

Measures to ensure efficient use of resources are implemented. For example, Computer Assisted Telephone Interviewing technology is being introduced progressively in relevant areas of Turkstat. There is currently no formal staff appraisal scheme. However, performance criteria are being developed for measuring staff performance and for career planning.

0.3 Relevance

0.3.1 The relevance and practical utility of existing statistics in meeting users’ needs are monitored.

Under the Statistics Law of Turkey (Article 20) a high level advisory body, the Statistical Council, was established to provide advice on the preparation and implementation of the Official Statistics Program. Its functions also include identifying areas in which official statistics are required and providing advice on future work programs. The Council includes Undersecretaries of Ministries, the Governor of the Central Bank of the Republic of Turkey (CBRT), and representatives from the Union of Chambers and Stock Exchanges, and academia.

A web-based “User Satisfaction Survey” is now conducted every six months. The results are not published but are used by Turkstat to assess how well user needs are being met. Previously, major user surveys were conducted periodically (e.g. 1996, 2001, and 2004).

User needs and the relevance of the statistical program are also assessed through the Annual Research Symposium, the National Statistics Congress, advisory committees and other ad hoc meetings with government officials.

In May each year Turkstat organizes an Annual Research Symposium focusing on methodology issues and statistical techniques, to which employees, academics and other experts from Turkey are invited to present papers. Foreign speakers might be invited for some special sessions.

Turkstat also assists in organizing the biennial National Statistics Congress, in cooperation with the Turkish Statistical Association and the Association of Statistics Graduates. The first congress was held in 1999.

A contact person is assigned to deal with requests or questions from users, and unpublished but non-confidential information is made available on demand. Turkstat also sends participants to meetings and seminars of regional and international organizations that examine the content of statistical programs.

0.4 Other quality management

0.4.1 Processes are in place to focus on quality.

Management is sensitive to issues regarding the quality of statistics. A special unit, the Data Quality Control Unit, has been established within Turkstat to monitor the quality of statistical series. Turkstat has developed a “Strategic Plan, 2007-2011” that is available on its website. The plan includes detailed strategic goals and objectives relating to all statistics produced by Turkstat. Progress towards meeting the strategic goals and objectives is being monitored regularly.

The Strategic Plan proposes the adoption of the Total Quality Management System from 2008.

0.4.2 Processes are in place to monitor the quality of the statistical program.

Under Article 41 of the Statistics Law of Turkey a Data Quality Control Committee has been established to review all aspects of statistical work undertaken within Turkstat. It is intended that Quality Reports will be prepared for statistical surveys from 2010.

0.4.3 Processes are in place to deal with quality considerations in planning the statistical program.

Turkstat is aware of tradeoffs in the production of statistics, especially in the context of its desire to meet EU statistical requirements. The Statistical Council plays an important role in the planning process as many of its members represent organizations that are both users and producers of official statistics.

Emerging data needs are gauged through user surveys, the Annual Research Symposium, the National Statistics Congress, advisory boards and other ad hoc meetings with government officials. Emerging data needs are at present considered to be largely those to meet EU accession.

1. Assurances of integrity

1.1 Professionalism

1.1.1 Statistics are produced on an impartial basis.

The national accounts are compiled on an impartial basis and there is no external pressure on the content or the release schedule of the national accounts. The technical independence of Turkstat is clearly enunciated in Article 17 of the Statistics Law of Turkey. No person or organization outside Turkstat can give instructions to the staff of Turkstat in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; or observance of statistical confidentiality.

The professional independence of Turkstat is further enhanced by the Statistics Law of Turkey (Article 23), which provides that the President of Turkstat reports directly to the Prime Minister, is appointed to a renewable five-year term by the Council of Ministers (with a two-term limit), and may not be removed from office (except for a disabling health condition).

Professionalism is promoted and supported within the agency. Staff recruitment is through a Public Service wide competition that includes both written and oral examinations. Upon arrival at the agency, new recruits take a two-month training course during which the mission and responsibilities of Turkstat are explained. The professional development of recruits is further enhanced by giving them the opportunity to work in various departments of Turkstat before being assigned to a permanent job. Recent recruits have also benefited from the opportunity to participate in a training scheme sponsored by the Statistical Office of the European Communities (Eurostat) involving a six month posting to a central statistical office in another country. Finally, promotion to the expert level (the highest in the hierarchy below the managerial level) requires an oral and written examination and the publication of a professional paper.

An internet based training facility has been established for use by new recruits, existing staff and staff of other government departments. This training program which was introduced in 2005 is conducted each Friday throughout the year. Each employee is required to complete approximately 100 hours of technical training in his statistical area. Participation is monitored randomly and the participant is required to complete an exam on the material that is covered by the course.

Staff are also given the opportunity to take part in lectures, conferences and professional meetings.

1.1.2 Choices of sources and statistical techniques, as well as decisions about dissemination, are informed solely by statistical considerations.

The choice of data sources and statistical techniques is guided solely by considerations of a statistical nature, within the constraints of resource availability and response burden. The timing and format of dissemination of statistical data are determined by Turkstat in view of statistical considerations. An advance release calendar for the coming year is issued in September each year.

1.1.3 The appropriate statistical entity is entitled to comment on erroneous interpretation and misuse of statistics.

Under the previous Law on Statistics, the SIS had the authority to comment on misinterpretation and misuse of statistics. The new Statistics Law of Turkey does not explicitly deal with this issue. However, national accounts staff have continued to check references to national accounts data in the media, and Turkstat requests clarification or correction when warranted.

1.2 Transparency

1.2.1 The terms and conditions under which statistics are collected, processed, and disseminated are available to the public.

The Statistics Law of Turkey is publicly available and is included on the Turkstat website in both Turkish and English. All Turkstat survey forms make reference to the Statistics Law of Turkey and to data confidentiality. Turkstat clearly identifies how to obtain information about statistical products, through identifying a contact person, with their postal address, telephone number and e-mail, in the various dissemination media.

1.2.2 Internal governmental access to statistics prior to their release is publicly identified.

The statistical information compiled by Turkstat cannot be given to any person or authority before it is made available to the public (Statistics Law of Turkey, Article 12). Quarterly national accounts are released to all users through a press release, on the website and on Turkey’s National Summary Data Page, at 10.00 a.m. on the pre-announced release date.

1.2.3 Products of statistical agencies/units are clearly identified as such.

Under the Statistics Law of Turkey, any data product prepared under that law must be identified as such. The name, logo and insignia of the agency appear in all publications, press releases and on the website. Likewise, national accounts data reproduced by other institutions, such as the Central Bank of the Republic of Turkey (CBRT), the State Planning Office (SPO), or the Ministry of Finance (MOF), are always attributed to Turkstat.

1.2.4 Advance notice is given of major changes in methodology, source data, and statistical techniques.

In the past major changes to the national accounts have only been made at the time of historical revisions about once a decade. Prior to the release of the upgraded quarterly national accounts in March 2008, Turkstat provided advice on its website concerning the proposed changes to be implemented in the quarterly national accounts.

1.3 Ethical standards

1.3.1 Guidelines for staff behavior are in place and are well known to the staff.

Apart from the provisions of the Statistics Law of Turkey prohibiting, and imposing penalties for, disclosure of confidential information, staff conduct is guided by the Law on Public Servants, of which all civil servants are aware. New staff are made aware of the guidelines and the mission and traditions of Turkstat when they join the agency and during the initial two-month training period.

2. Methodological soundness

2.1 Concepts and definitions

2.1.1 The overall structure in terms of concepts and definitions follows internationally accepted standards, guidelines, or good practices.

The revised Turkish national accounts series for the period from 1998 forward are compiled according to the 1995 European System of Accounts (1995 ESA) and are consequently consistent with the 1993 System of National Accounts (1993 SNA).

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices.

The Turkish System of National Accounts is compiled on a regular basis, annually since 1968 (with rough estimates going back to 1923), and quarterly since 1987. Prior to the latest major revision of the national accounts in March 2008 the system comprised the following elements:

  • Value added and gross domestic product (GDP) by activity, at current and constant prices (1987 = 100), annual and quarterly;

  • Expenditure on GDP at current and constant prices (1987 = 100), annual and quarterly;

  • GDP by income component, annual and quarterly, released once a year;

  • Compensation of employees by activity, annual and quarterly, released once a year;

  • Rest of the world accounts on an annual basis:

  • Occasional square input-output tables (most recently 1990, and 1998); and

  • Provincial GDP for the period 1987 to 2001.

In March 2008 revised annual and quarterly estimates were released for the period from 1998 following the implementation of the 1995 ESA in the Turkish national accounts. These cover the elements above except that the revised series for GDP by income components, compensation of employees by activity and provincial GDP have not yet been released. The base year for constant price estimates was moved to 1998 and chain volume measures were introduced using the “over the year” annual chaining technique. Input-output tables (including supply and use tables) for 2002 have also been released on a 1995 ESA basis.

Income and capital accounts for the total economy and for the institutional sectors are not compiled, and neither are financial accounts. However, a project has been commenced within the National Accounts Group to develop institutional sector accounts. Also a Working Group including representatives from Turkstat, CBRT, MOF and the Treasury has been established to develop financial accounts. In addition to the regular annual rest of the world accounts, detailed rest of the world accounts have been produced and published for the period 2000 to 2005, showing data in $US on a monthly basis. Further work is also planned to compile estimates for regional GDP (NUTS level 2) on a 1995 ESA basis.

The delimitation of the constituent units of the economy is in accordance with the 1993 SNA. Territorial enclaves in the rest of the world and free trade zones are included. The production boundary is in accordance with the 1993 SNA. In particular, own-account production of goods for own final consumption, output of goods for own-account fixed capital formation, mineral exploration, computer software, entertainment literary or artistic originals, and informal sector production (indirectly, through under-coverage adjustments in the 2002 benchmark year) are included. No explicit attempt is made to include illegal activities within the production boundary.

The asset boundary is generally in agreement with the 1993 SNA. A minor departure from the 1993 SNA relates to the treatment of all expenditure on military equipment and structures, as government final consumption expenditure, except for expenditure on the construction of dwellings provided to the families of military personnel which is included in gross fixed capital formation. Historical monuments are included in gross fixed capital formation. Valuables are also included in the benchmark estimates of capital formation for 2002, but are not separately identified on a quarterly basis.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices.

Institutional units and transactions are classified according to the 1993 SNA, although there are currently no institutional sector accounts. The industry classification is based on the Statistical Classification of Economic Activities in the European Community (NACE Rev 1.1). The product classification employed for the 2002 Supply and Use Tables is the Statistical Classification of Products by Activity in the European Economic Community (CPA). Household final consumption expenditure is classified according to the Classification of Individual Consumption by Purpose (COICOP). General government aggregates are not classified by the Classification of Functions of Government (COFOG) in the national accounts release. However, the Ministry of Finance now has responsibility for producing government finance statistics (GFS) and is publishing data classifying according to COFOG for central government from 2004 and for local government from 2007.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks.

The valuation rules used for recording flows and stocks are in accordance with the 1993 SNA. Market output is valued at basic prices. Agricultural output for own-use is valued at equivalent market prices. Other non-market output (of government and non-profit institutions, for instance) is valued at cost. Only the non-deductible part of value added tax (VAT) is included in the valuation of intermediate consumption. Information on insurance and freight for merchandise imports is available. Total exports and imports are valued on an f.o.b. basis.

Foreign Trade statistics are converted to new Turkish Lira using daily exchange rates. For imports of goods the selling rate is used, while for exports of goods the buying rate is used.

2.4.2 Recording is done on an accrual basis.

Transactions are recorded in accordance with the 1993 SNA, as far as possible on an accrual basis. In general, work in progress is recorded in the period in which it is produced. In principle, work-in-progress on construction is recorded in the period in which it is produced. However, work in progress is not recorded for the rearing of cattle or the cultivation of trees and orchards. Most general government transactions are initially recorded on a cash basis. However, where possible, explicit adjustments are made to record general government transactions on an approximate accrual basis. In addition, since the government as a rule does not make large retroactive payments to its employees, and public servants are paid on a monthly basis, data for compensation of employees is recorded on an approximate accrual basis.

The Public Financial Management and Control Law (Law no. 5018) makes the Ministry of Finance responsible for the preparation of GFS on a conceptually appropriate basis consistent with the Government Finance Statistics Manual 2001 (GFSM 2001) and the 1993 SNA. As this is a fairly recent development the Ministry of Finance has not yet fully taken over all aspects of the compilation of GFS data. From 2004, data for the central government are being compiled by the Ministry of Finance in accordance with GFSM 2001, and hence should be on an accrual basis. Similarly form 2007 data for local government are being compiled by the Ministry of Finance in accordance with GFSM 2001.

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices.

Transactions between establishments within the same enterprise are recorded on a gross basis.

3. Accuracy and reliability

3.1 Source data

3.1.1 Source data are obtained from comprehensive data collection programs that take into account country-specific conditions.

With the assistance of Eurostat a major project has been undertaken over the last several years to establish a reliable business register. The register now covers all industries except agriculture, fishing and defense. Administrative records from the taxation agency (Revenue Administration) are used to update the register at the end of April each year in respect of the previous calendar year. The annual updates take account of ceased businesses in a reasonably timely fashion. The coverage of the register was significantly expanded by conducting a field enumeration in preparation for the 2002 Census of Industry and Business Establishments. The register includes information on both enterprises and establishments.

The 2000 Building Census provided a significantly enhanced benchmark for estimates of the stock of dwellings, and hence for the estimates of actual and imputed dwelling rent, which are included in household consumption expenditure. Information on the number of electricity subscribers is used to extrapolate the number of dwellings for subsequent years. Data for average rents is obtained on a monthly basis from the Consumer Price Index (CPI), classified by type of house and heating system.

For agriculture, a census is carried out every ten years, and comprehensive data on production, sales and prices are collected by product and region through the extensive network of the Ministry of Agriculture. With assistance from Eurostat a number of projects are underway to improve the quality of agricultural statistics. There are also fairly comprehensive statistics on fishing.

The annual Structural Business Statistics survey covers most industries except agriculture and fishing. A new annual survey, the Survey of Industrial Production (PRODCOM), was commenced in 2004. It provides data on the quantity and value of products produced by the mining, and manufacturing industries. The monthly survey of manufacturing industry production is an important data source for the quarterly national accounts estimates of value added for manufacturing. A new quarterly survey of service industries is in the process of being developed to provide information on revenue and employment. Although this survey was commenced in 2005 the first results will not be published until about four years data are available. A new quarterly construction industry survey was also introduced in 2006 to provide data on production, orders, employment, wages and salaries, and working hours.

Building and construction used to be estimated on the basis of demand side data. Turkstat collects information on building permits (value and area in square meters, by type of building) and occupancy permits from all municipalities with a population of over 2,000. Most other construction activities (such as road building) are carried out by the State or public corporations, for which direct information is available. However, for the recent historical revision, a commodity flow approach was adopted in order to allow for construction activity that was not being covered by the permit system. The commodity flow approach resulted in a significant adjustment to the level of private gross fixed capital formation on construction in 2002 (up 33.5%). The commodity flow approach is used for all subsequent years although the measure based on permits is still calculated as a consistency check.

Turkstat surveys rely on scientific random sampling. Automated editing procedures are applied to questionnaires. Proper imputation methods are used to handle nonresponse, and expansion factors are derived scientifically, based on the sample design. Data collected in the annual structural business survey are sufficiently detailed to derive all key national accounts aggregates, namely output, intermediate consumption, fixed capital formation and changes in inventories.

Household surveys are comprehensive, covering both urban and rural areas and all socio-economic groups. They follow proper procedures with respect to sample design, edit and imputation procedures. Expansion factors are derived scientifically, based on the sample design. The Labor Force Survey (LFS) has been conducted monthly since 2000; although, data are disseminated quarterly (previously the LFS was conducted on a quarterly basis). The sample size has been increased to 13,000 households each month. The Household Budget Survey (HBS) used to be conducted at seven-year intervals; however, from 2002 the survey is being conducted annually. Data collected include purchases of consumption goods and durable goods, production for own-consumption, and purchases of valuables.

The Public Financial Management and Control Law (Law no. 5018) makes the Ministry of Finance responsible for the preparation of GFS on a conceptually appropriate basis consistent with the Government Finance Statistics Manual 2001 (GFSM 2001) and the 1993 SNA. As this is a fairly recent development the Ministry of Finance has not yet fully taken over all aspects of the compilation of GFS data.

Comprehensive government finance statistics are available on a monthly and an annual basis for the budgetary central government. Annual information on extra-budgetary activities of the central government and on local government units is also available. Turkstat is still conducting a quarterly survey of 600 local government units, but it is expected that this collection can be discontinued in 2009 when the Ministry of Finance is able to provide all the necessary data relating to GFS in a timely fashion. Detailed data are available to measure output, intermediate consumption, gross fixed capital formation and final consumption expenditure of general government.

3.1.2 Source data reasonably approximate the definitions, scope, classifications, valuation, and time of recording required.

Generally, source data are consistent with the definitions, scope, and classifications of the national accounts, and in terms of the time of recording, reference periods, and valuation principles. Taxes on production and imports, for instance, are estimated on an assessment basis, a reasonable approximation of taxation accruals. In addition, a specific adjustment is made to value added tax (VAT) data to more closely approximate an accrual basis. The fiscal year of the State, and all enterprises, is the calendar year, which tends to limit inconsistencies in the time of recording between the quarterly and the annual estimates. Annual questionnaires cover output, capital formation, and inventories, as well as limited details for intermediate consumption. Compilers are well aware that concepts and definitions underlying source data may differ from those required by the national accounts, and make necessary adjustments. From 2004, data for the general government sector are being compiled by the Ministry of Finance in accordance with GFSM 2001, and hence should be on an accrual basis.

The coverage of economic activities in terms of value added for most industries is very good.

3.1.3 Source data are timely.

Sub-annual surveys, annual and sub-annual administrative data, and all the other major statistics used in compiling the quarterly accounts are timely. Quarterly production data are published 60 to 70 days after the reference quarter.

There is a lack of timeliness in respect of the annual surveys (particularly the annual structural business survey). As a result, many are simply not used in the compilation of the annual national accounts. In practice, therefore, the national accounts’ compilers rely on extensive, detailed and reliable monthly or quarterly data from surveys and administrative sources. On the production side, for most activities, only quarterly estimates are compiled, and annual estimates are obtained by summation. Essentially, the annual compilation of the annual accounts involves incorporating the annual data on the revenue and expenditure of general government and public corporations, as well as any revisions to quarterly source data, which are generally quite small.

The 2002 Census of Industry and Business Establishments took many years to process, thereby causing delays in producing the results from subsequent annual business surveys. Results for the 2003 and 2004 annual business surveys were published in 2008.

Table A.

GDP in Terms of Production

article image

Since the estimation of household consumption, investment in machinery and equipment and private construction on the expenditure side of GDP relies on a commodity flow approach, it is carried out largely on a quarterly basis, with annual estimates obtained by summation.

With a few exceptions (e.g., rent, energy, communications, and part of transportation), household consumption is calculated through a commodity flow approach in which the split between intermediate consumption and value added, as well as between intermediate and final consumption expenditures by households with unincorporated businesses, is itself largely based on fixed coefficients, carried forward from the 2002 supply and use tables.

Table B.

GDP in Terms of Expenditure

article image

In the case of Turkey, the State and large (especially public) corporations account for a very significant share of output for some activities, and the coverage of these entities is excellent annually, and very good on a quarterly basis, especially for public sector establishments. The commodity flow approach probably yields reasonable results for private machinery and equipment, and construction. However, the use of a commodity flow approach without information on inventory changes, and the limited possibility to do crosschecks, is problematic for the measurement of household consumption, especially in the absence of independently derived annual estimates. Consequently, it would be desirable to develop direct estimates for household consumption expenditure on goods. Some data for this purpose are available from the new quarterly survey of service industries, the annual structural service industries surveys and the HBS. However, it may be necessary to develop other sources to obtain sufficient commodity detail. NAG hopes to be able to use data compiled from the bar coding system used by retailers to obtain very detailed information on retail sales of goods to households.

3.2 Assessment of source data

3.2.1 Source data—including censuses, sample surveys, and administrative records—are routinely assessed, e.g., for coverage, sample error, response error, and nonsampling error; the results of the assessments are monitored and made available to guide statistical processes.

Information is available about sampling errors where applicable, nonsampling errors (coverage, misclassification, measurement problems, nonresponse) are regularly reviewed, and adjustments and corrections made where warranted. High-value transactions are usually confirmed with respondents by post or by phone. Feedback is provided by the NAG to source data areas when there are concerns about the quality of particular statistics. The accuracy of administrative data and other secondary sources is routinely assessed, as is the accuracy of budgetary data, international trade, price statistics, and other sources used in compiling the accounts. Source data are also scrutinized for consistency with related data, and with national accounting concepts, definitions and classifications.

3.3 Statistical techniques

3.3.1 Data compilation employs sound statistical techniques to deal with data sources.

General procedures in survey divisions are sound. Computerized edits are performed to identify errors in the source data, and imputation techniques are used to account for nonresponse.

Production approach procedures

Output estimates are compiled at a sufficient level of industrial detail in the benchmark year (2002). In other years, estimates are prepared for about 30 economic activities, at the one- or two-digit level of the NACE classification. For manufacturing, the commodity flow system, which has been significantly enhanced during the historical revision, is used to adjust the ratio of intermediate consumption to output for the non-census years.

Estimates of intermediate consumption are compiled for the benchmark year at a sufficiently detailed level. However, fixed ratios (corresponding to the overall ratio of intermediate consumption to output at the one- or two-digit level) in 2002 have been applied for a number industries.

Output of owner-occupied dwellings is valued correctly as the estimated rent that tenants would pay for similar accommodation. The benchmark estimate is correctly extrapolated as the product of the stock of dwellings times the average rent.

Growing crops and rearing of livestock are recorded when harvested or slaughtered. Data on output and intermediate consumption in mining and manufacturing are adjusted for holding gains and losses on inventories.

For construction, the commodity flow approach at constant prices is used to estimate output using the input output ratio in 2002. Construction materials used in dwelling repairs are first deducted, as is cement that is assumed to be used in the production of concrete.

The perpetual inventory method is the conceptual basis for estimating consumption of fixed capital for all types of assets.

Government expenditure, except subsidies, has generally been available on a cash basis. Where possible, adjustments are made to approximate an accrual basis. Taxes on production and imports are initially valued on an assessment basis, using estimates reported by the MOF. The assessment basis is intermediate between cash and accrual valuation. An explicit adjustment is made to put VAT data on an accrual basis. Subsidies are valued on an accrual basis. From 2004, data for the general government sector are being compiled by the Ministry of Finance in accordance with GFSM 2001, and hence should be on an accrual basis.

Price indices are compiled at a detailed level and largely meet national accounts requirements with respect to concepts, coverage, and reference periods. Double-deflation is used for agriculture and manufacturing. Output volume indicators are used to extrapolate value added in mining, electricity, construction, wholesale and retail trade, hotels and restaurants, part of transport and storage, communications, and financial services. Single deflation of value added is employed for communications, and part of transport. Labor input measures from the LFS are used for business services, personal services, education, health and public administration and defense. Volume measures of taxes/subsidies on products are estimated by applying base-year tax rates to the volume of transactions, subject to a specific tax/subsidy. Output volume of trade margins is estimated by applying the base-year margin rates to the corresponding volume of sales. The base year for constant price estimates is 1998 and chain volume measures are compiled using the “over the year” annual chaining technique.

Expenditure approach procedures

Among GDP expenditure components, government final consumption, public gross fixed capital formation, external trade and a portion of household consumption (rents, nonprofit institutions, most spending on energy, transportation and communications) are estimated independently. The rest of household consumption and private investment in machinery and equipment and construction are derived using the commodity flow approach.

Household final consumption expenditure, external trade and investment in machinery and equipment are compiled according to the Classification of Broad Economic Categories, United Nations, 1989 (BEC), itself an aggregation of the Standard International Trade Classification, Revision 3 (SITC, Rev. 3). The BEC comprises some 20 categories, covering intermediate consumption, final consumption and capital goods. Government final consumption expenditure is now compiled according to COFOG. Gross fixed capital formation is compiled by type of asset.

With respect to volume measurement, specific price indices/deflators are used to deflate expenditure on GDP components. The deflation is done at a detailed level for household final consumption expenditure using COICOP categories, and using the Purchasing Power Parity (PPP) classification for gross fixed capital formation. Deflation of household final consumption expenditure is done indirectly by deflating independently the results of the commodity flow compilation, namely, imports, exports, trade margins, and taxes. A more generally accepted and straightforward approach would be to simply deflate the current price household consumption expenditure using the detailed components of the CPI, which would require some reworking of Turkstat’s compilation system for this component of GDP.

Exports and imports of goods are deflated using unit value indices calculated by the International Trade section of Turkstat. In some instances, volume measures are extrapolated and deflators are calculated implicitly. The household consumption implicit deflator is broadly consistent with the CPI. Government final consumption expenditure is obtained by deflating cost components of output for final use (broken down into employee compensation and “goods and services”).

The base year for constant price estimates is 1998 and chain volume measures are compiled using the “over the year” annual chaining technique.

None of the quarterly national accounts series are seasonally adjusted.

3.3.2 Other statistical procedures (e.g., data adjustments and transformations, and statistical analysis) employ sound statistical techniques.

Adjustments are made to source data when their underlying concepts are not consistent with national accounting principles. In addition, a systematic effort is made in the benchmark year to achieve an exhaustive measure of GDP, by including value added from informal and hidden legal activities. For the historical revision of the national accounts released in March 2008 significant improvements were made to the exhaustiveness of the measure of GDP in the benchmark year 2002. For a number of industries, including manufacturing, hotels and restaurants, other business services, recreation and cultural services and other services, an undercoverage adjustment was obtained by comparing employment in the Labor Force Survey (LFS) with employment from the 2002 Census of Industry and Business Establishments. This method is referred to as the “labor input method”. It is one of the methods recommended in Measuring the Non-Observed Economy – A Handbook. However, the undercoverage adjustment has been moved forward from 2002 using the level of activity in each industry rather than being calculated using LFS data for subsequent years.

For construction, the previously compiled estimates were compared with those derived from a commodity flow approach for construction materials. This comparison indicated a significant understatement in construction output as derived from the traditional sources. Consequently, the commodity flow method was used to estimate construction output for all years covered by the historical revision.

It is believed that small businesses are minimizing their reported income for taxation purposes by understating their income and overstating their operating expenses. Investigations, including sending surveys to accountants, are being undertaken in an effort to estimate the level of the resulting understatement of value added. However, at this stage no explicit adjustments are being made to value added on account of this type of misreporting.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Intermediate results are validated against other information, where applicable.

Data compiled from the main data sources are validated against other independent data to the extent possible.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated.

Estimates of household final consumption expenditure are compared with corresponding estimates derived from the annual HBS. This exercise has revealed a puzzling difference between the national accounts estimates and the HBS results for most categories of household final consumption expenditure, with the national accounts estimates being much higher. The extent of these differences was increased following the recent historical revision of the national accounts. Although both areas of Turkstat are conscious of this issue, a satisfactory explanation for the differences has not yet been found. Part of the explanation is believed to be due to a cultural behavior in Turkey not to reveal the full extent of ones income and expenditure, even to other family members.

3.4.3 Statistical discrepancies and other potential indicators of problems in statistical outputs are investigated.

The statistical discrepancy between GDP by activity and GDP by expenditure component is investigated. In this regard, however, the supply and use framework is of limited usefulness, as the estimates of supply and demand are not entirely independent. This arises because of the extensive use of the commodity flow approach to derive some major expenditure components of GDP. The discrepancy between GDP in terms of production and GDP in terms of expenditure is allocated entirely to the expenditure side, on the assumption that the production estimate is more reliable. In practice, this is done in the quarterly national accounts by deriving changes in inventories as a balancing item at both current and constant prices (changes in inventories therefore also includes the statistical discrepancy).

The supply and use framework was used to compile benchmark estimates of GDP for the year 2002. Work is in progress on the compilation of balanced SUT for 2003 and 2004. The intention is that SUT will be compiled on an annual basis.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used internally to inform statistical processes (see also 4.3.3).

Except for those occasions when major historical revisions are incorporated into the accounts, revisions are generally quite small, due to the current revision cycle of only one year, which does not provide time to incorporate results of major annual surveys. Systematic revision studies have not been undertaken.

4. Serviceability

4.1 Periodicity and timeliness

4.1.1 Periodicity follows dissemination standards.

Turkstat meets the Special Data Dissemination Standard (SDDS) with respect to the periodicity of the GDP estimates. Estimates of GDP are published quarterly at current and constant prices.

4.1.2 Timeliness follows dissemination standards.

Turkstat meets or exceeds the SDDS with respect to timeliness. Estimates for the first and fourth quarters are released 90 days after the reference quarter, while those for the second and third quarters are released 70 days after the reference quarter.

4.2 Consistency

4.2.1 Statistics are consistent within the dataset.

The estimates of GDP by activity and by expenditure component are internally consistent from a conceptual standpoint. The production measure is regarded as the official measure of GDP. Following the historical revision (covering the period from 1998), an independent expenditure measure of GDP is not produced in the quarterly national accounts, because no estimate is derived for changes in inventories. Instead changes in inventories and the implicit statistical discrepancy are combined as a single item. On an annual basis the same applies except for those years for which balanced supply and use tables have been compiled (to date only 2002).

GDP estimates at current prices, those at constant prices and the implicit deflators are consistent within the “value = volume × price” framework.

Quarterly and annual GDP estimates are consistent with one another from a conceptual standpoint. They are also quite close to one another prior to any proration to force equality between them, as annual figures are obtained by summation to a large extent.

4.2.2 Statistics are consistent or reconcilable over a reasonable period of time.

An historical revision of the Turkish national accounts was released in March 2008. The revision involved major improvements to the quality of the estimates, as well as the implementation of the 1995 ESA. The base year for constant price estimates was moved from 1987 to 1998. However, the revision has resulted in a major discontinuity in the national accounts time series. The revised series cover the period from 1998 forward.

The previously published series are still available on the Turkstat website up to 2007, third quarter. The annual series commence in 1923 while the quarterly series start in 1987. Generally, when changes in source data, methodology or statistical techniques are introduced, historical series are reconstructed as far back as reasonably possible. While this is also the intention on this occasion, there are, as yet, no specific plans for how to reconstruct the series for years prior to 1998.

4.2.3 Statistics are consistent or reconcilable with those obtained through other data sources and/or statistical frameworks.

The external trade data used to compile the expenditure measure of GDP are totally consistent with those shown in the balance of payments. This consistency is confirmed each quarter as part of the regular editing process.

4.3 Revision policy and practice

4.3.1 Revisions follow a regular and transparent schedule.

With the release of the 1998-2006 historical revision, Turkstat has announced a new revisions policy. In future, it is proposed to publish the preliminary quarterly GDP estimates approximately 70-90 days after the reference quarter, to revise the quarters of the current year, while data for the previous years will remain frozen until the next annual revision. Revisions will be conducted once a year for the most recent three or four years based on the availability of annual supply and use tables. From time to time, Turkstat will conduct historical revisions to correspond to major updates of international standards and/or the availability of new data sources. It will take some time before this new policy is fully operational.

The current revision policy, which has been in effect since the early 1990s, is familiar to users and has been quite stable from year to year. Revisions are usually small because data from independent annual sources are generally not incorporated into the accounts. The estimates of the first three quarters of the current year are revised at the time of the first estimation for the fourth quarter, and the estimates for all quarters of that same year are revised once more six months later. After that, as a rule, estimates for earlier years and quarters are not revised until a new benchmark year is estimated, every five to ten years. New source data tend to be incorporated at the time of major historical revisions to avoid breaks in the series.

4.3.2 Preliminary and/or revised data are clearly identified.

Preliminary and revised data are identified. Users are made aware that data are preliminary and subject to revision. The revised data are disseminated with the same level of detail as those previously published.

4.3.3 Studies and analyses of revisions are made public (see also 3.5.1).

Revisions to national accounts series are mentioned in the press release and on the website.

5. Accessibility

5.1 Data accessibility

5.1.1 Statistics are presented in a way that facilitates proper interpretation and meaningful comparisons (layout and clarity of text, tables, and charts).

Estimates, growth rates and shares are displayed in the press release by (i) activity at the one-digit level; and (ii) by major expenditure category, with additional detail for household consumption expenditure by major COICOP category, along with a short analytical commentary. It would be beneficial to users if some additional information could be shown in the release, including:

  • Income components of GDP;

  • Compensation of employees by industry;

  • Gross national income;

  • The sequence of accounts for the total economy;

  • More detailed components of gross fixed capital formation; and

  • Implicit price deflators.

Interpretation of the quarterly national accounts would be enhanced if all series were seasonally adjusted.

5.1.2 Dissemination media and format are adequate.

Data are first released via the press release and through electronic means. Longer time series appear in several publications: the Monthly Bulletin of Statistics of Turkstat, the Main Economic Indicators of the State Planning Organization (SPO) (monthly) and the annual national accounts publication (last issue 2000), as well as on the Turkstat website and the Central Bank of Turkey and SPO websites. More detailed series are available on demand and a designated contact person in the national accounts handles all enquiries.

5.1.3 Statistics are released on a preannounced schedule.

The Advance Release Calendar for the coming year is posted on the Turkstat website in September. A notice to this effect is published in the Monthly Bulletin of Statistics. The release calendar is also given in each national accounts press release.

5.1.4 Statistics are made available to all users at the same time.

The statistical series are released simultaneously to all interested users on the date and time specified in the advance release calendar. The press is not briefed in advance.

5.1.5 Statistics not routinely disseminated are made available upon request.

Non-published but non-confidential detailed data are made available to users upon request. However, users are no longer advised of this service in the quarterly national accounts release.

5.2 Metadata accessibility

5.2.1 Documentation on concepts, scope, classifications, basis of recording, data sources, and statistical techniques is available, and differences from internationally accepted standards, guidelines, or good practices are annotated.

Documentation on national accounts concepts and methodology was published in Gross National Product; Concepts, Sources, and Methods 1994. The metadata page on national accounts and other SDDS categories, which contains a summary description of methodology, was posted on the IMF’s Dissemination Standards Bulletin Board in September 1996. These summary methodologies and other related metadata are regularly reviewed and updated.

The NAG is in the process of producing the GNI Inventory which is required by Eurostat. This document will provide current information at a very detailed level on the concepts, sources and methods used to compile Turkey’s national accounts.

5.2.2 Levels of detail are adapted to the needs of the intended audience.

Metadata are available at different levels of detail.

5.3 Assistance to users

5.3.1 Contact points for each subject field are publicized.

All statistical releases and publications indicate the mailing address, telephone and fax numbers, and e-mail address of the specific divisions of the National Accounts Department that are responsible for the compilation of the statistics, as well as contact information for ordering the hardcopy publication.

5.3.2 Catalogs of publications, documents, and other services, including information on any charges, are widely available.

The list of Turkstat publications is updated daily on the Turkstat website. The price of various statistical products is clearly disclosed and assistance is provided in placing orders.

RECOMMENDATIONS

High priority

  • Integrate a regular annual estimation process into the existing—mainly quarterly based—extrapolation system for compiling the national accounts. To this end, construct annual supply and use tables in current and constant prices using data from the annual structural business survey and other available sources.

  • Compile a more independent measure of household final consumption expenditure, both annually and quarterly, using data from the annual business survey for retail and service industries, the new quarterly service industry collection, and other sources.

  • Collect information on changes in inventories (at least for the major stock holding industries–mining, manufacturing and wholesale and retail trade) on a quarterly basis.

  • Produce results from the annual structural business survey on a more timely basis.

  • Compile and publish the sequence of accounts for the total economy on an annual—and preferably quarterly—basis.

  • Seasonally adjust all published quarterly national accounts estimates.

Other

  • Reconstruct the GDP series on the new 1995 ESA basis as far back as possible.

  • Publish a series for gross national income in the next quarterly national accounts release (necessary data are available from the BOP), ensuring that migrant transfers are not taken into account.

  • Publish estimates for the income components of GDP and compensation of employees by industry as part of the quarterly national accounts release.

  • Publish more components of gross fixed capital formation (such as software, mineral exploration expenditure) in the quarterly national accounts.

  • Calculate explicit adjustments for the non-observed economy on an annual basis, rather than extrapolating the 2002 benchmark estimates.

  • Develop a methodology for adjusting value-added for the understatement of income and overstatement of expenses in data collected from businesses.

  • Develop and publish institutional sector accounts.

  • When available, use export and import price indices for goods for deflation purposes instead of unit value indices.

  • Conduct and publish periodic analyses of revisions to inform compilers and users of the national accounts.

  • Implement the new revisions policy announced in the document “Implementation of the ESA-95 in Turkey and the 1998-2006 Historical Revision” (included with the quarterly national accounts release on the Turkstat website).

  • Publish additional analytical series in the quarterly national accounts, including GDP per capita, GDP per person employed, and implicit price deflators.

  • Make the quarterly national accounts publication more user-friendly by: (i) publishing series in millions rather than thousands of YTL; (ii) avoiding percentage changes for “changes in inventories plus statistical discrepancy,” as these are not meaningful; and (iii) publishing shares of GDP in current prices only.

  • Advise users of the availability of unpublished national accounts series through the press release and the website.

  • Provide contact details for the National Accounts Group on the national accounts press release and on the website (in addition to the general contact detail already shown).

Table 1.

Turkey: Data Quality Assessment Framework (July 2003): Summary of Results for National Accounts

(Compiling Agency: Turkish Statistical Institute)

article image
article image

II. Consumer Price Index

0. Prerequisites of quality

0.1 Legal and institutional environment

0.1.1 The responsibility for collecting, processing, and disseminating the statistics is clearly specified.

The existence of an official statistical agency dates back to 1926 in Turkey. A centralized statistical system was established in 1930. The former statistical agency was reorganized to become the State Institute of Statistics (SIS) under the Law on Statistics, Law No. 53, of June 1962. That Law was supplemented by Decrees No’s 219, 357 and 403 in 1984, 1989, and 1990, respectively. A revised statistical law was prepared, with a view to making Turkey’s statistical provisions fully consistent with Eurostat guidelines in support of Turkey’s candidacy for European Union (EU) membership. The new law, the Statistics Law of Turkey No. 5429, was enacted on November 10, 2005. This new law totally replaces all previous legislation relating to the production of statistics in Turkey.

The new law establishes the basic principles and standards concerning the production and organization of official statistics. It provides for the creation of the Turkish Statistical Institute (TURKSTAT) and specifies its duties and authorities which are to collect, assess and disseminate all kinds of statistics relating to the economic, social, and cultural activities of the country, and to ensure coordination among other institutions and organizations that are involved in the statistical process prescribed in the Official Statistics Program (Article 1).

The Official Statistical Program (OSP) establishes the framework for official statistics that should be produced on subjects required at the national and international levels. Censuses and surveys are carried out within the framework of this Program. Taking into account existing resources, respondent burden, and cost-benefit analyses, the Program specifies the statistics needed to determine and monitor the situation of the country in the fields of the economy, social issues, demography, culture, environment, science, technology and any other required areas. The OSP is a detailed five year plan prepared by TURKSTAT with the benefit of advice from the Statistical Council (a high level advisory body that is technically part of TURKSTAT). The OSP needs to be approved by the Council of Ministers and its implementation is the responsibility of TURKSTAT. TURKSTAT produces an annual report on the implementation of the program (Article 3).

The technical independence of TURKSTAT is clearly enunciated in Article 17 of the Turkish Statistical Law. No person or organization outside TURKSTAT can give instructions to the staff of TURKSTAT in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; and observance of statistical confidentiality.

0.1.2 Data sharing and coordination among data-producing agencies are adequate.

TURKSTAT has a central role in coordinating the whole statistical system in Turkey. The duties and authorities of TURKSTAT are proscribed in Article 18 of the Statistics Law of Turkey. TURKSTAT is responsible for preparing the Official Statistical Program and for monitoring its implementation. TURKSTAT is responsible for determining the statistical methods, definitions, classifications and standards to be used in the production of official statistics (both within TURKSTAT and other institutions) in line with national and international norms. Each year TURKSTAT issues an advance release calendar for all official statistics.

0.1.3 Individual reporters’ data are to be kept confidential and used for statistical purposes only.

Article 13 of the Statistics Law of Turkey states that individual reporters’ data should not be provided to any administrative, judiciary or military authority, or person. Under Article 53 of the Statistics Law of Turkey staff who violate these provisions are liable to imprisonment and a (nominal) fine according to Article 258 of the Turkish Penal Code no. 5237.

Special aggregation rules have been developed to ensure that indirect disclosure of individual data does not occur when aggregations of data are presented. Staff review all data prepared for dissemination for possible indirect disclosure. Access to individual data is restricted to staff who require the information in the performance of their duties.

Article 14 of the Statistics Law of Turkey allows unit record data to be provided for research purposes provided that any information that could be used to identify individual respondents has been removed from the data file.

Steps are taken to secure the premises of the data-producing agency, and computer systems are password protected to prevent unauthorized access to data which is limited according to need.

0.1.4 Statistical reporting is ensured through legal mandate and/or measures to encourage response.

According to Articles 7 and 9 of the Turkish Statistical Law, TURKSTAT is authorized to request the information it deems necessary from all public institutions and organizations, and real and legal entities. All public institutions and organizations, and real and legal entities must furnish the requested information accurately and in the format and time period determined by TURKSTAT. TURKSTAT is authorized to investigate the accuracy of the information furnished, and to request additional information and documents from those concerned. The Law also prescribes that noncompliance can result in an administrative fine (Article 54). In practice, this provision is seldom applied, as serious efforts are made to create goodwill among data providers. In the case of households, penalties are never imposed. On the other hand, businesses that fail to provide the required information after repeated requests may be, and often are, fined.

To assist respondents to the annual business surveys to correctly complete their questionnaires, commonly used commercial accounting codes for relevant items have been included for each data item in the survey questionnaire. This practice was introduced for the survey in respect of reference year 2005 and continued in subsequent surveys.

0.2 Resources

0.2.1 Staff, facilities, computing resources, and financing are commensurate with statistical programs.

The CPI Team responsible for overseeing the compilation of the CPI is part of the Price Statistics Group within the Department of National Accounts and Economic Indicators. The Price Statistics Groups consists of two teams, the CPI Team and the PPI Team. The CPI Team is composed of one supervisor and 13 other professionals who are tasked with the review of price data collected by the 26 TURKSTAT regional offices. They are also responsible for review and analysis of the price indices compiled for all of the 26 regions and at the national level.

There are approximately 200 TURKSTAT employees in the regional offices involved in the collection and processing of price data for the CPI, the PPI, the Construction Cost Index and the International Comparison Program (ICP) from which purchasing power parities are compiled. Although the regional directors, supervisors and some other regional staff members are permanent staff, a large number the regional employees are contract employees with one-year contracts. The authorities indicate that adequate funding is available for the price statistics programs.

For permanent staff, turnover is not a problem especially for the CPI Team, where most of the staff members have more than ten years experience and no staff member has less than two years experience. All permanent staff and most contract staff hold college degrees both in Ankara and in the regional offices.

Information technology is an area that has experienced significant improvement in recent years. All staff members have individual computers that are adequate for work requirements. A state of the art Oracle database is used for data capture and index processing. Price data are entered into the database in the regional office via a direct internet link to Ankara. The compiled price series are also stored in the database. Backup and security measures are in place for all basic data and compiled indices. Excellent facilities are available for teleconferencing and internet-based training.

0.2.2 Measures to ensure efficient use of resources are implemented.

The activities of the Department of National Accounts and Economic Indicators, and their budgetary implications, are reviewed each month against the targets of the TURKSTAT work plan. After consultations with Group Leaders, the Department Head sends a report to management each month, which tracks performance of staff against a critical path for each task within the groups as well as the monthly use of budgetary allocations. The Finance Department of the TURKSTAT in turn regularly informs the departments regarding budget developments and possible additional allocations from unspent resources.

Programs are in place to minimize processing errors and improve the monitoring of price data for extreme values due to data collection and data entry errors. The data processing system allow for the comparison of price and index data across regions and over time. The time spent by price collectors to obtain the monthly price data is measured and compared to standards set through surveys by experts in the field. The price collection effort is subject to random audits to ensure that the highest standards of quality are being met.

0.3 Relevance

0.3.1 The relevance and practical utility of existing statistics in meeting users’ needs are monitored.

Under the Statistics Law of Turkey (Article 20) a high level advisory body, the Statistical Council, was established to provide advice on the preparation and implementation of the Official Statistics Program. Its functions also include identifying areas in which official statistics are required and providing advice on future work programs. The Council includes Undersecretaries of Ministries, the Governor of the Central Bank of Turkey, and representatives from the Union of Chambers and Stock Exchanges, and academia.

A web-based “User Satisfaction Survey” is now conducted every six months. The results are not published but are used by TURKSTAT to assess how well user needs are being met. Previously, major user surveys were conducted periodically (e.g., 1996, 2001, and 2004).

User needs and the relevance of the statistical program is also assessed through the Annual Research Symposium, the National Statistics Congress, advisory committees and other ad hoc meetings with government officials.

In May each year TURKSTAT organizes an Annual Research Symposium focusing on methodology issues and statistical techniques, to which employees, academics and other experts from Turkey are invited to present papers. Foreign speakers might be invited for some special sessions.

TURKSTAT also assists in organizing the biennial National Statistics Congress, in cooperation with the Turkish Statistical Association and the Association of Statistics Graduates. The first congress was held in 1999.

A contact person is assigned to deal with requests or questions from users, and unpublished, but non-confidential information, is made available on demand. TURKSTAT also sends participants to meetings and seminars of regional and international organizations that examine the content of statistical programs.

0.4 Other quality management

0.4.1 Processes are in place to focus on quality.

Management is sensitive to issues regarding the quality of statistics. A special unit, the Data Quality Control Unit, has been established within TURKSTAT to monitor the quality of statistical series. TURKSTAT has developed a “Strategic Plan, 2007-2011” that is available on its website. The plan includes detailed strategic goals and objectives relating to all statistics produced by TURKSTAT. Progress towards meeting the strategic goals and objectives is being monitored.

The Strategic Plan proposes the adoption of the Total Quality Management System from 2008.

0.4.2 Processes are in place to monitor the quality of the statistical program.

Under Article 41 of the Statistics Law of Turkey a Data Quality Control Committee has been established to review all aspects of statistical work undertaken within TURKSTAT. It is intended that Quality Reports will be prepared for statistical surveys from 2010.

0.4.3 Processes are in place to deal with quality considerations in planning the statistical program.

TURKSTAT is aware of tradeoffs in the production of statistics, especially in the context of its desire to meet EU statistical requirements.

Emerging data needs are gauged through user surveys, the Annual Research Symposium, the National Statistics Congress, advisory boards and other ad hoc meetings with government officials. Emerging data needs are at present considered to be largely those to meet EU accession.

1. Assurances of integrity

1.1 Professionalism

1.1.1 Statistics are produced on an impartial basis.

The CPI is compiled on an impartial basis and there is no external pressure on the content or the release schedule of the CPI. The technical independence of TURKSTAT is clearly enunciated in Article 17 of the Turkish Statistical Law. No person or organization outside TURKSTAT can give instructions to the staff of TURKSTAT in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; and observance of statistical confidentiality.

Professionalism is promoted and supported within the agency. Staff recruitment is through a Public Service wide competition that includes both written and oral examinations. Upon arrival at the agency, new recruits take a one-month training course during which the mission and responsibilities of TURKSTAT are explained. The professional development of recruits is further enhanced by giving them the opportunity to work in various departments of TURKSTAT before being assigned to a permanent job. Recent recruits have also benefited from the opportunity to participate in a Eurostat training scheme involving a six month posting to a central statistical office in another country. Finally, promotion to the expert level (the highest in the hierarchy below the managerial level) requires an oral and written examination and publication of a professional paper.

An internet based training facility has been established for use by new recruits, existing staff and staff of other government departments. This training program which was introduced in 2005 is conducted each Friday throughout the year. Each employee is required to complete approximately 100 hours of technical training in his statistical area. Participation is monitored randomly and the participant is required to complete an exam over the material that is covered by the course.

Every December two or three staff members from the TURKSTAT regional offices are brought into the central office for face-to-face technical training. These training sessions are conducted using teleconferencing with all appropriate regional staff. Upon return to the regional offices, the participants conduct face-to-face training as needed with the regional staff.

Staff members are also given the opportunity to take part in lectures, conferences and professional meetings.

1.1.2 Choices of sources and statistical techniques, as well as decisions about dissemination, are informed solely by statistical considerations.

The choice of data sources and statistical techniques is guided solely by considerations of a statistical nature, within the constraints of resource availability and response burden. The Price Statistics Group always refers to the latest internationally accepted methods for the choice of source data and statistical techniques. The implementation and adoption of these methods in Turkey is thoroughly discussed among staff, and final decisions are made public in advance of implementation. The timing and format of dissemination of statistical data are determined by TURKSTAT in view of statistical considerations. An advance release calendar for the coming year is issued in September each year.

1.1.3 The appropriate statistical entity is entitled to comment on erroneous interpretation and misuse of statistics.

Under the previous Law on Statistics, the SIS had the authority to comment on misinterpretation and misuse of statistics. The new Statistics Law of Turkey does not explicitly deal with this issue. However, prices staff have continued to check references to price index data in the media, and TURKSTAT requests clarification or correction when warranted. The practice of the Price Index Group is to draft a letter to the agency involved correcting the errors made. The letter is signed by the President of TURKSTAT, and the agency, usually a newspaper, is asked to publish the letter. Also, to prevent misinterpretation, the President regularly invites journalists for briefings on new developments and data problems.

1.2 Transparency

1.2.1 The terms and conditions under which statistics are collected, processed, and disseminated are available to the public.

The Statistics Law of Turkey is publicly available and is included on the TURKSTAT website in both Turkish and English. All TURKSTAT survey forms make reference to the Statistics Law of Turkey and to data confidentiality. TURKSTAT clearly identifies how to obtain information about statistical products, through identifying a contact person, with their postal address, telephone number and e-mail, in the various dissemination media.

1.2.2 Internal governmental access to statistics prior to their release is publicly identified.

There is no outside access to the CPI prior to its release.

1.2.3 Products of statistical agencies/units are clearly identified as such.

Under the Statistics Law of Turkey, any data product prepared under that law must be identified as such. The name, logo and insignia of TURKSTAT appear in all publications, press releases and on the TURKSTAT website. Likewise, CPI data reproduced by other institutions are always attributed to TURKSTAT.

1.2.4 Advance notice is given of major changes in methodology, source data, and statistical techniques.

Major changes in the CPI compilation system are discussed in advance with important user groups prior to the release of the index series that incorporate these changes.

1.3 Ethical standards

1.3.1 Guidelines for staff behavior are in place and are well known to the staff.

Apart from the provisions of the Statistics Act prohibiting, and imposing penalties for, disclosure of confidential information, staff conduct is guided by the Law on Public Servants, of which all civil servants are aware. New staff members are made aware of the guidelines and the mission and traditions of TURKSTAT when they join the agency and during the initial one-month training period.

2. Methodological soundness

2.1 Concepts and definitions

2.1.1 The overall structure in terms of concepts and definitions follows internationally accepted standards, guidelines, or good practices.

The determination and specification of CPI market basket items and the estimation of corresponding weights are based on concepts and definitions of household expenditures relating to consumption that are consistent with the 1993 SNA and the new CPI Manual.

Expenditure data on both services and commodities are tabulated in sufficient detail to permit analysis at all levels of detail for the five-digit Classification of Individual Consumption by Purpose Adapted to the Needs of Household Budget Surveys (CIOCOP-HBS).

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices.

The CPI covers all private households residing in the country irrespective of income, nationality, and social or residential status, including persons living in institutional households, such as students living in dormitories and military families.

Only monetary expenditures relating to household consumption are included in the CPI market basket weights. Hence, consumption of goods produced at home, such as home grown food, and housing services produced by homeowners (imputed rent) are excluded from the CPI market basket. Gifts are included for the purchasing households but not for the consuming households.

Up until the 2005 CPI revision, the Turkish CPI for urban households had a rental equivalence component, and the CPI for rural households had a component for own produced food. Starting in 2005 the two series were combined into one national CPI, and both rental equivalence and own produced food were excluded from the index on the basis that the new CPI would reflect only monetary expenditures related to consumption.

It should be noted that although own-produced goods and services, and imputed rent are excluded from the CPI market basket, estimates of the value of consumption for these items are available from the HBS. In addition, there is no alternative component to cover owner-occupied housing. The lack of a component to represent housing costs for homeowners in the CPI is a significant deficiency in the index, especially given that owner-occupants are estimated to constitute 70 percent of the households in Turkey. When rental equivalence (with a weight based on imputed rent for homeowners) is not used in the CPI, it is recommended that an alternative measure of the cost of housing for homeowners be used. One alternative is the measurement of the net acquisition of residential property supplemented by expenditures on additions and alterations. This would be the market basket weight which would be moved by a house price index.

In spite of the fact that the Eurostat Harmonized Index of Consumer Prices (HICP) does not have a component for owner-occupied housing due to a lack of being able to come to agreement as to how it should be measured, this does not justify the exclusion of this component from a national CPI. One simple alternative for Turkey would be the compilation of a separate CPI that includes rental equivalence for homeowners. This index could be published along with a number of separate indices that are compiled and disseminated to measure core inflation. The advantage of this approach is that the data for compiling this index are already available. The market basket weight would be derived from HBS tabulations on imputed rent, and this weight would be moved in the new separate index by the rent index from the current CPI.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices.

All classification and coding for expenditure data from the Household Budget Survey (HBS) and for the CPI market basket, monthly price survey, and index dissemination are done using a ten-digit national classification system based on the five-digit COICOP-HBS.

2.4.1 Market prices are used to value flows and stocks.

Both market basket weights and monthly prices used in the compilation of the CPI are valued at market prices, including VAT and applicable discounts. Detailed product characteristic data including terms of transactions are included as part of the specifications for monthly price collection.

2.4.2 Recording is done on an accrual basis.

Both the expenditure data from the HBS and the price data for the monthly price survey are recorded on an accrual basis.

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices.

The internationally recommended procedure for estimating the CPI market basket weights for durable goods, especially large durable goods such as vehicles, is to base the weights on net purchases (purchases net of sales) for these goods. For the current CPI, the market basket weights for durable goods, including vehicles, were estimated using the total value of purchases only. For most durable goods this is adequate, since sales of smaller durable goods tend to be insignificant. However, for vehicles, this procedure is not adequate since it is typical to have significant sales as well as purchases of vehicles by consumers.

The CPI Team has indicated that since the sales data for durable goods are available from the continuous HBS, the market basket weights for 2009 will be estimated using net purchases of durable goods.

3. Accuracy and reliability

3.1 Source data

3.1.1 Source data are obtained from comprehensive data collection programs that take into account country-specific conditions.

Beginning in January 2005 TURKSTAT began compilation and dissemination of a new CPI based on the 2003 Household Budget Survey (HBS) that covers all types of households in the economic territory of the country irrespective of income, nationality, and social or residential status.1 The new index replaced an urban CPI and a rural CPI that had market basket weights derived from the 1994 Household Income and Consumption Expenditure Survey. For the current CPI, a national index and 26 regional indices are compiled based on 26 separate regional market baskets with separate lists of items and weights. While the old CPI covered only 34 provinces, the current CPI covers all 81 provinces.

Beginning in 2003, the HBS has been conducted on a continuous basis. In 2003 the entire sample of 25,000 households were surveyed. However, starting in 2004 one third of the 25,000 households is being interviewed each year. The CPI market basket weights are estimated each year based on expenditure tabulations for the last three years of data that are available. The current weights for the new index are based on the HBS for 2004-2006. The reference base for the new index is calendar year 2003=100.

The households for the HBS are interviewed on a flow basis during the entire year. Each household is expected to complete 5 six-day diaries where all types of expenditures are recorded. During the diary month, the households also provide demographic information and income data as well a recall data on expenditures for durable goods and services for the previous eleven months.

The household sample for the HBS is selected using Census of Population data. Starting in 2007, TURKSTAT instituted the continuous Address Based Population Registration System. This system is now being used to construct the sampling frame for selecting new households for the continuous HBS.

The HBS household sample is selected using a three-stage sampling procedure based on probability sampling techniques at each stage.2 The response rate is approximately 70 percent. Substitute households with similar demographic and housing characteristics are used to substitute for non-respondents.

Data are also provided for household expenditure according to type of outlet (open market, supermarket, etc.), item varieties and seasonality for fruits and vegetable purchases. The weight of imputed rent was estimated on the basis of the market rent for equivalent rental units. However the current CPI does not have a rental equivalences component to represent owner-occupied housing.

Approximately 360,000 price quotations are collected for 454 basic items, consisting of 851 sub-items or specifications, from approximately 25,000 outlets in 81 provinces for the monthly price survey. Most of these prices are collected by personal visits from price collectors in Ankara and the 26 regional TURKSTAT offices. For fresh fruits and vegetables and fuels prices are collected on a weekly basis, and for most other items prices are collected twice monthly during the weeks including the 10th and the 20th of the month. Rents are priced monthly from a sample of 3945 rental units.

The outlets in the outlet sample are selected on the basis of experience of the price collectors and their knowledge of the local market. The main selection criteria in each locality are the coverage of the available shopping areas within each location and the degree to which different types of outlets are represented. The number of outlets sampled varies depending on the size of the locality and the type of item being priced. Once each year, usually in November and December, the outlet sample is reviewed and updated.

The outlet sample covers different types of outlets, from market stalls, craft shops, traditional shops, single-line retail shops, big shopping centers, restaurants, hotels and catalogues. Prices for fresh fruits and vegetables and fish are collected from market stalls as well as from supermarkets and traditional shops.

When possible Turkstat staff attend international training seminars and workshops and also monitor international statistical publications to ensure that up to date procedures are employed in data collection, data processing, classification, and index compilation.

3.1.2 Source data reasonably approximate the definitions, scope, classifications, valuation, and time of recording required.

The HBS is designed and conducted in close coordination with the compilation of the CPI. Since the definitions, scope and classification systems are the same for both datasets, the estimation of new CPI market basket weights on an annual basis using the continuous HBS is greatly facilitated.

The CPI monthly price survey is also designed to in manner that facilitates the compilation of the monthly CPI. The same COICOP-HBS classification system and the same concepts that are used for the market basket weights are used for index compilation.

3.1.3 Source data are timely.

Since the household sample for the continuous HBS is divided equally between three years, the market basket weights for the CPI are estimated using three years of HBS expenditure data. The current weights are based on the HBS for 2004-2006. This allows only one year for processing the HBS for the last year (of the three years) before the data are used in the CPI. This is an outstanding turn around cycle.

The monthly price survey is conducted and processed in a manner that allows the index to be compiled and disseminated on the third business day of the month following the reference month. Fresh fruits and vegetables and fish are priced weekly and most other components are priced biweekly. The collected price data are processed on a flow basis in order to allow for the processing of 360,000 prices per month to be entered into the database and properly reviewed. Turkstat is to be complimented on the timeliness and thoroughness of the process that is carried out each month.

3.2 Assessment of source data

3.2.1 Source data—including censuses, sample surveys, and administrative records—are routinely assessed, e.g., for coverage, sample error, response error, and nonsampling error; the results of the assessments are monitored and made available to guide statistical processes.

Each year the expenditure tabulations from the continuous HBS used to estimate the new market basket weights are thoroughly reviewed and evaluated for geographic consistency across the 26 regions and for atypical values. These new weights emerging from the HBS are also assessed against corresponding data from other sources such as the national accounts, the PPI, retail trade, export trade, and imports. It is apparent that for some components such as alcoholic beverages, tobacco, and some durable goods there is underreporting of expenditures. Weights for these items are estimated taking into consideration corresponding data from other sources.

It is generally acknowledged that the overall level of expenditures reported in the HBS understates significantly the actual expenditures, since corresponding production and sales data from the national accounts is higher. This, however, is not a significant problem for the estimation of CPI market basket weights since these weights are expressed in proportions (percentages) rather than in absolute values. Of course, there could be a problem if specific types of expenditures are more underreported than others. Wherever this was identified adjustments were made.

The mission engaged in a lengthy discussion with the HBS and prices staff in order to identify an obvious cause for the overall understatement of expenditures in the HBS. Turkstat seemed to be doing everything possible avoid the problem. The list of substitute households were held by the supervisors so that the interviewers could not choose the substitute households without making a good faith effort and so that they could not select the structure/household that looked to be the easiest to interview. Individual diaries were given to all household members making expenditures. Internal edits were made to determine whether, for example, households with automobiles had expenditures for benzene. In the same light household members were asked whether they were smokers, to determine whether there should be expenditures on tobacco. It was noted that the response rate on the five six-day diaries filled out by the households during the survey month did not seem to deteriorate over the month. Comparisons between reported income and total household expenditures were also made to determine where expenditure might be left out. The problem, of course, is that the income data, as with most household budget surveys, is not reliable either. The one overall conclusion was the because of cultural traditions, the Turkish people are very private about their incomes and expenditures, even among family members. It is obvious that in order to resolve this problem additional research will be required.

Since the household sample for the HBS is designed such that non-respondents are replaced with similar households, there is not a non-response problem per say. However, there could be a self-selection bias among those who choose not to participate. It is possible that the non-participating households have similar characteristic that are not taken into account with the household substitution process. Since the non-response rate is approximately 30 percent, this effect could be significant. Nonetheless, it appears that short of greatly expanding the sample and using extensive non-interview adjustment factors, there is not an obvious solution to this problem.

Prices data from the monthly price survey are thoroughly verified both in the regional offices as well as in the central office. Atypical prices are verified with the respondents. Cross regional checks are made for all price and index data. Independent audits of price collection are made on a random basis and reports of the conclusions of these audits are given to the supervisor of the price collector involved for an explanation.

3.3 Statistical techniques

3.3.1 Data compilation employs sound statistical techniques to deal with data sources.

In general, methodologically sound procedures are employed in estimating CPI market basket weights, processing monthly price data, and compiling the monthly CPI. Strict procedures are in place for reviewing and monthly price data to avoid coding, editing and tabulation errors. Atypical values are verified with the respondent and only change when it is determined that an error has occurred.

The market basket weights, the outlet sample, and the list of products and specifications are reviewed and updated each year. Since the market basket item list and the product specification are reviewed and updated each year there is no significant problem with disappearing products. However, when a specified product is not available in a given outlet, an imputed price is calculated using either the price change of another comparable product, the average price change of similar products, or the index change of the subgroup to which the product pertains. If the product specification continues to be unavailable a substitute specification is priced. For items that are permanently unavailable in a particular location, a replacement outlet that carries the same item is selected or a new variety that is similar in quality is selected. Appropriate quality adjustments are made using a variety of internationally recommended techniques.

For seasonally unavailable items, such as fresh fruits and vegetables and some clothing items, monthly weights are used. Missing prices are carried forward while their weights are shifted to similar items until the items are again available for pricing. This process is carefully monitored to identify any shift is the seasonal availability of these items.

3.3.2 Other statistical procedures (e.g., data adjustments and transformations, and statistical analysis) employ sound statistical techniques.

The weights for the CPI are estimated using the concept of monetary expenditures associated personal consumption. In most cases this conforms to the concepts of consumption expenditures outlined in the 1993 SNA and the CPI Manual. However, in some cases there are differences. Own produced consumption goods and services such as home grown food and imputed rent for homeowners are excluded from the CPI market basket weights in Turkey, since they are not monetary expenditures. It is common practice in many countries to exclude home produced food from the CPI market basket since these expenditures tend to be small, particularly in urban areas. However, in the case of owner-occupied housing, there is a problem with not including a component to represent the value of the housing services consumed by homeowners. In the case of Turkey, 70 percent of the households are owner-occupants. If rental equivalence is not used, it is recommended that net purchases of residential dwellings be used to represent housing costs for homeowners. Neither approach is used for the CPI market basket. In other respects, the recommendations concerning the valuation of consumption goods follow internationally recommended procedures. Purchases of goods and services are valued at the actual prices paid by households regardless of the method of payment.

Methodological techniques used to compile the CPI generally follow international recommendations. Elementary aggregates are compiled using geometric means. Higher-level aggregates are compiled using the Laspeyres index formula that compares prices in the current period with prices in December of the previous year. New market basket weights based on three years of HBS expenditure data are introduced each year. These weights are updated for price change between the HBS survey years and the December in which they are linked into the index. While the real base (the period from which the expenditure weight are drawn) of the index changes each year, the reference base for the CPI continues to be 2003=100. For the current CPI, the weights are estimated using expenditure data from the HBS surveys of 2004-2006.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Intermediate results are validated against other information, where applicable.

Validation of CPI indices is carried out on a regular basis by comparing these data with corresponding price and index data from the PPI, including agricultural price data, as well as export and import price data.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated.

Atypical index behavior is assessed on a continuous basis. Each month, the behavior of the CPI indices at all levels is compared across the 26 regions to help detect errors in data collection and/or index processing.

3.4.3 Statistical discrepancies and other potential indicators of problems in statistical outputs are investigated.

Geographical aggregation of indices to obtain the national index and imputation procedures for missing prices are carried out in a manner that do not produce inconsistent aggregate results regardless of the order of aggregation.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used internally to inform statistical processes (see also 4.3.3).

Each year thorough studies are conducted with regard to the revision of market basket weights, the outlet sample, the list of market basket items, and the product specifications for price collection. The basic source data for theses studies are expenditure tabulations for each of the 26 regions from the continuous HBS, the annual tourism survey, and a special annual survey for the institutional population. Additional sources, including retail sales data and industrial production data, are also used for certain items such as alcohol and tobacco product that tend to be underestimated by the HBS. One study carried out, with regard to market basket weights, is a comparison of expenditure shares by product and group across the 26 regions of the country.

Reports are written to inform the authorities of the results of these studies. These reports are then distributed to the regional office so that the index can be updated. This process also informs the data collection effort and the HBS. For example, the HBS household sample is being expanded due to insufficient expenditure data in some geographic areas.

4. Serviceability

4.1 Periodicity and timeliness

4.1.1 Periodicity follows dissemination standards.

The CPI is compiled and published monthly, thus meeting SDDS standards.

4.1.2 Timeliness follows dissemination standards.

The CPI is published on the third work day of the month following the reference month, thus exceeding SDDS standards.

4.2 Consistency

4.2.1 Statistics are consistent within the dataset.

Index aggregation procedures ensure that aggregates of regional indices at all levels of detail are consistent with corresponding indices at the national level.

4.2.2 Statistics are consistent or reconcilable over a reasonable period of time.

When a major index revision takes place, the old index is carried forward using price changes from the new series and weights from the old series. The new index is also compiled with the new weights going backward to the reference base for the new series. Continuous CPI data at the national level are available since 1980.

4.2.3 Statistics are consistent or reconcilable with those obtained through other data sources and/or statistical frameworks.

The CPI data are consistent with corresponding data from other national statistical series such as the RPI, the PPI, agricultural price statistics, and the national accounts.

4.3 Revision policy and practice

4.3.1 Revisions follow a regular and transparent schedule.

The market basket weights for the CPI are updated on an annual basis using tabulations on expenditure data from the continuous HBS. Given that under the structure of the continuous HBS one third of the household sample is interviewed each year, the CPI market basket weights are derived from three years of HBS expenditure data. The current weights for the CPI are based on the HBS data for the years 2004-2006. Although new weights are linked into the index each year, the reference base for the index is maintained as calendar year 2003=100. It should be emphasized that this transition to an annual chained CPI is a significant accomplishment eliminating the need for costly and time consuming periodic major revisions of the index. An important advantage of this type of index is that improvement to the CPI can be introduced on a flow basis each year. It also provides a framework for updating the sample of outlets for price collection, the list of market basket items, and the product specifications.

4.3.2 Preliminary and/or revised data are clearly identified.

All CPI indices are final when published. Since they are not subject to revision and/or correction, there are no preliminary indices.

4.3.3 Studies and analyses of revisions are made public (see also 3.5.1).

Although thorough revision studies are conducted when weights, outlets, items and specifications are revised each year using the results of the continuous HBS and other sources, these studies are technical in nature and are not really appropriate for dissemination. However, reports are written to the authorities and distributed to the regional office so that index can be updated.

5. Accessibility

5.1 Data accessibility

5.1.1 Statistics are presented in a way that facilitates proper interpretation and meaningful comparisons (layout and clarity of text, tables, and charts).

The principal results of the CPI are disseminated every month on the Internet and in the hardcopy “Press Release: Consumer Price Index Month/Year.” The press release includes: (1) A brief analysis of the important changes in the index; (2) A chart comparing the 12-month percent changes for the current and previous year; (3) A series of tables presenting percent changes for one month, 12 months, current month to December of the previous year, and a 12 month moving average of percent changes. These tables cover the 12 major COICOP groups, the 26 regions, historical indices for four years and special indices such as core inflation indices. In addition complete historical data are available on the TURKSTAT website. In spite of presenting a number of different types of percent changes, the tables are well designed and easily readable. This is enabled by using the same general format for all of the tables that are presented.

5.1.2 Dissemination media and format are adequate.

Both the hard copy monthly CPI press release and the more detailed tables CPI tables available on the Turkstat website have well designed formats. The website has site map that is very useful in locating the index data for each dataset.

5.1.3 Statistics are released on a preannounced schedule.

The CPI is released on the third business day of the month following the reference month of the index.

5.1.4 Statistics are made available to all users at the same time.

The CPI is simultaneously released to all users on the Turkstat website and at a press conference held at Turkstat headquarters each month. No advance information is provided to anyone outside of Turkstat.

5.1.5 Statistics not routinely disseminated are made available upon request.

Turkstat sponsors an excellent website that allows index users to download detailed historical price statistics data. The content/format of the tables can be determined by the index user. Monthly index data are available starting in 1980 for the national CPI. At both the national level and the regional level detailed index data are available for virtually all index series compiled by Turkstat on the reference bases of 1994 and 2003. Many of the series are extended forward and backward on these two reference bases. There is no charge for use of this service.

5.2 Metadata accessibility

5.2.1 Documentation on concepts, scope, classifications, basis of recording, data sources, and statistical techniques is available, and differences from internationally accepted standards, guidelines, or good practices are annotated.

Metadata on the CPI are available in both Turkish and English on the Turkstat website. In addition to providing detailed information on the current index, there are also metadata on historical index series. Although these metadata are relatively complete, it would be desirable to have a separate document giving a complete formal documentation of the methods employed for compiling the CPI.

5.2.2 Levels of detail are adapted to the needs of the intended audience.

Turkstat provides information on statistical series to suits the needs of different types of index users. In addition to detailed metadata, there is, for example, a brochure on frequently asked questions concerning the CPI and the PPI.

5.3 Assistance to users

5.3.1 Contact points for each subject field are publicized.

The Turkstat website lists the contact persons by each statistical dataset, and their telephone numbers and e-mail addresses for queries and data requests.

5.3.2 Catalogs of publications, documents, and other services, including information on any charges, are widely available.

The Turkstat website contains exhaustive lists of databases, index series, publications, and research documents. The excellent site map is extremely helpful in finding any type of information that is available. While most of the documents and index series can be downloaded without charge, for those statistical products which are sold, the information and prices on how to obtain the items is well presented.

RECOMMENDATIONS

  • Estimate the CPI market basket weights of durable goods using households’ net purchases.

  • Expand CPI coverage to include housing costs for owner-occupants, possibly by compiling a separate index similar to those used to measure core inflation.

  • Continue research to determine the cause of the underestimation of personal consumption expenditures by the HBS.

  • Prepare a detailed formal documentation of the methodology employed for compiling the CPI.

III. Producer Price Index

0. Prerequisites of quality

0.1 Legal and institutional environment

0.1.1 The responsibility for collecting, processing, and disseminating the statistics is clearly specified.

The existence of an official statistical agency dates back to 1926 in Turkey. A centralized statistical system was established in 1930. The former statistical agency was reorganized to become the State Institute of Statistics (SIS) under the Law on Statistics, Law No. 53, of June 1962. That Law was supplemented by Decrees No’s 219, 357 and 403 in 1984, 1989, and 1990, respectively. A revised statistical law was prepared, with a view to making Turkey’s statistical provisions fully consistent with Eurostat guidelines in support of Turkey’s candidacy for European Union (EU) membership. The new law, the Statistics Law of Turkey No. 5429, was enacted on November 10, 2005. This new law totally replaces all previous legislation relating to the production of statistics in Turkey.

The new law establishes the basic principles and standards concerning the production and organization of official statistics. It provides for the creation of the Turkish Statistical Institute (TURKSTAT) and specifies its duties and authorities which are to collect, assess and disseminate all kinds of statistics relating to the economic, social, and cultural activities of the country, and to ensure coordination among other institutions and organizations that are involved in the statistical process prescribed in the Official Statistics Program (Article 1).

The Official Statistical Program (OSP) establishes the framework for official statistics that should be produced on subjects required at the national and international levels. Censuses and surveys are carried out within the framework of this Program. Taking into account existing resources, respondent burden, and cost-benefit analyses, the Program specifies the statistics needed to determine and monitor the situation of the country in the fields of the economy, social issues, demography, culture, environment, science, technology and any other required areas. The OSP is a detailed five year plan prepared by TURKSTAT with the benefit of advice from the Statistical Council (a high level advisory body that is technically part of TURKSTAT). The OSP needs to be approved by the Council of Ministers and its implementation is the responsibility of TURKSTAT. TURKSTAT produces an annual report on the implementation of the program (Article 3).

The technical independence of TURKSTAT is clearly enunciated in Article 17 of the Turkish Statistical Law. No person or organization outside TURKSTAT can give instructions to the staff of TURKSTAT in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; and observance of statistical confidentiality.

0.1.2 Data sharing and coordination among data-producing agencies are adequate.

TURKSTAT has a central role in coordinating the whole statistical system in Turkey. The duties and authorities of TURKSTAT are proscribed in Article 18 of the Statistics Law of Turkey. TURKSTAT is responsible for preparing the Official Statistical Program and for monitoring its implementation. TURKSTAT is responsible for determining the statistical methods, definitions, classifications and standards to be used in the production of official statistics (both within TURKSTAT and other institutions) in line with national and international norms. Each year TURKSTAT issues an advance release calendar for all official statistics.

0.1.3 Individual reporters’ data are to be kept confidential and used for statistical purposes only.

Article 13 of the Statistics Law of Turkey states that individual reporters’ data should not be provided to any administrative, judiciary or military authority, or person. Under Article 53 of the Statistics Law of Turkey staff who violate these provisions are liable to imprisonment and a (nominal) fine according to Article 258 of the Turkish Penal Code no. 5237.

Special aggregation rules have been developed to ensure that indirect disclosure of individual data does not occur when aggregations of data are presented. Staff review all data prepared for dissemination for possible indirect disclosure. Access to individual data is restricted to staff who require the information in the performance of their duties.

Article 14 of the Statistics Law of Turkey allows unit record data to be provided for research purposes provided that any information that could be used to identify individual respondents has been removed from the data file.

Steps are taken to secure the premises of the data-producing agency, and computer systems are password protected to prevent unauthorized access to data which is limited according to need.

0.1.4 Statistical reporting is ensured through legal mandate and/or measures to encourage response.

According to Articles 7 and 9 of the Turkish Statistical Law, TURKSTAT is authorized to request the information it deems necessary from all public institutions and organizations, and real and legal entities. All public institutions and organizations, and real and legal entities must furnish the requested information accurately and in the format and time period determined by TURKSTAT. TURKSTAT is authorized to investigate the accuracy of the information furnished, and to request additional information and documents from those concerned. The Law also prescribes that noncompliance can result in an administrative fine (Article 54). In practice, this provision is seldom applied, as serious efforts are made to create goodwill among data providers. In the case of households, penalties are never imposed. On the other hand, businesses that fail to provide the required information after repeated requests may be, and often are, fined.

To assist respondents to the annual business surveys to correctly complete their questionnaires, commonly used commercial accounting codes for relevant items have been included for each data item in the survey questionnaire. This practice was introduced for the survey in respect of reference year 2005 and continued in subsequent surveys.

0.2 Resources

0.2.1 Staff, facilities, computing resources, and financing are commensurate with statistical programs.

The PPI Team responsible for overseeing the compilation of the PPI is part of the Price Statistics Group within the Department of National Accounts and Economic Indicators. The Price Statistics Groups consists of two teams, the CPI Team and the PPI Team. The PPI Team is composed of one supervisor, three statisticians, one civil engineer, and one assistant expert. These professionals are tasked with the review of price data collected by the 26 TURKSTAT regional offices. They are also responsible for review and analysis of the producer price indices that are compiled at a national level.

There are approximately 200 TURKSTAT employees in the regional offices involved in the collection and processing of price data for the CPI, the PPI, the Construction Cost Index and the International Comparison Program (ICP) from which purchasing power parities are compiled. Although the regional directors, supervisors and some other regional staff members are permanent staff, a large number the regional employees are contract employees with one-year contracts. The authorities indicate that adequate funding is available for the price statistics programs.

For permanent staff, turnover is not a problem especially for the Price Statistics Group, where most of the staff members have more than ten years experience and no staff member has less than two years experience. All permanent staff and most contract staff hold college degrees both in Ankara and in the regional offices.

Information technology is an area that has experienced significant improvement in recent years. All staff members have individual computers that are adequate for work requirements. A state of the art Oracle database is used for data capture and index processing. Price data are entered into the database in the regional office via a direct internet link to Ankara. The compiled price series are also stored in the database. Backup and security measures are in place for all basic data and compiled indices. Excellent facilities are available for teleconferencing and internet-based training.

0.2.2 Measures to ensure efficient use of resources are implemented.

The activities of the Department of National Accounts and Economic Indicators, and their budgetary implications, are reviewed each month against the targets of the TURKSTAT work plan. After consultations with Group Leaders, the Department Head sends a report to management each month, which tracks performance of staff against a critical path for each task within the groups as well as the monthly use of budgetary allocations. The Finance Department of the TURKSTAT in turn regularly informs the departments regarding budget developments and possible additional allocations from unspent resources.

Programs are in place to minimize processing errors and improve the monitoring of price data for extreme values due to data collection and data entry errors. The data processing system allow for the comparison of price and index data over time. The time spent by price collectors to obtain the monthly price data is measured and compared to standards set through surveys by experts in the field. The price collection effort is subject to random audits to ensure that the highest standards of quality are being met.

0.3 Relevance

0.3.1 The relevance and practical utility of existing statistics in meeting users’ needs are monitored.

Under the Statistics Law of Turkey (Article 20) a high level advisory body, the Statistical Council, was established to provide advice on the preparation and implementation of the Official Statistics Program. Its functions also include identifying areas in which official statistics are required and providing advice on future work programs. The Council includes Undersecretaries of Ministries, the Governor of the Central Bank of Turkey, and representatives from the Union of Chambers and Stock Exchanges, and academia.

A web-based “User Satisfaction Survey” is now conducted every six months. The results are not published but are used by TURKSTAT to assess how well user needs are being met. Previously, major user surveys were conducted periodically (e.g., 1996, 2001, and 2004).

User needs and the relevance of the statistical program is also assessed through the Annual Research Symposium, the National Statistics Congress, advisory committees and other ad hoc meetings with government officials.

In May each year TURKSTAT organizes an Annual Research Symposium focusing on methodology issues and statistical techniques, to which employees, academics and other experts from Turkey are invited to present papers. Foreign speakers might be invited for some special sessions.

TURKSTAT also assists in organizing the biennial National Statistics Congress, in cooperation with the Turkish Statistical Association and the Association of Statistics Graduates. The first congress was held in 1999.

A contact person is assigned to deal with requests or questions from users, and unpublished, but non-confidential information, is made available on demand. TURKSTAT also sends participants to meetings and seminars of regional and international organizations that examine the content of statistical programs.

0.4 Other quality management

0.4.1 Processes are in place to focus on quality.

Management is sensitive to issues regarding the quality of statistics. A special unit, the Data Quality Control Unit, has been established within TURKSTAT to monitor the quality of statistical series. TURKSTAT has developed a “Strategic Plan, 2007-2011” that is available on its website. The plan includes detailed strategic goals and objectives relating to all statistics produced by TURKSTAT. Progress towards meeting the strategic goals and objectives is being monitored.

The Strategic Plan proposes the adoption of the Total Quality Management System from 2008.

0.4.2 Processes are in place to monitor the quality of the statistical program.

Under Article 41 of the Statistics Law of Turkey a Data Quality Control Committee has been established to review all aspects of statistical work undertaken within TURKSTAT. It is intended that Quality Reports will be prepared for statistical surveys from 2010.

0.4.3 Processes are in place to deal with quality considerations in planning the statistical program.

TURKSTAT is aware of tradeoffs in the production of statistics, especially in the context of its desire to meet EU statistical requirements.

Emerging data needs are gauged through user surveys, the Annual Research Symposium, the National Statistics Congress, advisory boards and other ad hoc meetings with government officials. Emerging data needs are at present considered to be largely those to meet EU accession.

1. Assurances of integrity

1.1 Professionalism

1.1.1 Statistics are produced on an impartial basis.

The PPI is compiled on an impartial basis and there is no external pressure on the content or the release schedule of the PPI. The technical independence of TURKSTAT is clearly enunciated in Article 17 of the Turkish Statistical Law. No person or organization outside TURKSTAT can give instructions to the staff of TURKSTAT in relation to data sources, selection of statistical methods and procedures; form, content and time of dissemination; and observance of statistical confidentiality.

Professionalism is promoted and supported within the agency. Staff recruitment is through a Public Service wide competition that includes both written and oral examinations. Upon arrival at the agency, new recruits take a one-month training course during which the mission and responsibilities of TURKSTAT are explained. The professional development of recruits is further enhanced by giving them the opportunity to work in various departments of TURKSTAT before being assigned to a permanent job. Finally, promotion to the expert level (the highest in the hierarchy below the managerial level) requires an oral and written examination and publication of a professional paper.

An internet based training facility has been established for use by new recruits, existing staff and staff of other government departments. This training program which was introduced in 2005 is conducted each Friday throughout the year. Each employee is required to complete approximately 100 hours of technical training in his statistical area. Participation is monitored randomly and the participant is required to complete an exam over the material that is covered by the course.

Every December two or three staff members from the TURKSTAT regional offices are brought into the central office for face-to-face technical training. These training sessions are conducted using teleconferencing with all appropriate regional staff. Upon return to the regional offices, the participants conduct face-to-face training as needed with the regional staff.

Staff members are also given the opportunity to take part in lectures, conferences and professional meetings.

1.1.2 Choices of sources and statistical techniques, as well as decisions about dissemination, are informed solely by statistical considerations.

The choice of data sources and statistical techniques is guided solely by considerations of a statistical nature, within the constraints of resource availability and response burden. The Price Statistics Group always refers to the latest internationally accepted methods for the choice of source data and statistical techniques. The implementation and adoption of these methods in Turkey is thoroughly discussed among staff, and final decisions are made public in advance of implementation. The timing and format of dissemination of statistical data are determined by TURKSTAT in view of statistical considerations. An advance release calendar for the coming year is issued in September each year.

1.1.3 The appropriate statistical entity is entitled to comment on erroneous interpretation and misuse of statistics.

Under the previous Law on Statistics, the SIS had the authority to comment on misinterpretation and misuse of statistics. The new Statistics Law of Turkey does not explicitly deal with this issue. However, prices staff have continued to check references to price index data in the media, and TURKSTAT requests clarification or correction when warranted. The practice of the Price Index Group is to draft a letter to the agency involved correcting the errors made. The letter is signed by the President of TURKSTAT, and the agency, usually a newspaper, is asked to publish the letter. Also, to prevent misinterpretation, the President regularly invites journalists for briefings on new developments and data problems.

1.2 Transparency

1.2.1 The terms and conditions under which statistics are collected, processed, and disseminated are available to the public.

The Statistics Law of Turkey is publicly available and is included on the TURKSTAT website in both Turkish and English. All TURKSTAT survey forms make reference to the Statistics Law of Turkey and to data confidentiality. TURKSTAT clearly identifies how to obtain information about statistical products, through identifying a contact person, with their postal address, telephone number and e-mail, in the various dissemination media.

1.2.2 Internal governmental access to statistics prior to their release is publicly identified.

There is no outside access to the PPI prior to its release.

1.2.3 Products of statistical agencies/units are clearly identified as such.

Under the Statistics Law of Turkey, any data product prepared under that law must be identified as such. The name, logo and insignia of TURKSTAT appear in all publications, press releases and on the TURKSTAT website. Likewise, PPI data reproduced by other institutions are always attributed to TURKSTAT.

1.2.4 Advance notice is given of major changes in methodology, source data, and statistical techniques.

Major changes in the PPI compilation system are discussed in advance with important user groups prior to the release of the index series that incorporate these changes.

1.3 Ethical standards

1.3.1 Guidelines for staff behavior are in place and are well known to the staff.

Apart from the provisions of the Statistics Act prohibiting, and imposing penalties for, disclosure of confidential information, staff conduct is guided by the Law on Public Servants, of which all civil servants are aware. New staff members are made aware of the guidelines and the mission and traditions of TURKSTAT when they join the agency and during the initial one-month training period.

2. Methodological soundness

2.1 Concepts and definitions

2.1.1 The overall structure in terms of concepts and definitions follows internationally accepted standards, guidelines, or good practices.

In general, the definitions, the determination of items for monthly price collection, and the estimation of weights for the PPI are based on concepts consistent with the 1993 SNA and the PPI Manual. However, index compilation is done exclusively by commodity aggregation. The PPI is not compiled by economic activity or industry. Under an economic activity-based or industry-based PPI compilation system secondary goods are taken into account directly. The PPI Manual recommends that both product-based and industry-based PPIs be compiled.

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices.

The PPI covers agriculture, mining and manufacturing, and energy, fuel and water. Although the PPI covers all sales of production to the domestic market, sales of production for export are excluded. The service sector is also excluded. Turkstat plans to expand the coverage of the PPI to include exports and services in 2011.

Since the weights for the PPI at the two-digit NACE level (for some sectors at the three or four-digit NACE level), are based on data from the national accounts, all industrial sales turnover estimated in the national accounts, with the exception of production sold for export, is included in the weights. In other words, the coverage of the PPI reflects the sales from all enterprises, including those sales that are estimated/imputed for non-reporting enterprises in the industrial surveys used for compiling the national accounts.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices.

Turkstat has adopted European standards for the classification of economic activities and production. For the PPI, economic activities are classified using the Statistical Classification of Economic Activities in the European Community, Revision 1.1 (NACE, Rev 1.1). The product classification is based on the System for Collection and Dissemination of Statistics on the Production of Manufactured Goods (PRODCOM). The eight-digit product classification system associated with PRODCOM (PRODCOM List) is used for the classification of products in the PPI except that the national classification system includes three extra digits to provide for additional detail.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks.

In accordance with the 1993 SNA, weights and prices are generally valued at basic prices, factory-gate prices including VAT. However, in the case of agricultural products, due the difficulty of obtaining true farm-gate prices, the prices used are sometimes prices obtained at farmers’ markets, usually the first point of sale for the product. This is a common practice for agricultural prices.

2.4.2 Recording is done on an accrual basis.

Both producer prices and weights are valued on an accrual basis.

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices.

Transactions between establishments within the same enterprise are recorded on a gross basis for the determination of PPI derived from national accounts data.

3. Accuracy and reliability

3.1 Source data

3.1.1 Source data are obtained from comprehensive data collection programs that take into account country-specific conditions.

Beginning in January 2005 TURKSTAT began compilation and dissemination of a new annually chained PPI with weights from two years prior to the current year. For the 1995 index, the weights were estimated using data from annual tabulations of sales data from industrial and agricultural surveys conducted in 2003. For the current year the weights were derived from calendar year 2006 data. However, the reference base for the current PPI is still 2003.

Data for compiling the PPI are obtained from a number of sources. The first is the monthly Industrial Production Survey for Manufacturing and Mining. This survey covers directly approximately 80 percent of the value of production. Each year the monthly data a tabulated by product across all establishments. For each 10 digit PRODCOM code, the establishments are presented in descending order of the value of total domestic sales of the product. The establishments that account for approximately 80 percent of sales of the product are included in the PPI establishment sample.

For electricity, factory-gate sales are determined using data from the Electricity Generation Corporation, Inc., the Turkish Electricity Trade and Contracting Corporation, and the Electricity Distribution Corporation for Turkey. The data obtained accounts fro 100 percent of the sales of electricity. Water data are obtained directly from the municipalities and from the Agricultural and Environment Department.

Agricultural data on sales are obtained from the monthly Survey of Agricultural Production. Average price data are obtained on a monthly basis from the Agricultural and Environmental Department.

The sampling frame used for the monthly Industrial Production Survey for Manufacturing and Mining is the new Business Register that is constructed using data on establishments obtained from Ministry of Finance tax registry revenue reports.

Since the weight data for the PPI come from GDP estimates from the national accounts for higher levels of aggregation—NACE, two-digit level, and sometime three and four-digit levels, the PPI covers all product sales to the domestic market.

Approximately 13,200 price quotations are collected from approximately 1,725 establishments for 756 COICOP products. The number of monthly price quotations and establishments surveyed was increased significantly with the 2005 revision of the PPI.

3.1.2 Source data reasonably approximate the definitions, scope, classifications, valuation, and time of recording required.

Since the PPI, the industrial and agricultural surveys, and the national accounts in Turkey are compiled using Eurostat recommendations including the use of NACE, Rev 1 for classifying economic activities and the PRODCOM List for classifying products, the concepts, definitions, valuations and time of recording are consistent for the source data and the PPI.

3.1.3 Source data are timely.

The current weights for the PPI are for calendar year 2006. This is an outstanding turn around cycle since the sales data from the industrial and agricultural surveys are processed in only one year. In other words the data for calendar year 2006 are processed and tabulated during 2007 to be used for determining weights and new product and establishment samples in December of 2007 for compilation of the PPI for 2008.

The monthly price survey is conducted and processed in a manner that allows the index to be compiled and disseminated on the third business day of the month following the reference month. A monthly form is used to collect by mail, fax, or e-mail, prices referring to the 5th, 15th and 25th of each month. Approximately 13,000 price quotations collected from 1725 establishments are entered into the database and properly reviewed. Turkstat is to be complimented on the timeliness and thoroughness of the process that is carried out each month to compile the PPI.

3.2 Assessment of source data

3.2.1 Source data—including censuses, sample surveys, and administrative records—are routinely assessed, e.g., for coverage, sample error, response error, and nonsampling error; the results of the assessments are monitored and made available to guide statistical processes.

Product weights, the lists of products and establishments selected for price collection for the PPI are evaluated on a regular basis. Any irregularities are verified against the source data including the monthly industrial and agricultural surveys and the national accounts weight data that are provided for the compilation of the PPI. This process serves as feedback for improvement of the monthly industrial and agricultural surveys.

Prices data from the monthly price survey are thoroughly verified both in the regional offices as well as in the central office. Atypical prices are verified with the respondents. Consistency checks are made with corresponding price data from the CPI and export and import unit values.

3.3 Statistical techniques

3.3.1 Data compilation employs sound statistical techniques to deal with data sources.

In general, methodologically sound procedures are employed in estimating PPI weights, processing monthly price data, and compiling the monthly PPI. Strict procedures are in place for reviewing and monthly price data to avoid coding, editing and tabulation errors. Atypical values are verified with the respondent and only changed when it is determined that an error has occurred.

The index weights, the establishment sample, and the list of products and specifications are reviewed and updated each year. Since the weights and list of the products and product specification are reviewed and updated each year there is no significant problem with disappearing products. However, when a specified product is not available in a given establishment, an imputed price is calculated using either the price change of another comparable product, the average price change of similar products, or the index change of the subgroup to which the product pertains. If the product specification continues to be unavailable a substitute specification is priced. For items that are permanently unavailable in a particular location, a replacement establishment that carries the same item is selected or a new variety that is similar in quality is selected. Appropriate quality adjustments are made using a variety of internationally recommended techniques.

For seasonally unavailable items, such as fresh fruits and vegetables, monthly weights are used. Missing prices are carried forward while their weights are shifted to similar items until the items are again available for pricing. This process is carefully monitored to identify any shift is the seasonal availability of these items.

3.3.2 Other statistical procedures (e.g., data adjustments and transformations, and statistical analysis) employ sound statistical techniques.

Methodological techniques used to compile the PPI generally follow international recommendations. Weighted arithmetic means are used to combine basic price observations at the first level of aggregation. The weights used at this level are based on volume of sales data obtained from the establishments where the prices are being collected. Higher-level aggregates are compiled using the Laspeyres index formula that compares prices in the current period with prices in December of the previous year. New weights based on tabulations of industrial and agricultural surveys, and GNP data from the national accounts are introduced each year. These weights are updated for price change between the survey year and the December in which they are linked into the index. While the real base (the period from which weights data are drawn) of the index changes each year, the reference base for the PPI continues to be 2003=100. For the current PPI, the weights are estimated for calendar year 2006.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Intermediate results are validated against other information, where applicable.

Validation of PPI indices is carried out on a regular basis by comparing these data with corresponding price and index data from the CPI, and export and import price data.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated.

Unusual movements in the monthly index are investigated and either corrected or explained in the monthly press release.

3.4.3 Statistical discrepancies and other potential indicators of problems in statistical outputs are investigated.

Imputation procedures for missing prices are carried out in a manner that does not produce inconsistencies in the aggregate index over time.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used internally to inform statistical processes (see also 4.3.3).

Each year thorough studies are conducted with regard to the revision of PPI weights, the establishment sample, the list of products, and the product specifications for price collection. The basic source data for theses studies are the annual tabulations of sales by product and establishment derived form monthly industrial and agricultural surveys. Reports are written to inform the authorities of the results of these studies. These reports are then distributed to the regional office so that the index can be updated.

4. Serviceability

4.1 Periodicity and timeliness

4.1.1 Periodicity follows dissemination standards.

The PPI is compiled and published monthly, thus meeting SDDS standards.

4.1.2 Timeliness follows dissemination standards.

The PPI is published on the third work day of the month following the reference month, thus exceeding SDDS standards.

4.2 Consistency

4.2.1 Statistics are consistent within the dataset.

Since PPI tabulations are made only by product aggregation, and not by industry (economic activity), there is no issue with regard to internal consistency between the two types of aggregation. Aggregation procedures employed are consistent.

4.2.2 Statistics are consistent or reconcilable over a reasonable period of time.

When a major index revision takes place, the old index is carried forward using price changes from the new series and weights from the old series. The new index is also compiled with the new weights going backward to the reference base for the new series.

4.2.3 Statistics are consistent or reconcilable with those obtained through other data sources and/or statistical frameworks.

The PPI data are consistent with corresponding data from other national statistical series such as the CPI, the national accounts, and certain components of the export and import unit value indices.

4.3 Revision policy and practice

4.3.1 Revisions follow a regular and transparent schedule.

The PPI weights, the establishment sample for price collection, the list of products and the product specifications are updated annually. These updates are based on annual tabulations for the monthly industrial and agricultural surveys. The current weights for the PPI are based on value of sales data for 2006. Although new weights are linked into the index each year, the reference base for the index is maintained as calendar year 2003=100. It should be emphasized that this transition to an annual chained PPI is a significant accomplishment eliminating the need for costly and time consuming periodic major revisions of the index. An important advantage of this type of index is that improvement to the PPI can be introduced on a flow basis each year. It also provides a framework for updating the sample of establishments for price collection, the list products, and the detailed product specifications used for pricing.

4.3.2 Preliminary and/or revised data are clearly identified.

All PPI indices are final when published. Since they are not subject to revision and/or correction, there are no preliminary indices.

4.3.3 Studies and analyses of revisions are made public (see also 3.5.1).

Although thorough revision studies are conducted when the PPI weights, the sample of establishments, the list of products, and the product specifications are revised each year, these studies are technical in nature and are not really appropriate for wide dissemination. However, reports are written to the authorities and distributed to the regional office so that index can be updated.

5. Accessibility

5.1 Data accessibility

5.1.1 Statistics are presented in a way that facilitates proper interpretation and meaningful comparisons (layout and clarity of text, tables, and charts).

The principal results of the PPI are disseminated every month on the Internet and in the hardcopy “Press Release: Producer Price Index Month/Year.” The press release includes: (1) A brief analysis of the important changes in the index; (2) A chart comparing the 12 month percent changes for the current and previous year; (3) A series of tables presenting percent changes for one month, 12 months, current month to December of the previous year, and a 12 month moving average of percent changes. These tables cover NACE categories at the two-digit level and historical indices for four years. In addition, complete historical data are available on the TURKSTAT website. In spite of presenting a number of different types of percent changes, the tables are well designed and easily readable. This is enabled by using the same general format for all of the tables that are presented.

5.1.2 Dissemination media and format are adequate.

Both the hard copy monthly PPI press release and the more detailed tables PPI tables available on the Turkstat website have well designed formats. The website has site map that is very useful in locating the index data for each dataset.

5.1.3 Statistics are released on a preannounced schedule.

The PPI is released on the third business day of the month following the reference month of the index.

5.1.4 Statistics are made available to all users at the same time.

The PPI is simultaneously released to all users on the Turkstat website and at a press conference held at Turkstat headquarters each month. No advance information is provided to anyone outside of Turkstat.

5.1.5 Statistics not routinely disseminated are made available upon request.

Turkstat sponsors an excellent website that allows index users to download detailed historical price statistics data. The content/format of the tables can be determined by the index user. Monthly historical index data are available for the national PPI at the two-digit and three-digit NACE levels. Historical index data are also available on the 1994 reference base. Historical average price data are available at the ten-digit COICOP List level. There is no charge for use of this service.

5.2 Metadata accessibility

5.2.1 Documentation on concepts, scope, classifications, basis of recording, data sources, and statistical techniques is available, and differences from internationally accepted standards, guidelines, or good practices are annotated.

Metadata on the PPI are available in both Turkish and English on the Turkstat website. In addition to providing detailed information on the current index, there are also metadata on historical index series. Although these metadata are relatively complete, it would be desirable to have a separate document giving a complete formal documentation of the methods employed for compiling the PPI.

5.2.2 Levels of detail are adapted to the needs of the intended audience.

Turkstat provides information on statistical series to suits the needs of different types of index users. In addition to detailed metadata, there is, for example, a brochure on frequently asked questions concerning the CPI and the PPI.

5.3 Assistance to users

5.3.1 Contact points for each subject field are publicized.

The Turkstat website lists the contact persons by each statistical dataset, and their telephone numbers and e-mail addresses for queries and data requests.

5.3.2 Catalogs of publications, documents, and other services, including information on any charges, are widely available.

The Turkstat website contains exhaustive lists of databases, index series, publications, and research documents. The excellent site map is extremely helpful in finding any type of information that is available. While most of the documents and index series can be downloaded without charge, for those statistical products which are sold, the information and prices on how to obtain the items is well presented.

RECOMMENDATIONS

  • In addition to compiling a PPI by product aggregations, develop a PPI compiled by economic activity or industry. This PPI would take into account secondary products.

  • Expedite planned expansion of PPI coverage by compiling export indices.

  • Move forward with plan to further expand PPI coverage to include the service sector, possibly on a flow basis, beginning with public transportation and communication.

  • Prepare a detailed formal documentation of the methodology employed for compiling the PPI in Turkey.

IV. Government Finance Statistics

0. Prerequisites of quality

0.1 Legal and institutional environment

0.1.1 The responsibility for collecting, processing, and disseminating statistics is clearly specified.

Article 23 of the Establishment Law No. 540 authorizes the State Planning Organization (SPO) in the Office of the Prime Minister to collect data required for compiling macroeconomic statistics, including government finance data for the public sector, necessary for the preparation of Turkey’s Annual Programs. The SPO data on public sector finance are also used for budgetary preparation purposes. In addition, since 2001, the SPO has been responsible for preparing Turkey’s Pre-Accession Economic Programs (including public finances) and submitting them to the European Commission. Further, the SPO has been a key source of information on extrabudgetary funds, social security institutions, and local governments covered in the IMF program-designed fiscal statistics published by the Treasury and the SPO. The SPO publishes annual data on Turkey’s public sector finance on its website.

Articles 30 and 53 of the Public Financial Management and Control Law (PFMCL) No. 5018, enacted in December 2003, authorize the Ministry of Finance (MOF) to compile and disseminate government finance statistics for the general government and to impose sanctions on institutions that do not provide timely and comprehensive data to it. Under this authority, the MOF is to establish compilation standards and practices, including coverage, accounting framework, and methods of recording and valuation, for government finance statistics. Accordingly, the By-Law on General Budget Accounting and the By-Law on Central Government Accounting, introduced by the MOF in 2006 and 2007 respectively, provide the basic rules for government accounting. In addition, under the PFMCL, the MOF has the responsibility to publish monthly budgetary revenue and expenditure data for the central government and quarterly data on the local government. The MOF is also to report government finance statistics to the IMF. With the implementation of the PFMCL, the MOF has enhanced the coverage and periodicity of government finance statistics that it publishes on its website.

Under the Law on Regulating Public Finance and Debt Management and under Article 2a of the Law on the Structures and Duties of the Undersecretariat of Treasury, the Undersecretariat of the Treasury (Treasury) in the Office of the Prime Minister is responsible for compiling financing data for the budgetary central government. Combining the MOF’s data on revenue and expenditure of the budgetary central government and its own data on financing, the Treasury publishes data on the finance of the budgetary central government on a monthly basis. In addition, as noted above, it publishes on its website fiscal data as defined under previous IMF programs for Turkey. Further, the Treasury is responsible for compiling data on debt of the public sector, which are published on the Treasury website.

0.1.2 Data sharing and coordination among data producing agencies are adequate.

The MOF, the Treasury, and the SPO coordinate their activities and share data as required to carry out their responsibilities. Consultations also take place among agencies to reconcile differences. Closer coordination among the three agencies will streamline their data compilation efforts.

Pursuant to the PFMCL, as the MOF has assumed the responsibility of compiling data on the local government, which currently is also undertaken by the SPO, closer consultations with the SPO in the transition period will facilitate the MOF’s compilation of consolidated local government data with appropriate adjustments, thereby enhancing the quality of the data. Such consultations will also facilitate the MOF’s compilation of the general government data.

In addition, pursuant to the PFMCL as the MOF moves toward compiling and disseminating general government data on a timely, accurate, and comprehensive basis, protocols on data submission, data sharing, and quality control will need to be formalized beyond existing ones with the Treasury and extrabudgetary funds; in particular, they are to include local government entities and social security institutions. The MOF has initiated efforts in this area with training programs under way for local government officials.

0.1.3 Individual reporters’ data are to be kept confidential and used for statistical purposes only.

Data from individual public corporations are kept confidential and used solely for the compilation of statistical aggregates.

0.1.4 Statistical reporting is ensured through legal mandate and/or measures to encourage response

The PFMCL gives the MOF the right of access to data for statistical compilation.

0.2 Resources

0.2.1 Staff, facilities, computing resources, and financing are commensurate with statistical programs.

Existing staff resources are sufficient to carry out current responsibilities. The MOF’s implementation in 2006 of a standard chart of accounts for use by central and local government entities and the MOF’s establishment of the say2000i government accounting system also for use by central government and local government entities have enhanced the efficiency of the MOF’s compilation of government finance statistics; they have also contributed to the significant progress the MOF has made in improving the coverage and periodicity of its government finance statistics. Further, they have facilitated the MOF’s incorporation of international standards—the IMF’s Government Finance Statistics Manual 2001 (GFSM 2001) and the European System of Accounts 1995 (ESA 1995)— in its compilation of such data.

Staff involved in compiling government finance statistics have improved their skills through training. However, specific measures are not in place to encourage retention of such trained staff.

Computer and other technology resources are adequate to meet current needs.

0.2.2 Measures to ensure efficient use of resources are implemented.

There is no separate cost information on compilation of government finance statistics. The allocation of resources is part of the overall fiscal management process.

0.3 Relevance

0.3.1 The relevance and practical utility of existing statistics in meeting users’ needs are monitored.

As noted in 0.1.1. above, various datasets are compiled and disseminated to meet specific needs of users. The SPO compiles and publishes data on the public sector as part of its preparation of Turkey’s Annual Programs and other policy-related documents, as well as its contributions to the budget preparation. The MOF compiles and publishes detailed data on revenue and expenditure separately on the central government and on the local government; these data inform on the impact of fiscal policy on the various government sectors. The MOF also reports data in the GFSM 2001 presentation to the IMF for publication in the IMF Government Finance Statistics Yearbook. The Treasury compiles financing and debt data to inform policymakers on borrowing requirements.

Turkstat periodically undertakes users’ surveys to monitor the relevance of Turkey’s official macroeconomic statistics (including fiscal data) in meeting users’ needs.

0.4 Other quality management

0.4.1 Processes are in place to focus on quality.

Statistical production is an integral part of normal operations of responsible agencies. The focus is on the application of appropriate accounting rules, since accounting records represent the primary source data for various fiscal datasets.

0.4.2 Processes are in place to monitor the quality of the statistical program.

The Court of Accounts audits aggregate budgetary government financial data against data of individual government accounting offices at the end of the year.3 The Treasury and the MOF check the consistency of data on revenue, expenditure, and financing on an ongoing basis.

The PFMCL mandates the application and publication of uniform accounting standards, in line with international standards.

Government finance statistics are based on accounting records. Data reported to the IMF in the GFSM 2001 format are produced through the use of a bridge table drawing data from the national presentations.

0.4.3 Processes are in place to deal with quality considerations in the statistical program.

Steps have been taken to enhance data quality, including improved coverage, periodicity, and timeliness. Data requirements for (previous) IMF programs and accession to the European Union have provided impetus for the improvements in data quality.

A national rolling five-year statistical plan is in place to guide further developments, including those for government finance statistics. The plan calls for improving data quality and addressing new and emerging data needs.

Three major targets set for government finance statistics are:

  • the implementation of the PFMCL, which, among other items, authorizes the MOF to set government accounting rules in line with the International Public Sector Accounting Standards (IPSASs) and to collect data from relevant entities to compile government finance statistics at the general government level on a frequent and timely basis;

  • the introduction of accrual accounting and the migration to GFSM 2001 compilation framework; and

  • the introduction of the say2000i computerized accounting system at the central and local government levels to facilitate the collection of data over the Internet from government accounting entities and tax collection offices.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis.

Government finance statistics have traditionally been compiled and disseminated on an impartial basis. The PFMCL establishes the professional independence of the departments responsible for these statistics.

Staff involved in compiling government finance statistics continue to need substantial training to compile government finance statistics in the GFSM 2001 framework.

1.1.2 Choices of sources and statistical techniques, as well as decisions about dissemination, are informed solely by statistical considerations.

Staff are, in practice, free from political, or other influence, in choosing the most appropriate sources and methods for compiling government finance statistics.

1.1.3 The appropriate statistical entity is entitled to comment on erroneous interpretation and misuse of statistics.

Compilers of government finance statistics are allowed to explain technical aspects of the statistics and to respond publicly to misinterpretations of the statistics. Explanatory material, to aid in the interpretation of the statistics, is provided on a regular basis. Additional information is also provided when asked.

1.2 Transparency

1.2.1 The terms and conditions under which statistics are collected, processed, and disseminated are available to the public.

The compilation of government finance statistics is governed by laws and regulations related to the formulation of government financial policy and related planning processes. Article 53 of the PFMCL stipulates the periodicity of compilation: quarterly for general government data, monthly for central government data, and quarterly for social security institutions and the local government. The Official Statistics Program, issued by the Cabinet of Ministers, also sets forth the responsibilities of agencies, the periodicity of data to be produced and disseminated, and their advance release calendars.

1.2.2 Internal government access to statistics prior to their release is publicly identified.

Using authorized passwords, officials of the SPO and the Treasury have access to the say 2000i database that the MOF maintains. This information is being posted among Turkey’s metadata for the SDDS.

1.2.3 Products of statistical agencies/units are clearly identified as such.

Most government finance statistics are disseminated in agency-specific publications or websites. The Official Statistics Program defines roles of agencies in the production of statistics. Further clarifications for linkages of government finance statistics can help guide users in the understanding of differences in the published statistics and reduce the potential of data misinterpretations.

1.2.4 Advance notice is given of major changes in methodology, source data, and statistical techniques.

There have been major changes in methodology, source data, or statistical techniques in the recent years and advance notice has been given of some of these changes, including updates in Turkey’s metadata for the SDDS. Beginning with the data for 2006 for the budgetary central government, the MOF has started compiling and reporting to the IMF data presented in the GFSM 2001 framework. Beginning with data for 2007, the MOF has expanded such compilation and reporting to cover data on the central government. The MOF plans to expand such compilation and reporting to cover central and local governments for the year 2009, and to cover the general government (including social security) for the year 2010.

1.3 Ethical standards

1.3.1 Guidelines for staff behavior are in place and are well known to the staff.

Staff are bound by the general rules applying to all public service staff, which are set out in Government Staff Code 657. In addition, Article 52 of the PFMCL defines responsibilities of accounting and statistical staffs involved in data compilation and dissemination, requiring their adherence to international standards, as well as principles of integrity and reliability. However, at this time, there are no specific guidelines for behavior of staff involved in compiling and disseminating government finance statistics.

2. Methodological soundness

2.1 Concepts and definitions

2.1.1 The overall structure in terms of concepts and definitions follows internationally accepted standards, guidelines, or good practices.

Concepts and definitions of budgetary data are broadly consistent with those in the Manual of Government Finance Statistics 1986 (GFSM 1986). However, the MOF has started to implement the guidelines of the European System of Accounts 1995 (ESA 95) and the Government Finance Statistics Manual 2001 (GFSM 2001), including accrual recording of transactions.

The migration to the GFSM 2001 framework is included in Turkey’s rolling five-year statistical plan. A bridge table has been developed aligning national classifications to those set forth in the GFSM 2001. In addition, the coverage of the general government and its subsectors has been delineated in conformance with concepts and definitions of the ESA 1995 and the GFSM 2001. The MOF has introduced a new chart of accounts for use by central and local government entities. It has also established standard accounting rules, including the recording of transactions on an accrual basis, to be followed by central and local government entities. The say2000i automated accounting system accessible by central and local government entities provides a central data source for government finance statistics, facilitating the compilation and quality control of data.

The MOF has started applying the GFSM 2001 framework in compiling and reporting to the IMF for publication in the IMF’s GFSY data on the budgetary central government.4 It is in the process of reporting to the IMF GFSM 2001 data on the central government for the year 2007. The MOF plans to expand such compilation and reporting on the local government beginning with data for 2009 and to include social security institutions to cover the whole general government beginning with data for 2010.

The MOF has initiated the compilation of an inventory of the government’s nonfinancial assets beginning with acquisitions made in 2007; a partial balance sheet for the central government is being developed.

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices.

The data coverage meets the requirements of the Special Data Dissemination Standard, which calls for the dissemination of data on the general government operations, the central government operations, and the central government debt.

The data on the public sector compiled and disseminated on an annual basis by the SPO are comprehensive in scope. Its data on the general government cover the central government, local administrations, extrabudgetary funds, social security institutions, and revolving funds. Its data on the public sector cover the general government as described above and state-owned enterprise.

The MOF’s data on revenue and expenditure on the central government are comprehensive. The MOF data on revenue and expenditure of the local government are also comprehensive. Separate datasets disseminated on the MOF website on the central government and the local government, with no consolidated accounts of the two levels of government, however, pose the risk of users misinterpreting that revenue or expenditure might be double-counted in cases where intergovernmental transactions are not consolidated. Consolidated accounts of the general government need to be compiled and disseminated to the public to enhance data accuracy and reliability.

In addition, as noted in 2.1.1., the MOF has started compiling data on the central government in the GFSM 2001 presentations; however, it has yet to incorporate data on the local government and social security operations to compile data on the general government in the GFSM 2001 framework 5 Further, it has yet to disseminate on its website data that it compiles in the GFSM 2001 presentations and provide such data to Turkstat for use in the compilation of national accounts.

The Treasury’s data delineating financing (“below the line” items) as related to revenue and expenditure (“above the line” items) are comprehensive. The Treasury also disseminates fiscal data as defined in (previous) IMF programs.

With respect to debt data, the Treasury publishes on a regular basis not only data on debt of the central government, but also that of the public debt. Disaggregated data are available by location of issuance (at home or abroad), maturity, instrument, currency, and key creditors. The Treasury can enhance the scope of its coverage by disseminating an additional dataset on public debt compiled on a residency basis. The Treasury can collaborate this effort with the Central Bank of Turkey (CBT), whose data from banks are key sources for such information.

The SPO data shown in Turkey’s annual programs represent the main source of consolidated finance data on the general government and the public sector, which are published.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices.

The PFMCL defines the coverage of the general government. It also authorizes the MOF to develop an institutional list for the general government for statistical compilation purposes to meet accession requirements of the European Union. The list, to be finalized at year end of 2008, is to be used for compiling government finance statistics in 2009.

Following practices of EU member states, Turkey treats social security funds as a separate subsector of the general government and not as part of the central government.

Tax revenues received by the budgetary central government but earmarked for local governments are included in the revenue of the budgetary central government; when transfers are made to the local government, grants are recorded accordingly in the budgetary central government accounts and accounts of the local government. This reflects the fact that the budgetary central government has the discretion of whether to transfer the amount in full or in part.

Expenses are recorded on an accrual basis in accordance with the GFSM 2001.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks.

Generally, actual prices are used in government finance data. Where transactions are recorded on a due for payment basis (see below), market valuation, or the nearest proxy, is used.

Consistent with Government Finance Statistics 1986 (GFS 1986), debt liabilities are recorded at face value (i.e., the amount to be repaid at the end of the contract), rather than at market values.

2.4.2 Recording is done on an accrual basis.

All central government entities have adopted modified accrual accounting rules since 2006. Such accounting framework is being introduced in the local government and social security institutions and the process has not been completed.

Revenue and expenses (including interest payments) are recorded on an accrual basis, as well as on a cash basis.

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices.

Generally, government finance transactions are shown on a gross basis.

Tax refunds and rebates are correctly classified as (negative) revenue transactions.

3. Accuracy and reliability

3.1 Source data

3.1.1 Source data are collected from comprehensive data collection programs that take into account country-specific conditions.

The MOF’s say2000i automated accounting system collects comprehensive budgetary information on the central and the local governments, providing timely source data for the MOF’s monthly data on revenue and expenses of the central government and the local government.

The Treasury compiles public sector debt statistics based on its debt management database. Combining information in its own debt database and the MOF’s say2000i database, the Treasury compiles financing data for the central government.

Before 2008, the SPO conducted annual survey on the local government covering 905 municipalities (including all municipalities that have population greater than 20,000 and a sample of ones with population fewer than 20,000), all special provincial administrations (SPAs), all utilities and the Ilbank. For 2008, the SPO collects data via the Internet from all municipalities (3205 plus 16 metropolitan areas), all SPAs, all utilities, local administration units, and the Ilbank. The tradition of the SPO collecting data on the local government relates to the fact that Turkstat publishes its surveys on the local government with a two-year lag, and the SPO needs timely data to perform its responsibility for preparing estimates for Turkey’s Annual Programs and for contributing to various policy issues.

3.1.2 Source data reasonably approximate the definitions, scope, classifications, valuation, and time of recording required.

The PFMCL authorizes the MOF in its responsibility to compile general government statistics to align definitions, scope, classifications, valuation methods, and time of recording to international standards. Accordingly, the MOF has started applying the GFSM 2001 framework in compiling its government finance statistics.

Classifications

The MOF has developed a bridge table linking budgetary items to the GFSM 2001 classifications. The MOF has used the bridge table to derive data on the central government in the GFSM 2001 presentation and started reporting such data to the IMF for publication in the IMF Government Finance Statistics Yearbook, beginning with the data for the year 2006.

Basis for recording

The MOF has introduced the accrual-based recording for compiling its government finance statistics.

The PFMCL authorizes the MOF to establish accounting standards for the general government. The accrual accounting system introduced was developed under an agreement—Programmatic Financial and Public Sector Adjustment Loan—with the World Bank; it follows international public accounting standards.

3.1.3 Source data are timely.

The establishment of the say2000i system has provided a timely and comprehensive data source, since the system gathers information on financial transactions of the general government. The availability of comprehensive and timely data source has facilitated the MOF’s compilation and dissemination of monthly data on the budgetary central government and quarterly data on the local government.

3.2 Assessment of source data

3.2.1 Source data—including censuses, sample surveys and administrative records—are routinely assessed, e.g., for coverage, sample error, response error, and nonsampling error; the results of the assessments are monitored and made available to guide statistical processes.

The main validation of all government finance data is via the benchmarking of preliminary data to audited data for all agencies. No special assessments are made of the relationship between preliminary and final data.

Budget realizations are subject to validation assessments undertaken by the SPO as part of its work in preparing for the Annual Programs.

3.3 Statistical techniques

3.3.1 Data compilation employs sound statistical techniques to deal with data sources.

Comprehensive information is gathered by the MOF, the Treasury, and the SPO. As noted earlier, for 2008, the SPO’s collects data on the local government via the Internet, covering all municipalities, as opposed to its previous method of using sampling techniques to derive estimates for the local government.

With the introduction of accrual recording in compiling government finance statistics, beginning with the budgetary central government data for 2006, the MOF has revised its data on budgetary central government for the years 2000 through 2005 and disseminated them on its website. However, the MOF has not revised data on the local government and social security funds.

3.3.2 Other statistical procedures (e.g., data adjustments and transformations, and statistical analysis) employ sound statistical techniques.

The MOF’s separate presentations on its website of data on the central government and on the local government, unaccompanied by consolidated accounts to take account of intergovernmental transactions, has subjected the data to misinterpretations, including the appearance of double-counting of expenses, when certain expenses appear in both the central government and the local government data. The MOF is working on developing procedures on data adjustments and transformations to reflect properly intergovernmental transactions to produce consolidated accounts of the central government and the local government. It also needs to develop similar procedures to incorporate data on social security operations to derive data on the general government.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Intermediate results are validated against other information, where applicable.

No validations of intermediate government finance data are carried out.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated.

Not applicable.

3.4.3 Statistical discrepancies and other potential indicators of problems in statistical outputs are investigated.

Statistical discrepancies between the deficit/surplus and financing, and between financing and changes in gross debt, are investigated and resolved.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used internally to inform statistical processes (see also 4.3.3).

The SPO routinely reviews the coverage and accuracy of various datasets on government finance; it uses these analyses to inform its statistical processes to compile data on the public sector finance, which appear in Turkey’s Annual Programs, Pre-Accession Economic Programs, and Medium-Term Program. The SPO could leverage its expertise in this area to other responsible agencies to strengthen the coverage and accuracy of their datasets.

4. Serviceability

4.1 Periodicity and timeliness

4.1.1 Periodicity follows dissemination standards.

Periodicity of government finance statistics follows the SDDS and exceeds it in some areas. Specifically:

  • The MOF’s data on budgetary central government operations are published monthly. (In addition, Turkey publishes annual data on central government operations.)

  • The Treasury’s data on central government debt are published monthly, exceeding the SDDS prescription for quarterly data.

  • The SPO’s data on general government operations are published annually. (Turkey’s data go beyond the general government to cover the public sector, including state-owned enterprises.)

4.1.2. Timeliness follows dissemination standards.

Timeliness of government finance statistics follows the Special Data Dissemination Standard (SDDS) and exceeds the SDDS prescriptions in a number of areas. Specifically:

  • The MOF’s monthly data on budgetary central government operations are published within 15 days after the reference period (exceeding the 30-day lag prescribed in the SDDS). In addition, exceeding the SDDS requirements, its quarterly data on the local government are published within one quarter after the reference period (the SDDS does not require the dissemination of quarterly local government data).

  • The Treasury’s monthly data on central government debt (both domestic and external6) are published within three weeks after the reference period, exceeding the SDDS prescribed quarterly data with a quarterly lag.

  • The SPO’s annual data on general government operations are published within two-quarters after the reference period.

The Official Statistics Program provides detailed information on the timeliness and periodicity for various datasets, as well as advance release calendars for them.

4.2 Consistency

4.2.1 Statistics are consistent within the dataset.

Government finance statistics are internally consistent. The SPO makes adjustments to various datasets and supplements with additional information to compile comprehensive consolidated data on the general government and the public sector. The SPO could leverage its expertise in compiling consistent and consolidated data on the public sector finance to the MOF and other responsible agencies through the promotion of their compiling and disseminating consolidated data, in addition to the publication of separate datasets on subsectors of the general government. Such enhancements in data quality would, in turn, lessen the burden on the SPO in gathering additional data and undertaking extensive data reviews and adjustments in the preparation of its Annual Programs.

Government domestic and foreign borrowing and amortization data are derived from the system used by the Treasury to manage debt. This system records all debt contracts made by public sector agencies and tracks all transactions (borrowing, amortization, interest, fees and charges) associated with the contracts. Data on general government financial assets are derived from accounting records of cash holdings. This ensures consistency between changes in liabilities (debt), financial assets (cash), and financing transaction data.

Annual data are compiled by summing sub-annual data—there are no benchmarking adjustments required.

4.2.2 Statistics are consistent or reconcilable over a reasonable period of time.

Time series on government finance statistics are consistent over time—the MOF, the Treasury, or the SPO are able to explain all divergences from the expected trend. With the introduction of the GFSM 2001, the MOF has produced time series data on an accrual basis on the budgetary central government going back to year 2000. The MOF plans to develop time series data on an accrual basis on the central government; however, it does not plan to develop such data on an accrual basis on the local government.

4.2.3 Statistics are consistent or reconcilable with those obtained through other data sources and/or statistical frameworks.

There are no general government national (sectoral) accounts for Turkey. Turkstat collects data from the MOF and the SPO and modifies these data as required for input into the national accounts. However, no reconciliation is carried out between government finance and national accounting aggregates.

Domestic debt statistics are compiled and disseminated on a net basis, excluding interest; however, in banking statistics, domestic debt stock is shown in market prices, produced from the balance sheet of banks.

Fiscal data are not fully reconciled with monetary data on government accounts with the banking system and with balance of payments data (viz., external debt data). Consistency checks among fiscal data, monetary statistics, and balance of payments statistics are seldom undertaken by the MOF and the CBT.

4.3 Revision policy and practice

4.3.1 Revisions follow a regular and transparent schedule.

The publication of government finance statistics follows a regular schedule whereby initial data are preliminary and are subsequently replaced by final data. Final data are subject to subsequent revisions, but this does not occur on a routine basis.

4.3.2 Preliminary and/or revised data are clearly identified.

Preliminary data are not clearly identified, but are notified by a general comment on the publication; an example is “Unless otherwise indicated, data are preliminary when first released.”

4.3.3 Studies and analyses of revisions are made public (see also 3.5.1).

No time series or analyses of revisions to data, or of the relationship between preliminary and final data, are published.

As noted earlier, the SPO routinely reviews the adequacy of data produced by other agencies and uses such reviews to inform its statistical processes, including the collection of additional data where it deems necessary. The SPO’s reviews of data could be made public to guide users in understanding how such analyses inform statistical processes.

5. Accessibility

5.1 Data accessibility

5.1.1 Statistics are presented in a way that facilitates proper interpretation and meaningful comparisons (layout and clarity of text, tables, and charts).

Government finance statistics are disseminated in various datasets and by a number of agencies (the SPO, the Treasury, and the MOF). There is, however, no single publication or central database dedicated to providing an overview of a complete set of government finance statistics; and users need to navigate through a number of separate publications or websites to assemble the full range of such data. Further, linkages among the different datasets are not apparent to users.

As the PFMCL has designated the MOF as the agency for compiling and disseminating government finance statistics, the MOF could develop a central electronic fiscal data dissemination bulletin board on its website to provide users’ easy access to various datasets through hyperlinks. The bulletin board could be prefaced with salient features of the various datasets noting differences and linkages; it should also provide guidance for data reconciliations; and responsible agency contacts should also be identified.

In addition, the English version of the MOF website is less comprehensive than its Turkish version. For the Treasury and the SPO, the English and the Turkish versions of their respective websites are equally comprehensive.

5.1.2 Dissemination media and formats are adequate.

In addition to posting data on their respective websites, the MOF, the Treasury, and the SPO publish their data in various documents, including the Public Accounts Monthly Bulletin, the Government Gazette, budget documentations, the SPO’s Monthly Main Economic Indicators and Annual Programs, and the Treasury’s Public Debt Management Report.

Government finance statistics in the GFSM 2001 presentation, a new dataset the MOF has compiled for statistical reporting to the IMF beginning with data for 2006, are not posted on the MOF website; they are published in the IMF Government Finance Statistics Yearbook. The MOF has indicated that it plans to post the GFSM 2001 data on its website in due course.

5.1.3 Statistics are released on a pre-announced schedule.

An advance release calendar (a year ahead) that gives precise release dates for key fiscal datasets is disseminated on the Turkstat’s website (www.die.gov.tr/turcat/sdds/sdds.htm). The MOF is working on developing additional advance release calendars for the high-frequency GFSM 2001 data that it plans to compile and report to the IMF.

5.1.4 Statistics are made available to all users at the same time.

Data on the budgetary central government are released simultaneously to all interested parties in the press release Central Government Budget Realization, issued by the MOF.

Central government debt statistics are released simultaneously to all interested parties by the Treasury in its monthly and quarterly press releases on Outstanding Domestic Debt and External Debt Stock, as well as in its monthly Public Debt Management Report.

General government statistics are released simultaneously to all interested parties in the SPO’s Monthly Main Economic Indicators and its Annual Programs.

5.1.5 Statistics not routinely disseminated are made available upon request.

Nonpublished (but nonconfidential) sub-aggregates are made available upon request. In addition, a freedom of information act (viz., Information Access Law No. 4982) has been implemented to provide the public their access to unpublished data.

5.2 Metadata accessibility

5.2.1 Documentation on concepts, scope, classifications, basis of recording, data sources, and statistical techniques is available, and differences from internationally accepted standards, guidelines, or good practices are annotated.

Explanatory notes accompany data published by the MOF, the Treasury, and the SPO. On the MOF website, metadata are more comprehensive on the Turkish version than on the English version.

Summary descriptions of methodology and classification for data covered under the SDDS are available on the IMF’s Dissemination Standards Bulletin Board.

However, more detailed metadata are needed to inform policymakers and market participants on major differences among the various datasets. Consistent with 5.1.1.above, there is a need for the publication of a user guide to Turkey’s government finance statistics, with clear explanations on the scope, concepts, definitions, classifications, and methodology for individual datasets, as well as on their linkages and differences. Information on ways to reconcile the different datasets would also be useful to users. Such documentation could be posted on the electronic data dissemination bulletin board as envisioned in 5.1.1. above. (See also 1.2.1.)

5.2.2 Levels of detail are adapted to the needs of the intended audience.

More data and metadata should be made available on the English version of the MOF website.

5.3 Assistance to users

5.3.1 Contact points for each subject field is publicized.

Information on the contact person for each of the data categories covered under the SDDS, including the person’s telephone and facsimile numbers and e-mail address, is posted on the IMF’s Dissemination Standards Bulletin Board. Such persons can provide assistance to users of the statistics or guide them to appropriate technical experts for the data. Contact information is also available on other datasets.

5.3.2 Catalogues of publications, documents, and other services, including information on any charges, are widely available.

Detailed lists of their respective datasets are shown on the websites of the MOF, the Treasury, and the SPO.

RECOMMENDATIONS

High priority

  • To enhance accuracy and reliability of data, the MOF needs to compile and disseminate consolidated accounts of the general government on a priority basis in the near term. Consolidated data will properly adjust for intergovernmental transfers and other transactions to eliminate the appearance of double-counting of such activities. In light of the SPO’s extensive experience in compiling consolidated accounts of the public sector finance, close coordination between the MOF and the SPO will help the MOF to publish such data in the near term.

  • To adhere to the international standard on the coverage, the MOF needs to incorporate data on social security institutions to compile and disseminate data on the general government. The coverage of social security institutions is also important because they represent a significant aspect of fiscal operations; the lack of their coverage in the MOF data diminishes the usefulness of such data in reflecting the share of the government sector in the overall economy. Existing data on revenue and expenditure of social security institutions available to the MOF provides the key source data for the coverage of such operations. In building capacity in compiling general government data in the near term and on a priority basis, the MOF could benefit from the extensive experience the SPO has in compiling consolidated accounts on the finance of the general government and the public sector.

  • In light of the disparate datasets published by the MOF, the Treasury, and the SPO, a clear mapping of the linkages of the various datasets is needed to guide users navigate through them. Clear explanations on the salient features of the various datasets should accompany the mapping, along with hyperlinks to the data. To enhance data accessibility, the MOF can establish a fiscal data dissemination bulletin board on its website launching such mapping and hyperlinks. To further enhance data accessibility, the MOF needs to make available on the English version of its website corresponding datasets it posts on the Turkish version.

  • To enhance data consistency, the Treasury, in collaboration with the CBT, could compile and disseminate an additional data series on the public sector’s external debt based on the residency concept, so that such data can be reconciled with the data on the public sector’s liabilities to non-residents, as shown in Turkey’s international investment position data compiled by the CBT.

Other

  • Since the SPO reviews the adequacy of macroeconomic data produced by various agencies and undertakes to collect additional information where it deems necessary, it could leverage its expertise in data reviews and analyses to inform statistical processes of other data producing agencies.

  • The MOF could share with Turkstat the data its compiles in the GFSM 2001 presentations so that such data could be incorporated in Turkstat’s data on national accounts.

  • To further assist the users in understanding the differences of the various datasets and in reconciling them, the SPO could publish its experiences in reviewing and analyzing data to inform its statistical processes in compiling consolidated public sector data covering the budgetary central government, extrabudgetary funds, the local government, social security institutions, and state-owned enterprises. Such information would also facilitate users’ analysis of the SPO data presented in Turkey’s Annual Programs, the Pre-Accession Economic Program, and the Medium-Term Program.

V. Balance of Payments Statistics

0. Prerequisites of quality

0.1 Legal and institutional environment

0.1.1 The responsibility for collecting, processing, and disseminating the statistics is clearly specified.

The compilation and dissemination of the balance of payments (BOP) statistics are the responsibility of the Central Bank of the Republic of Turkey (CBRT) and are governed by the terms and conditions of the CBRT Law of January 1970 (Law 1211), as amended by Law 4651 of April 2001. The CBRT Governor is appointed by the Council of Ministers for a five-year term and may be reappointed at the expiration of this term (Article 25). Article 43 of the CBRT Law authorizes the CBRT to directly request and collect all statistical information relating to the financial system and other statistical information that shall be deemed necessary for the surveillance of developments in the economy and the BOP, from banks, other financial institutions, and persons. The latter includes legal persons and households. This article also authorizes the CBRT to investigate and supervise the accuracy of the aforesaid information, and to request additional information and documents. The last provision of Article 43 establishes that the CBRT may publish the statistical information. Within the CBRT, the BOP Division of the Statistics Department7 is responsible for the compilation and dissemination of BOP and international investment position (IIP) statistics, as well as the data template on international reserves and foreign currency liquidity (reserves template), private sector’s long-term external debt, and total short-term external debt.

The Statistics Law of Turkey (Law 5429) of November 2005 reinforces the responsibility of the CBRT for the compilation and dissemination of BOP and the other above mentioned statistics. This responsibility is highlighted in the five-year Official Statistics Program (OSP) created by this Law; which also identifies the frequency and periodicity for data collection and dissemination, main data sources, and the methodological framework.8

0.1.2 Data sharing and coordination among data-producing agencies are adequate.

Procedures and working arrangements among agencies for timely provision of data to the CBRT for BOP compilation are adequate. In compiling the BOP, the CBRT draws on its data collection system, mainly based on foreign currency transactions data through resident banks; surveys conducted by the CBRT and Turkstat; foreign trade data provided by Turkstat; and administrative data sources.

Working arrangements with Turkstat are in place to assure timely and effective flow of data on (i) foreign trade compiled by Turkstat as specified in the Statistics Law, and (ii) surveys on travel and shuttle trade conducted by Turkstat since 2003. Likewise, working arrangements are in place for the Treasury to consolidate and disseminate total external debt data on a quarterly basis. To this end, the Treasury compiles long-term public external debt under the terms of the 1994 Treasury Establishment Law, and the CBRT provides to the Treasury data on private sector’s long-term external debt since October 2001 and total short-term external debt.

In implementing the OSP 2007-2011, the CBRT maintains regular meetings with other data producing agencies seeking to avoid duplication of effort and to take into account reporting burden for BOP compilation.

0.1.3 Individual reporters’ data are to be kept confidential and used for statistical purposes only.

Article 43 of the CBRT Law establishes that the CBRT shall not publish, disclose, or submit statistical information having a private and personal nature to an official authority or private body other than the Banking Regulation and Supervision Agency (BRSA). This article also states that collected statistical information shall not be used for purposes other than statistics and shall not be considered as evidence. Furthermore, Article 35 of the CBRT Law establishes that CBRT personnel are obliged to observe the secrecy of matters pertaining to the CBRT or to persons and institutions dealing with the CBRT that they acquire within their official capacities and due to their positions. This obligation continues to be binding even after the staff leave the CBRT. Personnel of the CBRT are subject to the provisions stipulated in the Code of Conduct for the damages they cause in connection with their duties.

The Statistics Law of Turkey (Law 5429) reinforces the confidentiality provisions. Article 13 of the Law includes a number of provision such as (i) confidential data can be accessed only by those involved in the production of official statistics only if such access is needed for performing their duties; (ii) data for a particular item is considered confidential if the number of respondents is less than three, or one of two of respondents are dominant even if the number of respondents is three of more; (iii) the confidential data cannot be delivered to any administrative, judiciary, or military authority or person, and can not be used for purposes other than statistics or as an instrument of proof; and (iv) civil servants and other staff in charge of compiling and processing these data are obliged to comply with this rule; even after the related personnel leave their duties and posts. Furthermore, Article 53 establishes that civil servants violating the confidentiality provisions of the Statistics Law shall be punished according to Article 258 of the Turkish Penal Code 5237 and subject to judiciary fines. The “Regulation of Procedures and Principles of Data Confidentiality and Confident Data Security in Official Statistics” was issued in June 2006.

In line with these provisions, surveys used for compiling BOP statistics indicate that collected information is confidential and only for statistical purposes; only authorized staff of the Statistics Department have access to data of individual banks and individual data from other respondents before they are aggregated for publication; and the computer systems prevent unauthorized access to individual data through password protection.

0.1.4 Statistical reporting is ensured through legal mandate and/or measures to encourage response.

Article 43 of the CBRT Law authorizes the CBRT to (i) directly request and collect statistical information relating to financial system and other statistical information necessary for the surveillance of developments in the economy and the BOP, from depository corporations, other financial corporations, and persons and (ii) directly request information, balance sheets, and reports of depository corporations, other financial corporations, and persons, and to investigate and supervise the accuracy of these information.

Banks and other financial corporations operating in Turkey are also required to submit their annual balance sheets and income statements along with the reports of their boards of directors and auditors to the CBRT (Article 43, paragraph 1). In line with Article 43 (paragraph 4), the CBRT has determined in circulars and guidelines that reporting forms are required to be submitted to the CBRT within one month after the end of the reference month.

Article 68 (III.b) of the CBRT Law stipulates penalties for noncompliance with reporting requirements, including fines from 500 to 1,000YTL (equivalent to 330 to US$670), and may suspend or restrict institutions’ operations authorized by the CBRT Law in the event of nonresponse. The Statistics Law also establishes the obligation to submit data for statistical purposes, including legal sanctions and similar fines for breaches of confidentiality.

The response rates have been high; therefore, the CBRT has not enforced penalties. The BOP Division seeks to encourage response by creating respondents’ goodwill, answering questions promptly, emphasizing the importance of BOP statistics for economic policy formulation, and providing assistance in completing and submitting forms. New data requests are preceded by pilot studies.

0.2 Resources

0.2.1 Staff, facilities, computing resources, and financing are commensurate with statistical programs.

The BOP Division is responsible for producing and disseminating BOP statistics (see 0.1.1). The Division is comprised of 23 professionals who are responsible for compiling the BOP, IIP, reserve template, private sector’s long-term external debt, and short-term external debt. Given these arrangements, staffing is adequate to perform existing tasks. Nevertheless, increasing in five the number of qualified staff of the BOP Division is of great importance for further expanding the set of supporting surveys and implementing upcoming methodological guidance (BPM6) in due course.

The level of professional expertise of the staff is high, and the qualifications of the staff are also adequate. Staff turnover is not high. Twenty staff members are university graduates, and nine have a Masters degree in economics, some of whom were fully financed by the CBRT. Eight staff members have participated in BOP courses and three staff members in external debt courses organized by the IMF in Washington, D.C. and the Joint Vienna Institute.

The BOP Division is equipped with computers, one for each staff member, which are LAN connected. In general, computer resources are adequate to perform existing tasks. Excel datasheets are the main software used for compiling BOP and IIP statistics. Nevertheless, the CBRT database (DB2) is already being used for compiling private sector’s long-term external debt, short-term external loans, and annual inward FDI survey data. The CBRT is working toward transferring the quarterly portfolio investment assets survey and the BOP from Excel files to the CBRT database.

0.2.2 Measures to ensure efficient use of resources are implemented.

In September of each year, the CBRT undertakes a review of its resources in the context of budgetary planning for the next three years.

The Accounting Department conducts a quarterly review to assess the cost-effectiveness in using resources to conduct work undertaken in each unit. Reviews are also conducted by the Human Resources Department to improve the work within each unit. Departments prepare their budgets for the next three years. New technology for data processing and/or data management is always tested by computer and systems analysis experts.

When necessary, the CBRT requests guidance from the Fund and OECD on the methodology and compilation of BOP/IIP statistics. Country experiences are also taken into consideration.

0.3 Relevance

0.3.1 The relevance and practical utility of existing statistics in meeting users’ needs are monitored.

The CBRT closely monitors users’ needs by taking into account their questions and requests to the Statistics Department. In line with the Information Access Law of October 2003, data needs of users are monitored through electronic mail services. The Law requires that the public institutions respond to the users queries within 15 days after receiving them. In this regard, both foreign and domestic users may send queries about BOP statistics directly to the CBRT and they generally receive responses within 15 days. In addition, the BOP Division intends to conduct a users’ survey among the subscribers to BOP and IIP statistics in the near future.

Staff of the BOP Division participate actively in regional and international meetings and conferences, including (i) OECD meetings, such as the Working Group on International Investment Statistics and the Working Party on International Trade in Goods and Trade in Services Statistics, (ii) the Coordinated Direct Investment Survey (CDIS) seminar organized jointly by the IMF and Eurostat, (iii) meetings on Eurostat BOP Working Party and the Travel Technical Group, (iv) BIS meetings on the locational/consolidated banking statistics and on the Handbook on Debt Securities Statistics. Occasionally, lectures and press conferences are organized in which outside experts and the media are invited to participate.

0.4 Other quality management

0.4.1 Processes are in place to focus on quality.

The CBRT recognizes that official statistics must have the confidence of users and exercises quality control during data production and dissemination. Validation procedures for assessing the plausibility or reasonableness of reported data are undertaken and reviewed with the reporting institutions to verify for possible misclassifications and miscalculations. All levels of the Statistics Department participate actively in the review of data prior to publication.

Data are checked for misreporting, and the CBRT reverts to the reporting institution when necessary. Staff from the BOP Division visit reporting institutions once or twice a year to review processing for data provision. Training programs in basic methodology and compilation programs are also conducted upon requests from the reporting institutions. For example, a two day workshop was conducted in Istanbul in early 2008, where staff from research departments of about twenty commercial banks participated.

0.4.2 Processes are in place to monitor the quality of the statistical program.

The Statistics Department verifies that data reporting practices are consistent with the regulations and that well-defined procedures are in place to ensure quality in the compilation process. Crosschecks are conducted and discrepancies are investigated.

During the first months of 2008, the work of the BOP Division was subject to an internal audit conducted by the CBRT’s Internal Audit Department.

For problems in data collection, the reporting institutions are informed and guided by the staff of the Statistics Department. The CBRT uses the media to monitor information on large external debt transactions and large-scale direct investment operations.

The Statistical Council established by the Statistics Law monitors the implementation of the five-year Official Statistics Program and provides advice when needed. The Governor of the CBRT is a member of the Statistical Council. The Council meets at least once a year.

0.4.3 Processes are in place to deal with quality considerations in planning the statistical program.

The five-year Official Statistics Program, prepared and reviewed annually by Turkstat in consultation with the CBRT, monitors the consistency with international standards, future plans, and advance release calendars.

CBRT management recognizes the tradeoffs among the different dimensions of data quality. Timeliness is regarded as one of the most important dimensions. This is reflected in the compilation and dissemination of monthly BOP statistics and annual IIP data with monthly IIP indicators being produce since early 2008. To accommodate this in the event of delays in reporting, some data are published as provisional.

Subscribers to BOP and IIP statistics automatically receive these data through links to the CBRT website. The number of subscribers has increased from about 2,000 in 2001 to 3,850 subscribers in November 2008. Improvements of data quality are taken into account in forward planning. Comparing with 2001, considerable data quality improvements were achieved. For example, annual data on inward FDI and quarterly data on portfolio investment assets are collected through surveys conducted by the BOP Division. The CBRT also enhanced the coverage of services through the implementation of surveys.

1. Assurances of integrity

1.1 Professionalism

1.1.1 Statistics are produced on an impartial basis.

Article 4 of the CBRT Law establishes that the CBRT shall enjoy absolute autonomy in exercising the powers and carrying out the duties granted by the Law under its own responsibility. The terms and conditions under which the BOP statistics are produced promote the professional independence of the CBRT.

Professional competence plays a key role in recruitment and promotion policies. When major changes in methodology occur, the CBRT publishes methodological notes on the CBRT website and/or in regular quarterly publications. Staff participates regularly in regional and international meetings and conferences related to their technical work responsibilities (see 0.3.1).

1.1.2 Choices of sources and statistical techniques, as well as decisions about dissemination, are informed solely by statistical considerations.

The choice of source data and statistical techniques for BOP compilation is based solely on statistical considerations made by the CBRT staff.

1.1.3 The appropriate statistical entity is entitled to comment on erroneous interpretation and misuse of statistics.

The Statistics Department comments on erroneous interpretation of its statistics if deemed truly significant and seeks to prevent misinterpretation or misuse of statistics by providing explanatory notes in its publications, website and in press releases.

1.2 Transparency

1.2.1 The terms and conditions under which statistics are collected, processed, and disseminated are available to the public.

The laws and regulations under which BOP statistics are collected, processed, and disseminated are available to the public, including the references to the obligation to compile and disseminate statistics, and the confidentiality of individual responses. The CBRT Law is available on the CBRT website9, and both the Statistics Law of Turkey and the 5-year Official Statistics Program are available on the Turkstat website10, as well as in hardcopy formats. These documents are published in Turkish and English.

1.2.2 Internal governmental access to statistics prior to their release is publicly identified.

As indicated in Turkey’s BOP metadata on the IMF’s Dissemination Standards Bulletin Board (DSBB), a limited number of senior CBRT staff have access to data prior to their publication. No one outside the Statistics Department of the CBRT has access to the data before their release to the public.

Article 12 of the Statistics Law establishes that all institutions participating in the OSP shall take measures to ensure easy and equal access to all users and to assure that no person or authority have access to this information before its public release.

1.2.3 Products of statistical agencies/units are clearly identified as such.

Statistical products of the CBRT, including BOP statistics, are well identified in the OSP 2007–2011 (see 0.4.3).

BOP publications, such as the monthly Balance of Payments Statistics and the quarterly Balance of Payments Report, are clearly identified as a CBRT product, acknowledging Turkstat and the Treasury, among other institutions, as data sources through footnotes that accompany the relevant tables.

1.2.4 Advance notice is given of major changes in methodology, source data, and statistical techniques.

In mid-November 2008, the CBRT included a separate section of “Announcement about future changes” in the CBRT website to provide users with advance notice of major changes in methodology, source data, and statistical techniques for BOP and IIP data (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html). Previously, the CBRT informed of major changes in methodology and compilation methods of BOP and IIP when implemented.11

1.3 Ethical standards

1.3.1 Guidelines for staff behavior are in place and are well known to the staff.

The Code of Conduct for CBRT Staff provides broad ethical guidelines for CBRT staff. It regulates recruitment of personnel, their rights, duties and responsibilities, as well as disciplinary rules and rewards. In addition, the Code of Organization and Duties determines the area of work and responsibilities of each department and division of the CBRT. These guidelines are fully available to staff. When new staff join the CBRT, they participate in an orientation course that covers these guidelines thoroughly.

2. Methodological soundness

2.1 Concepts and definitions

2.1.1 The overall structure in terms of concepts and definitions follows internationally accepted standards, guidelines, or good practices.

Concepts and definitions used to compile the BOP statistics are in broad conformity with the guidelines of the BPM5. Resident institutional units are defined in conformity with the BPM5; including the residency of offshore entities and enterprises operating in free trade zones, which is attributed to the economy where they are located.12

BOP transactions are defined according to the guidelines of the BPM5, including (i) the definition of direct investment following the 10 percent ownership rule, (ii) the definition of goods according to the recommended breakdown; and (ii) the recording of financial leasing as loans in the financial account.

Foreign currency deposits held with the CBRT by nonresident Turkish citizens (“Dresdner deposits”) and commercial banks deposits at the CBRT from reserve requirements denominated in foreign currency are treated in line with the requirements of the template on international reserves and foreign currency liquidity (reserve template).

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices.

Turkey’s BOP statistics broadly cover all resident-nonresident transactions according to the BPM5; including transactions of free trade zones located in Turkish territory, shuttle trade, reinvested earnings and intercompany lending, portfolio investment assets of the private sector, and trade credit. The CBRT includes in the BOP and IIP statistics the transactions and stock data of debt securities, respectively, according to the residency of the instrument holder rather than to the economy where the securities were issued, in accordance with the guidelines of the BPM5 and the External Debt Statistics: Guide for Compilers and Users.13

The scope of the following transactions in the BOP could be further strengthened: (i) external deposits of the nonbank private sector, (ii) equity and debt securities external assets, (iii) noncash loans disbursements to the public sector, and (iv) capital transfers. These areas are not considered significant and are being kept under review.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices.

The classification and sectorization of Turkey’s BOP/IIP statistics are largely in compliance with the BPM5 since early 2002. Institutional units are classified in four sectors (monetary authority, general government, banks, and other sectors). The monetary authority corresponds to the CBRT. Other sectors covers nonbank financial corporation, nonfinancial corporations (both public and private), and households. Consequently, government guaranteed external debt transactions are attributed to the institutional sector of the borrower, and public enterprises’ claims and liabilities are attributed to other sectors.

Some deviations from the guidelines of BPM5 include (i) goods for processing and repairs on goods are classified under services on a net basis, rather than as goods14, (ii) financial derivatives are commingled in portfolio investment; and (iii) general government debt securities issued abroad and held by residents are recorded as an increased in external assets rather than as a decrease in external liabilities. These areas are not considered significant and are being kept under review.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks.

Turkey’s BOP statistics are compiled in U.S. dollars. Transactions in other currencies are converted into U.S. dollars following guidelines outlined in the BPM5. When appropriate, transactions are valued at the exchange rate prevailing on the day of transaction; otherwise, at the correspondent weighted average exchange rates for the reference period. Stocks in other currencies are converted into U.S. dollars at the end-month correspondent exchange rate. Where transaction estimates are derived from stock data, the stock data are revalued into their original currencies, when available, and then the change in original currency is converted into U.S. dollars at the average exchange rate for the reference period.

Valuation rules used for recording transactions follow the principle of market valuation as recommended by the BPM5. Goods are valued on a f.o.b. basis, services are valued at the price paid for the services provided, tradable financial instruments are valued using the price quotation from markets, and nontradable instruments are recorded at nominal value.

Some deviations from the BPM5 valuation rules include (i) general government’s bonds issued abroad are valued at nominal value rather than at market value; nevertheless, adjustments based on the residency of the instrument holder rather than to the economy where the securities were issued are made at market value; and (ii) monetary gold is not valued at current market prices but a fixed price at the end of each quarter for the subsequent three-month period.15

2.4.2 Recording is done on an accrual basis.

BOP accounting, where possible, is done on an accrual basis. However, some BOP transactions are recorded on a cash basis. Transactions of goods are recorded according to customs documents, i.e., at the time they legally enter the country. Interest, dividends, and profits are recorded on a cash basis. Reinvested earnings on FDI are recorded in the periods when earnings are accrued to the direct investor in proportion to participation in the equity of the enterprise. Services are recorded when paid not when provided. Most transactions in the financial account are recorded when foreign exchange is paid or received. No adjustments are made to approximate accruals; for example, by applying income yields to the value of assets and liabilities using various types of the extrapolation techniques.

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices.

Grossing/netting procedures are broadly consistent with BPM5. Most current account items are recorded on a gross basis. Financial account items are mainly recorded on a net basis, separately for the individual asset and liability components.

3. Accuracy and reliability

3.1 Source data

3.1.1 Source data are obtained from comprehensive data collection programs that take into account country-specific conditions.

The data sources are broadly sufficient to compile major items of the BOP statistics. Since 2001, the CBRT has made significant progress in developing new/improved surveys to replace/supplement the banks’ foreign exchange records for a number of BOP items.

Banks’ foreign exchange records are collected from commercial banks and the CBRT. The Monthly Foreign Exchange Position Report submitted by the banks to the CBRT consists of codes established to classify all foreign exchange transactions, both for inflows and outflows. Transactions for a number of services, income, and current transfers are not identified and/or classified appropriately in the banks’ foreign exchange records. Therefore, surveys are conducted for some of these items to overcome this drawback.

Currently, the primary data sources to compile BOP statistics are:

Box 1. Balance of payments statistics—Main data sources

• Foreign trade data compiled by Turkstat

• Banks’ foreign exchange records

• Surveys conducted by Turkstat:16

  1. Shuttle trade (since 2003)

  2. Departing nonresident visitors—Travel receipts (since 2003)

  3. Arriving citizens—Travel expenditures (since 2003)

• Surveys conducted by the CBRT:

  1. Annual inward FDI survey (since 2001)

  2. Quarterly portfolio investment assets survey (since 2006)17

  3. Surveys on services

    • a. Air transportation—Resident operators (since 2005)

    • b. Air transportation—Nonresident airlines (since 2005)

    • c. Insurance services (since 2006)

    • d. Telecommunication services (since 2003)

    • e. News agencies services (since 2006)

    • f. Catering (since 2005)

    • g. Ground handling (since 2005)

    • h. Construction services in 200818

• Survey conducted by the Treasury

  1. Annual outward FDI survey in 200817

• Administrative sources:

  1. FDI data compiled by the Treasury to cross check FDI data from banks’ records

  2. Public sector debt stock data provided by the Treasury

  3. Debt securities held by nonresidents provided by custodian banks

  4. Equity securities held by nonresidents provided by Central Registry Agency

  5. Ministry of Labor and Social Security for compensation of employees (debit).

  6. General Directorate of Land Registry and Cadastre for real estate bought by nonresidents (recorded in FDI).

The financial press is closely monitored for information on international transactions, especially for financial account transactions such as FDI and portfolio investment. Goods: The main source of information for imports c.i.f. and exports f.o.b. is the monthly data on foreign trade flows compiled by Turkstat from customs documents and from major enterprises when customs documents are not available. The CBRT makes coverage and classification adjustments to goods trade data compiled by Turkstat to comply with BPM5 (see 3.3.2).

Transportation: Data for passenger fares, excess baggage, and other transportation receipts and expenditures regarding air transportation are collected through surveys from domestic and foreign airline operators. Turkstat provides freight and freight insurance expenses, disaggregated by the residency of operations and means of transportation, from customs declarations, rather than using a fixed coefficient based on previous periods.

Travel: Data are based on sample quarterly surveys conducted by Turkstat, in cooperation with the Ministry of Culture and Tourism and the CBRT. The surveys cover the average expenditure of nonresident foreign visitors and nonresident Turkish citizens (deemed as “tourism revenues”) and the average expenditure of residents on their trips abroad (deemed as “tourism expenditures”). The expenditures of nonresident Turkish citizens are included in the BOP since 2003. Business and personal travel are identified separately.

Other services: The CBRT surveys on services (identified in Table 2) are the main source data for aircraft operational leasing, telecommunication, insurance and reinsurance other than freight insurance, and news agencies services. For the remaining other services, the banks’ foreign exchange records are the main data source. Currently, the CBRT is conducting a survey on construction services, and a study on embassies and consulates, which results have not yet been included in the BOP statistics.

Table 2.

Turkey: Data Quality Assessment Framework (July 2003): Summary of Results for the Consumer Price Index

(Compiling Agency: Turkstat)

article image

Income: Income data on FDI, portfolio investment, and other investment are derived from the banks’ foreign exchange records. Data on wages earned by nonresidents working in Turkey for one year or less have been recorded under compensation of employees since 2005. Data on reinvested earnings from FDI in Turkey are obtained from the annual inward FDI survey conducted by the CBRT. Data on reinvested earnings from FDI abroad are not included in BOP/IIP statistics (authorities consider that these figures are not material).

Current transfers: The general government item includes cash grants received by the government through the banks’ foreign exchange records, and payments of Turkish citizens residing abroad to the government in order to be exempted from the compulsory military service. Noncash grants to the government are not regularly recorded in the BOP. In case of natural disasters, data on noncash grants were obtained from the Turkish Red Crescent (as it was the case in 1999). Data on workers’ remittances are derived from banks’ foreign exchange records and comprise the foreign exchange remittances by nonresident Turkish citizens sent through resident banks and converted into domestic currency. Earned premiums and claims received vis-à-vis paid premiums and claims incurred for nonlife insurance are recorded under the credit and debit items in current transfers of other sectors, respectively.

Direct investment: Data on equity capital are obtained from bank’s foreign exchange records and data on reinvested earning from the FDI survey (for inward investment). Since January 2003, FDI data include real estate investment transactions of nonresidents, based on records of the General Directorate of Land Registry and Cadastre, which shows the figures of traded real estate with a country breakdown. Intercompany loans from direct investors abroad to the resident FDI enterprises are included under the other capital component. The Treasury is currently conducted an annual survey on outward FDI, which results have not yet been included in the BOP statistics. Stock data for inward FDI are compiled by the CBRT through an annual survey and stock data for outward FDI are compiled from banks’ accounting records for banks and from Treasury’s records for the nonbank sector.

Portfolio investment: Data on equity and debt securities are derived from the banks’ foreign exchange records and the CBRT, without a separate identification for security assets, but with a breakdown into assets and liabilities and a further breakdown by sectors. Bonds issued abroad by the Treasury are obtained from the CBRT records. Data on domestic debt securities held by nonresidents are derived from stocks data provided by custodian banks on security-by-security basis, and adjusted for foreign exchange and market price changes. Since 2005, data on nonresidents’ equity security investments in Turkey are obtained from the Central Registry Agency Inc. of Turkey (CRA). Previously, these data were obtained from the Istanbul Stock Exchange (ISE). In contrast with flow data recorded in the BOP, stock data on portfolio investment assets separately identify equity and debt securities. Since 2006, these data are obtained from end-investors through a quarterly survey conducted by the CBRT; this survey was conducted annually during 2001-2005. This quarterly survey is also used for the CPIS conducted by the IMF.

Financial derivatives: These instruments are not adequately compiled in BOP statistics.

Other investment: Data on loans, currency and deposits, and other assets/liabilities are obtained from banks’ foreign exchange records. Data on short-term trade credit for imports and exports are provided by Turkstat, and their payments are estimated by using a moving average method. Data on disbursements and repayments of loans are received from banks in their Monthly Foreign Exchange Position Report along with detailed credit information forms. The coverage of public external debt disbursements in the BOP statistics is mainly limited to information on cash drawings provided by banks’ foreign exchange records.

Reserve assets: Changes in reserve assets are derived from the CBRT’s foreign exchange records. The Monetary and Financial Division of the Statistics Department is responsible for compiling the Central Bank reserves data on a weekly basis. The BOP Division is responsible for the compilation and dissemination of the reserve template on a monthly basis.

In general, data collection based on surveys is adequate. Table 2 below presents information on the surveys currently used to compile BOP and IIP statistics, including the agency that conducts the survey, the year when the survey was first conducted, the periodicity and coverage, the sample size, and the surveys’ responses. In general, the response rate is quite high. Currently, there are two additional ongoing surveys, which results have not been yet included in the BOP/IIP statistics: (i) the annual survey on outward FDI conducted by the Treasury and (ii) the quarterly survey on construction service conducted by the CBRT. The coverage of inward FDI surveys may be expanded in order to include information on direct investment-intercompany lending.

Surveys for Compiling BOP and IIP Statistics in Turkey

article image

A = Annual, Q = Quarterly, and M = Monthly.

For Turkstat surveys, sample size corresponds to 2008Q3. The sample size varies due to seasonality.

Considering only largest companies and for portfolio investment excluding nonfinancial enterprises.

These surveys were previously conducted by private enterprises on behalf of the CBRT.

Survey conducted annually by the CBRT for 2001-2005.

Ongoing surveys.

Source: CBRT information provided to the 2008 ROSC mission.
3.1.2 Source data reasonably approximate the definitions, scope, classifications, valuation, and time of recording required.

Definition, scope, classification, valuation, and time of recording of source data are to a large extent in line with BPM5. Specific procedures are used to improve the methodological foundation of the information received by the CBRT from various data sources. Notably, the ITRS only provides data on cash basis, although according to BPM5, the BOP statistics are to be compiled on an accrual basis.

3.1.3 Source data are timely.

Monthly foreign trade statistics are disseminated by Turkstat within four weeks after the end of the reference month. Commercial banks submit the Monthly Foreign Exchange Position Report to the CBRT via diskette or e-mail within 21 days after the end of the reference month. Data from enterprise surveys are also timely. If exceptional delays occur, the CBRT quickly notifies the banks and/or the enterprises responsible for the delay.

3.2 Assessment of source data

3.2.1 Source data—including censuses, sample surveys, and administrative records—are routinely assessed, e.g., for coverage, sample error, response error, and nonsampling error; the results of the assessments are monitored and made available to guide statistical processes.

Administrative records of banks’ foreign exchange transactions are periodically assessed to identify and overcome, to the extent possible, problems of coverage and misclassification, particularly those related to transactions of services, income, and current transfers. Surveys on travel services and shuttle trade are also assessed periodically. Since 2001, the survey data collection has acquired greater importance.

The accuracy of surveys data is routinely assessed for non response and sample errors. The CBRT confirms high-value transactions with individual respondents through the reporting bank and/or directly for surveys’ data. There is reporting threshold for requiring verification that transactions have been correctly classified in reports submitted by banks. Large fluctuations in the values of reported transactions are investigated by the CBRT.

3.3 Statistical techniques

3.3.1 Data compilation employs sound statistical techniques to deal with data sources.

The CBRT uses data management procedures that include computerized edit checks to identify coding and other errors in source data, and to minimize processing errors of editing and coding. The monthly data on banks’ foreign exchange records are checked by the CBRT on original currency and U.S. dollar equivalents through macros within the excel file. Discrepancies on those codes that link to other reporting forms are immediately notified to the bank responsible. Monthly crosschecks are performed for the database frame for related data, such as private external debt and intercompany debt liabilities. Once all corrections are made, data are consolidated and included in the monthly BOP statistics. In general, the reporting forms are clear, easy to complete, and appropriate for computer processing.

3.3.2 Other statistical procedures (e.g., data adjustments and transformations, and statistical analysis) employ sound statistical techniques.

The CBRT employs a number of statistical techniques to adjust the source data, including:

  • Imports c.i.f. are adjusted to f.o.b. basis using information of freight and insurance provide by Turkstat.

  • Shuttle trade: Since 2003, data on shuttle trade are obtained from quarterly surveys conducted by Turkstat at the border gates. The average expenditure for the products to be sold abroad derived from the survey results are multiplied by the number of foreign visitors who are involved in shuttle trade, which are estimated by using the percentage of foreign visitors who declare that they purchased products in bulk to be sold abroad in the survey for departing nonresident visitors.

  • Freight credit and debit data are revised based on data provided by Turkstat through the customs declarations since 2003. Previously, freight data were derived from internationally accepted fixed ratios of total export and import values disregarding modes of transportation. Currently, the freight and insurance expenses are disaggregated according to the residency of operators and means of transportation.

  • Travel revenues are estimated by expanding the average expenditures of visitors departing Turkey by the population of nonresident foreign visitors and nonresident Turkish citizens. For estimating travel expenditures, the average expenditure of residents traveling abroad is expanded by the population of resident Turkish citizens.

  • Trade credit for exports and imports, are estimated by using a moving average method for repayments, based on customs documents provided by Turkstat.

  • Securities (debt and equity) issued domestically bought or sold by nonresidents are estimated from positions data obtained from the custodian banks for debt securities and from the CRA for equity securities. Debt securities are provided on a weekly basis for Government Domestic Debt Instruments (GDDI), while equity securities data are provided on a daily basis. Both data provide the nonresidents’ holdings on a security-by-security basis. Flows data are calculated from positions by eliminating the price and foreign exchange rate changes.

  • Compensation of employees is estimated based on the number of nonresidents who were granted working permits for one year of less by the Ministry of Labor and Social Security.

Coverage adjustments are made by the CBRT to account for (i) goods purchased in Turkey by shuttle traders for resale abroad estimated through surveys conducted by Turkstat, and (ii) goods purchased from /sold to nonresidents by enterprises operating in foreign trade zones. Classification adjustments are also made to account for (i) nonmonetary gold (ii) goods procured in ports by carriers, and (iii) freight and freight insurance (see Table 3).

Table 3.

Turkey: Data Quality Assessment Framework (July 2003): Summary of Results for the Producer Price Index

(Compiling Agency: Turkstat)

article image

Foreign Trade Adjustments Million of US dollars

article image

Source: Turkstat, merchandise trade data.

Source: CBRT, balance of payments data.

Source: Direction of Trade Statistics (DOTS).

Data on reinvested earning for the current year are estimated based on the data obtained from the FDI survey for the previous year until data for the corresponding year are obtained.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Intermediate results are validated against other information, where applicable.

Data obtained from banks’ foreign exchange records are checked against other independent data sources, including (i) information in the financial press to verify high-value direct investment, external debt, and other transactions; (ii) public external debt stock data provided by the Treasury; (iii) imports of gold provided by the Istanbul Gold Exchange Market; (iv) banks’ balance sheets; and (v) CBRT database for debt and equity securities.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated.

The behavior of some variables is crosschecked with related indicators: (i) interest payments and receipts are periodically assessed in relation to corresponding stock data; (ii) discrepancies between merchandise trade and the associated financial flows from the ITRS are reviewed; (iii) data on freight earnings are regularly assessed in relation to the value/volume of the trade flows. In addition, the compilation of annual IIP starting in 2002 has facilitated the reconciliation of balance of payments data with changes in the corresponding stock data collected for external debt and for other elements of the IIP. A regular detailed BOP/IIP reconciliation (in tabular form) could further strengthen validation procedures.

3.4.3 Statistical discrepancies and other potential indicators of problems in statistical outputs are investigated.

The CBRT is alert to the level of the errors and omissions and seeks to understand them by crosschecks between the BOP statistics and monetary statistics, and between the current and financial account of the BOP.

Data reconciliations are not conducted with foreign trade data reported by Turkey’s trading partners to the IMF Direction of Trade Statistics (DOTS) and creditors’ data included in the Joint External Debt Hub (JEDH).

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used internally to inform statistical processes (see also 4.3.3).

Due to strict timetables for the delivery of monthly BOP statistics, revisions are carried out routinely. Preliminary estimates are needed when the periodicity and/or timeliness of some data sources are insufficient or when some reporters exceed deadlines. Monitoring revisions is considered a normal part of the current compilation practice. The results of adjustments made in the statistical processes are considered in compiling data for subsequent periods. When major revisions are made, the results are well documented, including the causes of revision, methods to incorporate new data sources, and estimation of the magnitude of the changes made.

4. Serviceability

4.1 Periodicity and timeliness

4.1.1 Periodicity follows dissemination standards.

The CBRT disseminates monthly BOP data at the CBRT website; thereby, exceeding the quarterly periodicity prescribed by the SDDS for BOP statistics. Turkey also disseminates a range of other external sector statistics with the periodicity prescribed by the SDDS. IIP statistics are disseminated annually by the CBRT; external debt statistics are disseminated quarterly by the Treasury; and merchandise trade and the reserve template data are disseminated monthly by Turkstat and the CBRT, respectively. Data on international reserves are disseminated weekly by the CBRT; thereby, exceeding the monthly periodicity prescribed by the SDDS. In April 2008, the CBRT initiated the dissemination of a monthly IIP indicator with the same disaggregation provided for annual IIP data.19

4.1.2 Timeliness follows dissemination standards.

The timeliness of the BOP statistics exceeds SDDS requirements. Turkey’s monthly BOP statistics are disseminated six weeks after the end of the reference month (rather than three months as prescribed by the SDDS). The timeliness of other external sector data complies or exceeds SDDS requirements, including IIP data (with timelines of two quarters rather than three quarters), and merchandise trade data (with timeliness of one month rather than eight weeks) also exceeds SDDS requirements.

4.2 Consistency

4.2.1 Statistics are consistent within the dataset.

Concepts, definitions, and classifications for producing monthly, quarterly, and annual BOP statistics are the same. Over the long run, the net errors and omissions item has not been large, but has been unstable over time. Net errors and omissions of Turkey’s BOP have been equivalent on average to 0.5 percent of GDP and 1.1 percent of the total goods’ exports and imports in the past eight years.

4.2.2 Statistics are consistent or reconcilable over a reasonable period of time.

Consistent detailed annual BOP statistics are available from 1984. When changes in source data, methodology, or techniques are introduced, historical series are reconstructed as far as reasonably possible. Major changes in economic trends are announced and explained in the monthly report Balance of Payments Developments and in the quarterly Balance of Payments Report posted on the CBRT website

4.2.3 Statistics are consistent or reconcilable with those obtained through other data sources and/or statistical frameworks.

BOP and IIP statistics are broadly consistent. Minor differences between BOP and IIP data—mainly due to the recording of equity and debt security assets and general government bonds issued abroad and held by residents—are closely monitored.

BOP and monetary statistics are broadly consistent and regular reconciliation processes are in place, including official reserve assets and banks’ foreign assets and liabilities data.

The CBRT, as the financial agent of the Treasury, has comprehensive data on general government external debt operations and through banks’ foreign record obtains data for public sector enterprises. Nevertheless, data on noncash disbursements of government loans are not fully covered. Regarding national accounts statistics, the CBRT regularly provides BOP data to Turkstat for the compilation of national accounts. There are limited regular data reconciliation exercises between BOP and GFS (disseminated by the Treasury) and national accounts statistics (disseminated by Turkstat).

The total external debt data disseminated by the Treasury are reconcilable with the correspondent external debt items of the IIP data disseminated by the CBRT. The CBRT follows the residency criteria in line with BPM5 and the External Debt Guide, while the Treasury records general government debt securities based on the place of issuance (Table 4). Consequently, for end-2007, Turkey’s total external debt data based on residency amounts to US$ 264.7 billion and to US$ 247.1 billion if based on place of issuance. The difference (US$ 17.7 billion) corresponds to government debt securities domestically issued but held by nonresidents (US$ 32.2 billion) minus government debt securities issued abroad and held by residents (US$ 14.5 billion).

Table 4.

Turkey: Data Quality Assessment Framework (July 2003): Summary of Results for Government Finance Statistics

(Compiling Agencies: Ministry of Finance, the Treasury, and the State Planning Organization)

article image
article image

Turkey—Total External Debt Position Million of US dollars

article image
Table 4a:

External Debt - General Government - Bonds

article image

Disseminated at the Treasury website, linked to the NSDP, and reported to the World Bank.

Data included in the IIP posted at the CBRT website, linked to the NSDP, and reported to STA.

See Table 4b below.

See table 4c below.

Memorandum item:

article image
Table 4b.

Government Bonds Issued Abroad

article image

Recorded as government’s external debt by Treasury.

Excluded from government’s external debt in IIP statistics (compiled by the CBRT-market value).

Data disseminated at the CBRT website (IIP dataset, Table 10).

Table 4c.

Government Domestic Debt Instruments (GDDI) Held by Nonresidents 1/

article image

Data disseminated at the CBRT website (IIP dataset, Table 5).

Memorandum

article image

4.3 Revision policy and practice

4.3.1 Revisions follow a regular and transparent schedule.

In mid-November 2008, the CBRT publicized its revision policy for BOP and IIP data at the CBRT website (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html). When disseminating the current month data, BOP statistics for the current year and previous year are revised. Also, at the end of the current year, previous five years’ BOP data are subject to revision. Except for this general rule, when large-scale revision is needed in any data that compose the BOP statistics, previous years’ statistics would be revised by informing the public at large. The revision policy for IIP data has also been recently publicized by the CBRT (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html).

4.3.2 Preliminary and/or revised data are clearly identified.

Preliminary data are identified in BOP data releases; however, better identification is needed when data are revised. The revised data are disseminated with the same level of detail as previously published for the preliminary data.

4.3.3 Studies and analyses of revisions are made public (see also 3.5.1).

When major revisions to BOP statistics are made, the revisions are announced on the CBRT website. For example, in early 2004, revisions included (i) changes in the recording of travel revenues and workers remittances due to the inclusion of survey data on expenditures made by Turkish citizens residing abroad with the corresponding adjustment in workers’ remittances; and (ii) the estimation of freight revenues and expenditures due to the identification of the residency of the operator and means of transportation. A brief document explaining these revisions and identifying their impact in the current account of the BOP and in their corresponding components was posted on the CBRT website simultaneously.

5. Accessibility

5.1 Data accessibility

5.1.1 Statistics are presented in a way that facilitates proper interpretation and meaningful comparisons (layout and clarity of text, tables, and charts).

The BOP statistics are published according to the standard components of the BPM5, and with time series. Data are published with various levels of detail, and presented in a clear manner, with charts and tables to facilitate analysis.

BOP data are disseminated on the CBRT website through the monthly report on Balance of Payment Developments, which includes 37 predefined Excel tables. The CBRT Quarterly Bulletin includes time series data for BOP statistics (monthly data for the current and previous years and annual data for the previous 11 years) with a detailed breakdown of the BOP standard components.

Analyses of current period developments are included in the monthly report Balance of Payments Developments and in the annual International Investment Position Report prepared by the Statistics Department, and in the quarterly Balance of Payments Report, prepared by the CBRT Research Department.

5.1.2 Dissemination media and format are adequate.

The CBRT prepares monthly BOP data releases that facilitate re-dissemination in the media. Detailed and summary monthly data on the major components of the balance of payment are available, in both Turkish and English

Current statistics and longer time data series can be accessed through the CBRT Electronic Data Delivery System (EDDS), which is an electronic database with a very powerful query facility.

5.1.3 Statistics are released on a preannounced schedule.

An Advance Release Calendar (ARC) for the next 12 months is disseminated on the CBRT website and on Turkstat website. The ARC is updated once a year in September.

5.1.4 Statistics are made available to all users at the same time.

BOP data are released simultaneously to all interested parties by posting them on the CBRT website (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html). The data available on the website are monthly and cumulative. The statistics are released punctually, that is according to the preannounced schedule.

5.1.5 Statistics not routinely disseminated are made available upon request.

Nonpublished but nonconfidential sub-aggregated data are promptly made available upon request.

5.2 Metadata accessibility

5.2.1 Documentation on concepts, scope, classifications, basis of recording, data sources, and statistical techniques is available, and differences from internationally accepted standards, guidelines, or good practices are annotated.

The conceptual framework and compilation practice for balance of payment statistics in Turkey are presented in the document Balance of Payments Methodology and Compilation Practice in Turkey available to the public in Turkish and English on the CBRT website (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html). The document provides (i) a detailed explanation of relevant concepts, definitions, and recording principles—including examples—for adequate recording of transactions in the BOP; and (ii) a description of the main data sources and compilation practices for the main BOP components. The CBRT disseminates a document for the compilation of IIP statistics, but this document is less comprehensive than the BOP document.

Turkey’s BOP and IIP metadata are also posted on the IMF’s Dissemination Standards Bulletin Board (DSBB) Metadata on BOP and IIP statistics are updated regularly.

5.2.2 Levels of detail are adapted to the needs of the intended audience.

Explanatory notes on methodology are included in the BOP and IIP publications prepared by the Statistics Department. More specialized information notes have been prepared and made public, particularly when significant changes in methodology and compilation practices have occurred.

5.3 Assistance to users

5.3.1 Contact points for each subject field are publicized.

In line with the Information Act of 2005, the CBRT provides prompt and knowledgeable service and support to users of BOP statistics. In mid-November 2008, the CBRT included in the CBRT website a contact point (e-mail, telephone, and address) for users of BOP and IIP statistics (http://www.tcmb.gov.tr/odemedenge/odmainengyeni.html).

A contact point for IIP statistics has also been included in the CBRT website in mid-November 2008. Contact persons, address, telephones, facsimiles, and e-mails are also identified for BOP and IIP statistics on the IMF’s DSBB.

5.3.2 Catalogs of publications, documents, and other services, including information on any charges, are widely available.

The CBRT website contains a list of publications and documents related to BOP statistics, including research papers (only in Turkish) written by staff of the BOP Division, that can be obtained free of charge. All dissemination formats are also identified on the IMF’s DSBB.

RECOMMENDATIONS

High priority

  • Improve the coverage of noncash disbursements for government loans and nonbank external deposit assets; the recording of bonds issued abroad held by residents, financial derivatives, and equity/debt security assets; and further develop/improve surveys on FDI (quarterly) and services.

  • Extend market valuation to all tradable debt instruments, and record on an accrual basis investment income in the BOP and external debt in the IIP.

  • Compile regularly a detailed BOP/IIP reconciliation table and publish a summary reconciliation table in 2009.

  • Allocate more staff resources to compilation of BOP and IIP for further expanding the set of supporting surveys and implementing the upcoming BPM6 in due course.

Other

  • Disseminate quarterly IIP data (as encouraged by SDDS) and comprehensive IIP metadata.

  • Assess regularly the consistency of BOP/IIP statistics, including with data from trading partners (DOTS) and creditors’ data (JEDH).

Progress in Implementing the 2001 ROSC Recommendations on BOP Statistics

article image
article image
Source: CBRT, Statistics Department.
Table 5.

Turkey: Data Quality Assessment Framework (July 2003): Summary of Results for Balance of Payments Statistics

(Compiling Agency: Central Bank of the Republic of Turkey)

article image