Bulgaria: Detailed Assessments Using the Data Quality Assessment Framework (DQAF)
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

This Report on the Observance of Standards and Codes Data Module provides a review of Bulgaria’s data dissemination practices against the IMF’s special data dissemination standard, complemented by an in-depth assessment of the quality of the national accounts, consumer price index, producer price index, government finance, monetary, and balance-of-payments statistics. Bulgaria has adopted a restructuring program, aimed at stabilization and significant improvements in fiscal and monetary statistics. Bulgarian statistics have been relevant, consistent, and available on a timely basis with good frequency.

Abstract

This Report on the Observance of Standards and Codes Data Module provides a review of Bulgaria’s data dissemination practices against the IMF’s special data dissemination standard, complemented by an in-depth assessment of the quality of the national accounts, consumer price index, producer price index, government finance, monetary, and balance-of-payments statistics. Bulgaria has adopted a restructuring program, aimed at stabilization and significant improvements in fiscal and monetary statistics. Bulgarian statistics have been relevant, consistent, and available on a timely basis with good frequency.

I. National Accounts

0. Prerequisites of quality

0.1 Legal and institutional environment

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

The Law on Statistics, published in the Official Gazette No. 57 of June 25, 1999 and amended in No. 42 of April 27, 2001, governs the activities of the National Statistical Institute (NSI) as well as other government statistical bodies. The NSI disseminates data under Chapter 5 of this Law, which requires statistical bodies to inform the public about the results from statistical surveys through mass media, own publications, and other means. Chapter 3 of the Law on Statistics establishes the National Statistical Council, a body attached to the NSI president and responsible for the annual National Program for Statistical Surveys (NPSS). This program covers all significant statistical work undertaken by government agencies, including the national accounts, and is approved by the Council of Ministers.

The Bulgarian Annual Statistical Plan incorporates the NPSS of the NSI and other surveys that are of interinstitutional significance. For each specific statistical survey included in the NPSS, the following items must be stated: the purpose, contents, and coverage of the survey; the methods for conducting the survey, information sources, and forms of data collection; the body or bodies conducting the survey, obligations for data provision by the respondents, and the budget for the survey.

The NSI undertakes most of the data collection itself, the main exceptions being that the Bulgarian National Bank (BNB) calculates the balance of payments (BOP) and provides data on banking statistics, and the Ministry of Finance (MOF) compiles the state budget data. The NSI is the only official producer of national accounts data, and this position has never been challenged. The national accounts are prepared in the National Accounts Division (NAD) of the NSI’s National Economic Accounts and Balances Directorate, in cooperation with other divisions. Also, within the NSI, the Basic Statistics Directorate organizes the surveys and aggregates the submitted data. The same directorate produces the final results of the surveys (by activities), after grossing up the results and making adjustments for nonresponse.

The main responsibility of the NAD is to apply the national accounts system, based on the European System of Accounts, 1995 (ESA 95), to the current statististical practice of NSI. To that end, the NAD staff support other divisions within the NSI and the National Statistical System as a whole.

0.1.2 Data sharing and coordination among data producing agencies are adequate

In pursuance of the Law on Statistics the NSI is responsible for coordinating all national statistical activity. The main incentives to encourage coordination among agencies include the sharing of information, participation in joint statistical activities, access to statistical expertise, and influence on the setting of priorities. The tools to achieve coordination are as follows:

  • the NPSS statistical infrastructure—classifications, registers, definitions, indicators;

  • councils, which improve relations with users, data suppliers, and other bodies of statistics, that is, the National Statistical Council, High Statistical Council, Methodological Council, Editing Council, Methodology and Training Council, IT Council; and

  • agreements with bodies of statistics and other institutions, protocols, rules, methods, and ordinance.

The Law on Statistics specifies the obligations of legal entities to submit information in response to surveys, covered in the NPSS and conducted on an obligatory basis.

The NPSS covers statistical information required for the observation and compilation of the Bulgarian economic, demographic, social, and environmental activities and other activities, taking into account the availability of sources, respondents’ burden, and cost-effectiveness. The NPSS includes a plan for the NSI statistical surveys and surveys with interinstitutional significance. The NPSS covers regularly conducted statistical surveys of established public significance, other statistical surveys for the relevant year, and particular surveys related to the international obligations undertaken by the Republic of Bulgaria. Exhaustive censuses are conducted according to the specific laws enacted for that purpose. The NPSS contains detailed characteristics of the surveys in terms of:

  • theme of the survey;

  • contents and object of the survey;

  • method of conducting the survey;

  • sources and form of data collection;

  • body/bodies conducting the survey;

  • obligation for data provision; and

  • budget for the survey.

The actual funds needed for implementing the NPSS are provided from the State Budget. These funds are allocated to the bodies undertaking the surveys. There is no information available for the coordination of these funds, or the efficiency of their use, for monitoring. The NPSS is published in the Official Gazette.

Annual operating plans for statistical activity have been developed in conformity with the NPSS. These plans contain detailed information about the allocation of tasks between different departments and divisions, including aspects of data exchange.

In accordance with the Law on Statistics there are two principal councils, i.e., the High Statistical Council and the National Statistical Council. The High Statistical Council, a consultative body to the NSI president, provides opinions and recommendation on the NSI plan for statistical surveys, drafts of legislative instruments in the field of statistics, national statistical standards, national classifications, nomenclatures, and methods. This council, consisting of 35 members, held two meetings in 2002, when the following issues were discussed: the NPSS for 2003, the medium-term program, and the updating of the statistical classifications.

The National Statistical Council is the body responsible for drafting and implementing the NPSS; it is attached to the NSI president and consists of 18 members. The roles of the National Statistical Council are to adopt a long-term strategy for the development of statistical activity, draft the NPSS, and approve a list of the structural units in the government administration that are bodies of statistics, as well as rules on their activity.

A third council—the Methodological Council—is responsible for implementing the principles of statistical activities incorporated in the Law on Statistics—impartiality, reliability, effectiveness, adequacy, publicity, and confidentiality. It is also involved in the development of methodologies, the planning of statistical surveys, and the dissemination of statistical information. The council consists of 15 members—of which five are from the NSI and the rest are from scientific, research institutions, nongovernment organizations, and other government agencies. In general, the councils are a useful mechanism for controlling statistical activities and they meet regularly.

Also, to improve coordination and information between state bodies, agreements have been made for data exchange between the NSI and the main ministries and other state institutions. These agreements specify the obligations of the NSI and another institution in providing statistical information, as well as other forms of cooperation in the field of statistics. The terms of data provision are consistent with the requirements of the advance release calendar (ARC) for dissemination of the results of the statistical surveys.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

Confidentiality of reporting is guaranteed under Article 25 of the Law on Statistics, which states that individual and personal data received may be used only for statistical purposes. Article 22 requires that all respondents to statistical surveys are informed of their rights and obligations, including the guarantee of data confidentiality. Data on individual reporters can only be provided to third parties if written permission has been received. In addition, Article 26 of the Law on Statistics states that statistical bodies and their employees cannot make public or provide (1) individual statistical data; (2) statistical data that may be combined in a way that leads to the identification of an individual person; and (3) statistical data that summarize information for less than three units or in which the relative part of one unit is more than 85 percent of the total volume. Article 55 states that any person not complying with these provisions will be liable to a fine of up to 6,000 Bulgarian leva or BGN (about US$2,000). One accidental breach of the confidentiality rules is known to have taken place, and appropriate disciplinary action was taken. All computers and databases are password controlled to restrict access to authorized staff. Destruction of unwanted survey forms is tightly controlled. Formal procedures are in place to authorize such destruction, and this is undertaken by specially licensed contractors.

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

Article 20 of the Law on Statistics requires all organizations and individuals to submit data covered by the NPSS. Article 52 establishes a penalty of up to 6,000 leva (about US$2,000) for noncompliance with reporting requirements (including misreporting). If someone refuses to supply data, attempts are made to encourage participation. However, continuing refusal can result in prosecution, and such cases are publicized.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

At the end of 2002 the NSI had 1,689 staff. Of this total, 1,609 had permanent contracts and 89 had temporary contracts. In comparison with 2001 the number of permanent staff had decreased by 190 employees.

Twenty-one people work in the NAD, which is made up of three sectors, but the structure is quite flexible to make the best use of the experience of the staff. The national accounts division consists of the following units: (1) gross domestic product (GDP) calculations from the production and expenditure approaches—both quarterly and annually; (2) institutional sector accounts, GDP from the income approach, and the production accounts for general government; and (3) supply and use tables, and regional GDP estimates.

The workload of the NAD is increasing because of Bulgaria’s application to join the European Union (EU). The NAD is obliged to implement all of the recommendations flowing from the participation of the NSI in the Eurostat pilot projects. It is also required to shorten the deadlines for the quarterly and annual accounts in accordance with the Eurostat data transmission program. The problems of this extra work are compounded by the high turnover of staff. Since 1997 two-thirds of the experienced staff have left, and have had to be replaced by new entrants.

Low pay is a significant problem for recruiting and retaining well-qualified staff. The NSI is classified as a category 3 institution rather than category 2, such as the MOF. Category 2 would allow the NSI to improve staff pay. So, following a Eurostat recommendation, a proposal to change its status was made at the end of 2001. Most agencies supported this application, except for the MOF and the Council of Ministers. Both of these bodies said there were insufficient funds available at the time. The NSI issued another proposal at the end of 2002 but has not yet received a response.

In 2002, the Council of Ministers decided that the NSI would have to vacate its current office building in Sofia because the office space was required for other purposes. Surprisingly, and disturbingly, suitable alternative accommodation for all of the NSI’s staff was not supplied. So, as of January 2003, some of the NSI staff are still in the old office, while attempts are made to find alternative accommodation. This situation has a serious impact on staff morale. It also demonstrates a serious lack of interest in statistics throughout the higher reaches of government. Computing for the national accounts is currently adequate and has been supported under the PHARE program.

0.2.2 Measures to ensure efficient use of resources are implemented

Funds from the central budget are assigned to the NSI each year on the basis of the State Budget of the Republic of Bulgaria. The main financial problem facing the NSI relates to limited funds for its operations. It is extremely difficult to cover the most pressing expenditures, related to conducting statistical surveys and studies, as well as overheads such as electricity, telecommunications, postal, and other external services. This calls for tight controls on expenditure, so that the NSI can run all of the surveys and studies included in the NPSS and disseminate the necessary statistical information.

A first set of performance indicators was developed in 2000, according to the specific objectives of Statistics about Statistics. The new performance indicators system follows the philosophy of the Statistical Capacity Building Indicators (PARIS 21 Task Team).1 The indicators cover seven dimensions:

  • effectiveness and usefulness;

  • resources—staff, budget, infrastructure;

  • transparency;

  • timeliness;

  • accuracy and relevance;

  • respondent burden; and

  • international cooperation.

For each of these indicators the approach is to review the data available (including identification of data gaps) and their usefulness as indicators of quality. This information is then expressed as time series in order to be able to see trends over time. This process is still ongoing.

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

The NSI, under Article 7(1) of the Law on Statistics, coordinates the state statistical activity by means of collaboration with the other bodies of statistics in drafting and implementing the NPSS. It is also responsible for ensuring methodological uniformity in the surveys.

The NSI recognizes the importance of the quality of statistical data. It has adopted the quality dimensions of statistical information in accordance with the definition of quality accepted by Eurostat. The topic of quality of statistical information is contained in the Strategy for Development of Statistics and in the Priority Tasks for the period 2001–2005.

Using the General Data Dissemination System (GDDS) as a framework for the development of statistics contributes to the improvement of the production and dissemination of the statistical information in accordance with the basic parameters of the system: that is, coverage, periodicity, and timeliness of data, quality of data, integrity of data, and public access. Regular revisions are made to the national metadata for the GDDS, which are included in the Dissemination Standards Bulletin Board (DSBB) of the IMF. Work towards adoption of the Special Data Dissemination Standard (SDDS) of the IMF is continuing.

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

The statistical divisions of the NSI continuously address the issue of the quality of the data received from statistical surveys. Since 2000 the Statistical Methods and Standards Division has been engaged in studying and elaborating issues of statistical data quality. For instance:

  • the theme of quality has been given priority in the NPSS since 2000;

  • quality assessment is addressed within the development of national statistical practice in line with the requirements of the European Statistical System; and

  • focusing on quality is involved in all stages of the production and dissemination of statistical data. The main publication of NSI—the “2001 Bulgarian Statistical Yearbook”—can be presented as an example of this approach. The dissemination of the methodologies of the statistical surveys (in every chapter of the “Yearbook”) is one of the requirements of data quality. Also covered is the increasing use of administrative sources of data to reduce respondents’ burden.

User satisfaction with data quality has been researched. This has involved periodic marketing research on user satisfaction, surveys on respondent burden, and the attitude of businesses towards statistical confidentiality.

The NSI has adopted the principles of the Quality Declaration of the European Statistical System and the relevant Eurostat Leadership Group. Recommendations on quality, which will be gradually applied to different surveys, will be considered in line with the accepted criteria established by the European Statistical System. These quality reports are in compliance with the European Commission (EC) and Council Regulations for the respective surveys.

Because of the importance of national accounts statistics within the macroeconomic statistics, Bulgaria puts particular importance on their quality. While Eurostat plays a key role in improving the quality of statistics in candidate countries by providing methodological and technical assistance, the countries themselves retain the ultimate responsibility for upgrading the statistical system to the European standards. Steps have been taken by the NAD to meet the quality requirements of national accounts data in terms of reliability, exhaustiveness, and correspondence with ESA 95 and related EU regulations. This involved the identification of shortcomings in statistical techniques, methods, sources of information, and accounts. This was achieved through a number of Eurostat pilot projects, e.g., on exhaustiveness of GDP estimates and the assessment of sources and methods for GDP compilation. The assessment is strictly orientated towards the ESA 95 rules and requirements of member states in the framework of the GNP Committee. Eurostat will carry out the second phase of the pilot project on exhaustiveness in 2003, under PHARE 2000.

0.3.3 Processes are in place to deal with quality considerations, including trade-offs within quality, and to guide planning for existing and emerging needs

The NSI holds regular meetings with users that consider the trade-offs between timeliness and quality of the national accounts. There are also surveys designed to get a wider perspective of user needs. The NPSS process considers the balance between supplying statistics as early as possible and ensuring that the disseminated data are of good quality. There is continuing pressure from Eurostat to improve on the timeliness of data. However, moves in this direction are made only if they can be achieved by more efficient procedures rather than by reduced quality of the statistics.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

There is no external interference in the compilation and dissemination of NSI statistics. The 1999 Law on Statistics establishes the autonomy of the NSI. It requires that “statistical activity shall be carried out in compliance with the principles of the rule of law, impartiality, reliability, efficiency, adequacy, publicity, and confidentiality.” It also established the role of the National Statistical Council in drafting the NPSS, which, along with the rules for dissemination of statistical information products and services, ensures that objective principles are applied in the collection, processing, and dissemination of official statistics. The NSI strives to adhere to objective international methodological recommendations in the preparation of its economic statistics.

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

The NSI makes choices about sources of data and statistical techniques solely on the basis of professional statistical considerations and in the light of international standards. Data sources are soundly based on household and business surveys. There is no evidence of outside interference.

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

Key data are presented and explained at regular monthly press conferences. One of the objectives is to prevent any misunderstandings or misinterpretations by users. In all cases of incorrect interpretation of data, the NSI responds publicly to provide the correct interpretation.

1.2 Transparency

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

Users have full access to information about the conditions under which statistical data are collected, processed, and disseminated, as well as about release timetables, confidentiality of individual data, and other specific points. This information is contained in basic documents such as the Law on Statistics, the rules for dissemination of statistical information, ARCs, and the catalog of NSI statistical publications. All of these products are reproduced both on the NSI website and in its printed and electronic publications.

The fundamental principles of official statistics, methodologies, classifications, nomenclatures, mission statements, etc. are regularly released in the NSI publication Statistics Magazine, in both printed and electronic form, as well as on the website.

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

Under new dissemination rules, national accounts data are provided to relevant government ministers (under embargo conditions) for briefing purposes up to one hour prior to the embargo time. This new rule is contained only on the NSI website, in Bulgarian. It is not yet reflected on the IMF GDDS metadata. The NSI is taking immediate action to present the new rule in English on the website and to update the GDDS metadata.

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

The NSI is clearly identified on all its printed and electronic products through the standard inclusion of the NSI logo together with a box containing publishing information and other insignia. In the case of joint publications with another institution, inclusion of the organization’s name and logo is obligatory, and the source of the information is stated in the methodological notes and tables. A basic requirement when giving permission for NSI statistics to be reproduced is that the NSI must be acknowledged as the source of the information.

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

The NSI informs users about the current status, and all significant changes, in relation to methodology, sources, and data collection techniques for all the surveys in the NPSS. In-depth information is provided to the academic community and practicing statisticians through the institute’s Statistics Magazine. Similar information is also presented on the NSI website.

1.3 Ethical standards

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

In addition to itemizing specific responsibilities, the job descriptions for each NSI staff member contain a list of the laws with which he/she needs to be familiar, i.e., the Law on Statistics, the Law for the Civil Servant, the Law for the Administration, and the Code of Behavior of the Civil Servant. While no special training course is provided to new staff on these laws, before they sign their employment contracts they are required to sign a declaration not to release individual data that they access during the course of their everyday work.

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

From 1991 onwards, the national accounts for Bulgaria are broadly compiled according to the recommendations of the ESA 95, which is consistent with the System of National Accounts 1993 (1993 SNA). The data for the years 1980–90 were calculated using the conversion tables, known as a document F-20 developed by UNSO, for adjusting the main macroeconomic aggregates of the Material Product System to those of the System of National Accounts 1968.

2.2 Scope

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

Most of the 1993 SNA (ESA 95) tables and accounts that the ISWGNA2 determined as the minimum requirement for implementation are compiled on a regular basis. The only exception is sectoral financial accounts, which are under development. The tables and accounts that are currently produced are:

  • annual value added and GDP at current and constant prices by activity;

  • annual expenditures of GDP at current and constant prices;

  • annual value added components at current prices by activity;

  • sequence of accounts for the total economy, and for each sector (up to the capital account) with an annual frequency; and

  • annual rest of world accounts (up to net lending).

All of the 1993 SNA (ESA 95) tables and accounts that the ISWGNA determined as recommended for its implementation are compiled on a regular basis. These are:

  • quarterly value added and GDP at current and constant prices by activity;

  • quarterly expenditures of GDP at current and constant prices; and

  • annual supply and use tables.

On a regular basis, Bulgaria also elaborates regional GDP conforming to the NUTS23 and NUTS3 levels.

The national accounts in principle cover the economic activity of all Bulgarian resident units, in conformity with the ESA 95. However, residency is established after six months, rather than the one year specified in the ESA 95. Nonetheless, this variance is not considered to have a significant impact on the figures. The accounts cover the whole territory of Bulgaria, including embassies in other countries and free-trade zones.

The production boundary covers most economic activities of resident units that produce goods and services in line with the ESA 95. So, in particular own-account, research and development, plus fixed capital formation on own-account, are included. Also covered are the own-account production of goods and housing services by household producers.

However, estimates for illegal activities are not currently included, although a special study to derive figures for the narcotics trade was undertaken for 1997.

Tangible assets include productive agricultural goods, buildings, machinery and equipment, transport vehicles, and other instruments and equipment. However, all defense-related assets are currently treated as assets, even those of a military nature that should be classified as current expenditure. Fortunately, this does not affect the size of total GDP since the correct treatment would result only in a reduction to gross fixed capital formation (GFCF) and a counterbalancing increase in general government final consumption. Valuables are not currently treated as assets, though the amounts involved are likely to be small. However, within the framework of the Eurostat pilot project on GFCF some investigative work on valuables will be undertaken. This will involve (1) the examination of existing practices, including trying to ascertain the kind of information currently being collected, or (2) development of new estimates. A key issue, which remains to be resolved, is how, in practice, valuables can be distinguished from nonvaluables, for example using some threshold. It should be noted that this issue is also unlikely to affect total GDP, because valuables purchased by households will be included in household final consumption.

Intangible assets cover owners’ rights (patents, licenses; know-how, trademarks); research and development; concessions; mineral exploration; computer software; and other intangible assets. The films, sound recordings, manuscripts and tapes containing drama and musical performances, radio and TV programs, literary and artistic output, etc. are regarded as entertainment, literary, or artistic originals. Where originals are produced and sold, the price paid is correctly scored as GFCF. If the original is retained by the producer, the valuation, which strictly depends on the expected benefits to the producer, is usually based on the costs of production. The quarterly figures for work-in-progress do not cover agriculture even though this is recommended in the ESA 95. However, this is consistent with the recommendations of many national accounts experts because of the difficulty of deriving these figures, particularly prior to the harvest in the current year. Even so, this issue will be addressed by a Eurostat pilot project.

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 used in the compilation of the Bulgarian national accounts are in broad conformity with internationally recommended systems. The ESA 95 is followed in classifying institutional units and transactions.

Since 1996 the production and generation of income accounts by activities are compiled according to the National Classification of Economic Activities, which is consistent with the Standard Classification of Economic Activities within the European Communities, Revision 1 (NACE, Rev. 1). Beginning in 2001, NACE, Rev. 1, Classification of Products by Activity and the List of Products of the European Communities (PRODCOM) are applied directly as national classifications. The Classification of Individual Consumption by Purpose (COICOP) is used for household expenditures, the consumer price index (CPI), and the retail trade survey. The Classification of Functions of Government (COFOG) has not yet been implemented but is planned for the medium term. However, the existing classification of the functions of general government, used by the MOF, is not significantly different from the COFOG. In particular, the existing general government data provide the information necessary for the individual/collective split of consumption expenditures. Foreign trade data are classified according to the 9-digit level of the National Custom Tariff based on the Combined Nomenclature. Capital formation is broken down into six categories according to the Pi6 level defined in the ESA 95.

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 ESA 95. Market output is valued at basic prices before any taxes on products (import duties, excises, VAT, and the tax on insurance premiums) are added and before any subsidies on products are deducted. Changes in inventories of finished goods and changes in inventories of work-in-progress should, in principle, be valued at the prices prevailing at the time they were produced. In practice, they are valued at the prices of the reporting time, i.e. the beginning and the end of the period in question. This is why additional adjustments for holding gains are undertaken. Intermediate consumption is valued at purchaser’s prices and is also adjusted for holding gains. Gross output for nonmarket producers is derived from the production costs incurred in providing the services, by summing up the components of intermediate consumption, compensation of employees, and consumption of fixed capital.

Final consumption of households, general government and nonprofit institutions serving households (NPISHs) is valued at purchaser prices. GFCF is also valued at purchaser prices (including all costs associated with ownership transfer), if they are bought. However, and if they are produced for own consumption, the figures are recorded at basic prices, although in this case there is no difference from purchaser prices. Owner-occupied housing is valued at comparable rental values, and imports and exports are valued free on board.

Household production for own final consumption covers agricultural goods produced and consumed by those households and is valued at the relevant selling prices. Nonagricultural and nonfood industrial products produced by households for own use such as clothing, furniture, etc. are not included in gross output, in line with the recommendations of the ESA 95.

2.4.2 Recording is done on an accrual basis

In principle, accrual accounting is followed for all accounts. According to the accounting law and standards, all units are obliged to record their transactions on an accrual basis. However, figures for general government are still recorded on a cash basis. The NAD adjusts the figures for taxes on production onto an accruals basis using data from the tax authorities. Similar adjustments are made to the figures for subsidies on production using data from the financial reports of the receiving enterprises. The MOF is planning to produce budget data on an accruals basis by the year 2005.

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

The production figures are based on reported data from establishments, and their transactions with other 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 collected from comprehensive data collection programs that take into account country-specific conditions
Registers

The Unified Register of Economic Units in Bulgaria (BULSTAT) is used as a common sample frame for all enterprise surveys. This register is organized in a decentralized way in the 28 Regional Statistical Offices. It covers all legal entities including foreign-owned ones, natural persons registered as sole proprietors, partnerships, budgetary units, and NPISHs. The Register Department of the NSI is responsible for maintaining the BULSTAT. According to the Statistics Law, all units have to register with the BULSTAT. As identification characteristics, the register contains information about the main activity according to the 4-digit level of NACE, Rev. 1, name, address, and legal status. The BULSTAT contains three levels of units—the enterprise, any establishments with complete bookkeeping, and any local units (but only wages and salaries are observed for local units). It does not cover professionals such as lawyers (who are exempted by a special law), street and market traders, and temporary and occasional workers.

A Statistical Register is derived from the BULSTAT, to keep track of the currently registered units for statistical purposes. This is so that dead units can be removed from the register. In addition to the identification characteristics, the Statistical Register contains information about the main activity of the unit, the number of employees, turnover, and long-term tangible assets.

All units are included in the register when they are newly created and removed when they report their closure. New units have to register because no legal economic activity is possible without a registration number. There is no formal mechanism for reporting dead units. However, larger businesses that have double-entry bookkeeping are required to provide quarterly statistical reports to the NSI. So, if a particular unit fails to supply such a report, the appropriate Regional Statistical Office checks to see if it is still in existence. Small units (unincorporated enterprises with single entry bookkeeping) are checked by the tax authorities, not by the Regional Statistical Offices.

Special procedures exist for registering legal entities controlled by the Ministry of Defense, the Ministry of the Interior and other similar institutions. Professionals (advocates, doctors, dentists, teachers, and artists) have to apply for a tax number and are registered in a special tax register. For all agricultural producers, including private households, a special register is built-up from the Ministry of Agriculture and Forestry Register, the Social Insurance Register, and the last census on population, housing funds, and agricultural farms.

Monthly checks are made with other registers, such as the Social Insurance Register and the VAT Register. The resulting register is considered to be very sound. This was confirmed by a one-off street by street census of all businesses, however small, undertaken by the regional offices in the year 2000.

Statistical surveys

Statistical units (enterprises and establishments) are classified to the appropriate industry according to their principal activity, even if they have any secondary activity. The statistical surveys are conducted using the Statistical Register. Statistical reports are collected from the whole enterprise as well as from any establishments. Specific questionnaires are designed for different activities and also for units having double or single entry bookkeeping.

All units (enterprises and establishments) are obliged to respond to the surveys under the Law on Statistics. Public enterprises generally have a 100 percent response rate, but it is less for private firms and varies by type of activity. The Basic Statistics Divisions make estimates for the nonresponse based on other statistical surveys.

The main statistical surveys conducted by the NSI and used for national accounts purposes are as follows:

  • annual bookkeeping statements, provided by all public and private enterprises with double-entry bookkeeping, with special annexes for revenues and expenditures; acquisition of tangible fixed assets; revenues from sales of construction activity; reports on retail and wholesale trade outlets; etc.;

  • less detailed annual statistical reports on revenues and expenditures provided by private firms with single-entry bookkeeping;

  • annual reports of insurance enterprises, collected by the NSI and based on the specific accounting standard of insurance undertakings; additional information is provided by the yearbook on insurance businesses compiled by the Insurance Supervision Directorate in MOF;

  • annual reports of nonprofit institutions based on their specific accounting standard, which allows output to be split between market and nonmarket activities;

  • annual reports of investment and pension funds, which are separate institutional units;

  • the PRODCOM survey that is exhaustive annually, conducted for 10,000 enterprises engaged in mining, manufacturing, electricity, gas, and water supply activities. Information is collected for about 3,000 different industrial products. Monthly, a sample survey is carried out covering a quarter of the universe and observing 950 products. According to the requirement of the PRODCOM Regulation, the deadline for publication of the data is six months, but at present they are available 10 months after the end of the reference year. Monthly data are published within 40 days;

  • agricultural balances, containing very detailed information about the supply of goods and their use in both quantity and value terms—compiled by kind of production unit for state and co-operative agricultural farms, private farms, and households;

  • annual reports on labor statistics, providing information about the average annual number of employees, wages and salaries, and working hours—split by economic activity and type of ownership for double-entry bookkeeping units;

  • the Labor Force Survey (LFS), used to estimate hidden and self-employed activities. The methodology of the survey has been developed according to the International Labor Organization recommendations and usual international practice. The survey covers a sample of 24,000 noninstitutional households in the whole territory of the country. The reference period is one calendar week (the week before the survey). The data are collected from a two-stage cluster sample of the population aged 15 years and over;

  • material balances in quantity and value terms for about 200 commodities groups, constituting intermediate consumption of enterprises and establishments; and

  • the Household Budget Survey (HBS), which is a continuous survey covering all households in the country. The survey uses a two-stage scientific random sampling technique: first, selection of population census areas, and second, selection of households. Institutional households are not included in the sample. All households participate in the survey for one year. At present the sample covers 3,000 households. All households are included in the survey on a voluntary basis. The HBS sample is self-weighting, i.e. all households have an identical probability of selection. Households that refuse to participate are substituted by households with similar characteristics. A special investigation was undertaken to confirm that this procedure did not introduce any bias into the sample. The HBS collects information on monetary expenditure on food products, nonfoods, services, income from wages and salaries, social security benefits, households own production, etc. Since 1999, households’ expenditures are classified according to COICOP.

NSI also carries out sample statistical surveys. A very important use of these surveys is to control and correct the data from the exhaustive statistical surveys and to cross-check the information from the statistical reports. Furthermore, they allow additional verification of the estimates for the private sector.

The following table summarizes for each NACE, Rev. 1 section the main basic sources for the annual accounts estimates and an assessment of their coverage. Taken over the total of all activities, the coverage of the main sources should be considered to be good.

Main sources for the annual accounts

article image
article image
article image

Other sources of information used for quarterly national accounts are a:

  • special statistical survey, “Financial Results,” which is exhaustive for all enterprises in the public sector with representative samples for private units in all industries other than agriculture;

  • statistical survey of insurance undertakings;

  • labor statistics report;

  • retail trade survey; and

  • report on investment of enterprises.

The available price information—producer price index (PPI), CPI, unit value indices (UVIs) etc.—is adequate for compiling the quarterly and annual accounts.

Main sources for the quarterly

article image
Administrative data

Monetary and banking statistics are provided by the BNB. This covers annual and quarterly balance sheets and income statements of the BNB and the commercial banks. Additionally, annual annexes are supplied containing more detailed breakdowns of some indicators.

BOP data are produced by the BNB in accordance with the fifth edition of the Balance of Payments Manual (BPM5).

The consolidated budget is supplied quarterly and annually by the MOF, broken down by function and for each subsector of general government. Annually, the execution of the state budget is very detailed—by separate units, activities and revenues, and expenditure items.

Quarterly customs statistics on imports and exports of goods are used for checking the direct exports of enterprises and the external turnover of trade firms.

Income declarations collected by MOF are the main source for output produced by households not recorded in the business register, lawyers and other professionals, street traders, etc.

Tax authorities’ information is used to convert VAT and excises data onto an accruals basis using information on the delay between when the taxes fall due and when they are paid to the government.

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

Source data closely conform to the definitions, scope, classifications, valuation, and time of recording required for the national accounts. The NSI is consulted on any changes to the accounting standards, which means that financial reports meet the needs of the national accounts.

3.1.3 Source data are timely

The data collection system within the NSI is strongly controlled and provides timely data for the elaboration of the national accounts. The Basic Statistics Divisions deliver their quarterly information between 45–55 days after the end of the reference quarter. This is with the exception of the fourth quarter of every year, when the data become available within 80 days because of the extra processing required. Administrative information from the MOF, BNB, etc. is delivered about 60 days after the end of each quarter. The elaboration of the quarterly GDP figures also benefits from the fully computerized system created by the NAD. This meant that in 2001 the deadlines for the national accounts for the first three quarters of the year were reduced to 80 days after the end of the relevant quarter. This is closer to the Eurostat requirement of 70 days. The estimates for the fourth quarter, and hence the first annual totals, only become available 115 days after the end of the year.

More comprehensive annual figures become available 18 months after the accounting period, based on annual information from statistical surveys, bookkeeping records, and administrative data sources. The final estimates based on balanced supply and use tables are produced 36 months after the year in question.

3.2 Statistical techniques

3.2.1 Data compilation employs sound statistical techniques
Production approach procedures

The industrial breakdown of the annual GDP estimates is based on the 3-digit level of NACE, Rev. 1. Intermediate consumption for each 3-digit industry is further broken down into 3-digit products using data from the latest supply and use tables. This is the only use of historical ratios, and these are only three years old.

The calculation of agricultural gross output is supplemented by figures for household primary and secondary production and consumption. These are based on estimates of areas under cultivation and livestock estimates produced by region. Intermediate consumption includes estimates of the value of seeds, fertilizer, livestock feed, power, and repairs. Output used for own intermediate consumption is excluded from both output and intermediate consumption.

For other economic activities, various methods are used to supplement the standard reports and surveys to arrive at estimates for total output and intermediate consumption. Estimates are prepared at the 3-digit level of the NACE, Rev. 1 for nonresponse; in most cases, average values for similar enterprises are calculated and used for the estimation for nonresponding enterprises. Government budget data cover all activities that should be classified as government services. Estimates for holding gains are made by activity branches and for raw materials, work-in-progress, finished goods, and goods for resale.

Own-account construction and production of major equipment are not recorded as work-in-progress, but only when completed.

The imputed rent of owners’ occupied dwellings is evaluated on the basis of number of dwellings and market rent. A lot of improvements in the data collected for actual rents have been implemented regarding the stratification and details of rented dwelling stock. However, there are still concerns about the quality of the estimates, because the rental market of dwellings is less than 10 percent of the total dwelling stock. So, the NSI is in process of implementing an alternative method—the user cost approach recommended by Eurostat.

Consumption of fixed capital is not estimated by the recommended perpetual inventory method, due to the lack of long time series on asset data. So, the estimates are derived from depreciation figures reported in the financial reports.

Volume measures

For constant price estimates, the recommendations of the ESA 95 are followed. For measuring gross value added of most market producers, the double deflation method is used, that is, output and intermediate consumption are deflated separately. However, some use is made of volume projection, when appropriate. In general, Paasche type deflators are used to estimate the production account components. The constant price figures are expressed as annual chained indices.

The output of the industrial branches is deflated at the 2-digit level of NACE, Rev. 1. The deflator presents a combination of domestic PPIs and export UVIs. The Business Statistics Directorate provides information on domestic market PPIs at the 3-digit level. They are Laspeyres type with fixed 1995 base year. NAD recalculates them to derive price changes over the average of the previous year. From the monthly surveys, information on the domestic sales and direct exports is available. They are used to weight the PPIs and the export UVIs provided by the external trade statistics division and to obtain implicit deflators of gross output at the 3-digit level. These indices are weighted to the 2-digit level by applying current period sales as weights in order to approximate a Paasche deflator. So, in effect the volume estimates are being derived at the 3-digit level of NACE, Rev. 1.

The latest supply and use tables are used to derive a structure of all intermediate inputs (domestically produced and imported) by branch. The resulting structure is used to weigh the corresponding PPIs and import UVIs to obtain an aggregate index for each industry. Agricultural output at constant prices is estimated using direct valuation of quantities in the previous year prices; intermediate consumption is derived by deflation.

The relevant business statistics divisions supply the deflators for trade, transport, and communications. The market output of the other services is calculated at constant prices using the corresponding CPIs.

The trade margin (output) at present is calculated at constant prices using double deflation. Combined CPI, PPI, and UVI indices are used as deflators for purchases by the wholesale and retail trade, adjusted for change in inventories. In the framework of the Eurostat pilot project on constant prices, growth in output derived from a detailed deflation of sales will be applied.

The output of nonmarket services and construction at constant prices is the sum of the deflated value added and intermediate consumption. The wages and salaries index for each industry is used for value added. In the case of intermediate consumption, a suitable price index is applied to each of its components.

The implicit deflator for domestic demand is used to deflate financial intermediation services indirectly measured (FISIM). Additional work will be undertaken in cooperation with the PHARE FISIM project to derive estimates based on the projection of the deflated values of the relevant loans and deposits.

The volume measures of net taxes on products are estimated by applying the base year tax rates to the volume of the transactions that are subject to each tax or subsidy.

Expenditure approach procedures

The estimate of GDP by final expenditure is compiled independently from the production data. The estimates are in accordance with the domestic concept and in compliance with the ESA 95, and these principals have been applied to the time series back to 1991. However, the national accounts follow the more conceptually correct national concept from 1996. This introduces a break in the figures, which is necessary to maintain full consistency with the BOP figures.

In some cases the estimated values of some specific transactions from the production side are applied directly to the related categories of final demand—imputed rent of owner occupied dwellings, production of households for own final use, and own account construction of residential buildings. Some of the data from the demand side are equally important in determining the adjustments for exhaustiveness applied to the production figures—for example the HBS information about expenditure on food is the basis for adjustments to the data for the relevant manufacturing industries and retailers. In some cases the custom statistics for imports and exports of goods are used to adjust the production figures obtained from business statistics.

Household final consumption expenditure

Since 1997, household final consumption expenditure is based on the 4-digit level of the COICOP. The NAD uses various sources to build up the estimates of household final consumption expenditure. Balance of payments data are used to make adjustments for expenditure by tourists as required by the national concept, which has been followed since 1996. The HBS is used for food products and retail trade statistics for nonfood products. This guarantees an independent estimate of consumption. Branch statistics data are used for other items such as services, including consumption of water, electricity, and heat energy. Retail trade data have also been presented in accordance with COICOP since the year 2000. It is important to mention that different sources can be chosen for different COICOP items. The NAD has more than one estimate for most items, which are used for comparison purposes. COICOP classes are also adjusted for the purchases of businesses, income in kind, and household production for own final consumption.

NPISH final consumption expenditure

The final consumption expenditure of NPISHs is derived from the corresponding gross output figures, adjusted for incidental sales.

Government final consumption expenditure

The government final consumption expenditure is calculated by subtracting the value of sales, which are included in the intermediate consumption or in household final expenditure, from the gross output of general government. The distinction between final and intermediate consumption is based on detailed information from the budget. COFOG has not yet been introduced. However, the activity detail of the National Classification of the Functions of the State Budget provides the information necessary for the individual/collective split of consumption expenditure.

Gross fixed capital formation

GFCF is measured as expenditure on the acquisition of fixed assets (the actual expenditure on construction activity, improvements of existing assets, cost of delivery and installation of fixed assets, and progress payments by customers), less the disposal of existing fixed assets.

For tangible fixed assets, data are obtained through the annual report on investments and cover all units outside the household sector. Until 1998, the data included expenditure on the acquisition of secondhand assets, so adjustments had to be made to exclude them. The commodity flow approach is used for cross-checking the consistency of the GFCF data for construction, machinery, and equipment. Data on sales to the domestic market of those branches producing investment goods and information from customs statistics on imports of investment goods are also used.

Changes in livestock used for production are derived from the agricultural balances. Expenditure on acquisition of assets by households is estimated on the basis of information on the number of permissions for individual construction of dwellings.

Estimates for intangible fixed assets are based on the specific annex to the annual balance sheet.

Changes in inventories

Measurement of change in inventories is in line with the ESA 95 recommendations. It is based mainly on quarterly information.

The estimate for the enterprises sector is based on the quarterly survey “Financial Results” and the annual balance sheets. Data are available in value terms for the stocks at the beginning and at the end of the period. Changes in inventories for enterprises are derived as the difference between the inventory levels at the beginning and end of each quarter. They are separately estimated by stage of processing for all industries and adjusted for holding gains. Current and constant price data are derived simultaneously.

HBS data and agricultural balances are used to estimate the change in inventories for unincorporated activities of households and agricultural producers. As the data for agriculture are available in volume terms, it is not necessary to make adjustments for holding gains.

Imports and exports of goods and services

The exports and imports of goods are valued free on board and are calculated on the basis of customs declarations. Those data are provided by the External Trade Statistics Division of the NSI at the 8-digit level of the Combined Nomenclature, which has a transition key to the national classification of products and NACE, Rev. 1. The same figures are included in the BOP.

The data on exports and imports of services cover construction, transport, communication, financial and insurance services, operating leasing, culture, science, sports, maintenance of embassies, etc. This breakdown is provided by the BOP. Until 2000 it was only prepared in U.S. dollars and was converted into Bulgarian leva by the NAD. Since 2001, the BOP also provides the figures in leva, based on actual transaction exchange rates.

Volume measures

The household final consumption expenditure estimates at constant prices are derived by applying the relevant CPIs for the corresponding COICOP items at the 3-digit level. The CPIs used are chain-linked, Paasche type indices covering the whole territory of the country.

The final consumption expenditures of general government and NPISHs are deflated with the implicit output indices of the corresponding branches.

GFCF is obtained at constant prices by applying appropriate Paasche price indices—the implicit construction output deflator, domestic PPIs, and import UVIs for investment goods; and the implicit deflator of the output of agricultural productive goods.

Depending on the average stock-holding period the value of inventories by type and by branch is estimated at average previous year prices using specific monthly or quarterly price indices. Raw materials are deflated with a combination of PPIs and import UVIs. The latest supply and use tables are used to derive a structure of all materials that can be held in stock (including imports) by industry. Work-in-progress and finished goods are deflated with indices derived from the relevant PPIs, weighted from the 3-digit to 2-digit level with corresponding current period weights. For the electricity and service branches, the overall price index for manufacturing industry is used, because their stocks will be the result of secondary activities. Goods for resale are deflated with monthly and quarterly PPIs for manufactures.

The exports and imports of goods are deflated with the corresponding Paasche type UVIs. Since 2001, the Foreign Trade Statistics Division has improved the quality of these indices. The price index is calculated from the average prices of selected commodities for each 3-digit NACE, Rev. 1 class. The import of services is deflated with the harmonized CPI of EU-15. Exports of services are derived at constant prices by applying the corresponding national CPI.

Specific quarterly compilation techniques

Quarterly accounts for Bulgaria have been prepared since 1994. GDP is calculated from the production and expenditure approaches, and at current and constant prices. The constant price figures are expressed as time-series of chain-linked indices. The estimates are based largely on direct surveys and apply the same methodology as the annual national accounts.

The quarterly estimates of the production account follow the structure of the annual versions. However, as is common in other countries, the level of detail is less than for the annual accounts. Also, work-in-progress is not estimated quarterly for agriculture, because no usable data are available. Adjustments are made for holding gains; exhaustiveness—people working without labor contract, based on the LFS; income in kind; and underreporting.

Seasonally adjusted GDP is calculated on an experimental basis and is published, but only as a graph. The model used is Tramo/Seats (Demetra interface proposed by Eurostat). The data are available from 1995, but only for total GDP. The calculation of seasonally adjusted output and expenditure components is planned.

Three sets of national accounts are compiled for each quarter:

  • 1. First quarterly accounts estimates published 80 days after the corresponding period are compiled on the basis of the quarterly basic statistical information available. The GDP estimates from the production and expenditure side are balanced on the basis of the use matrix framework. However, the statistical discrepancies between the two methods are not entirely removed.

  • 2. The preliminary annual estimates can involve revisions to the estimates for the first three quarters.

  • 3. Finalized quarterly estimates can be reconciled with the final annual accounts. Reconciliation with the annual figures makes use of the recommended Denton technique.

In practice, there are two series of finalized estimates for each year—the first set is reconciled with the final annual figures, but before finalizing the supply and use tables. The second set is achieved when the last revision of the annual accounts has taken place, after constructing the supply and use tables for the corresponding year. Unlike the preliminary data, with final revised estimates, reconciliation is done in line with the balancing procedure of quarterly aggregates, and all statistical discrepancies are removed.

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

A large number of procedures are applied in national statistical practice to adjust the source data and to improve the coverage, definitions, and valuation in conformity to international guidelines. The estimates from the production and the expenditure approaches are balanced through a supply and use framework.

The NAD has participated in the Eurostat pilot project on exhaustiveness, which aims to ensure the exhaustive coverage of the production system as defined in the ESA 95. Specific adjustments are made for informal and underground activities. Some investigation of illegal activities has also been undertaken, such as the transactions involved in drug production, distribution, and consumption. However, these estimates have not yet been incorporated into the accounts, owing to concerns about their quality.

According to the definitions of the European Commission Decision on exhaustiveness, concerning the estimation of the legal activities that are not directly observed in statistical surveys, the NSI effort has, until now, been mainly focused on the following areas:

  • survey nonresponse;

  • evasion of the payment of taxes and/or social security; and

  • exemption from the obligation to submit information to the tax and social security authorities.

For nonresponse, adjustments are prepared at the 3-digit level of NACE, Rev. 1. The number of nonresponded units, their activities, and number of employees are determined. Then the average revenue from sales per employee is calculated from the data of responded units in the same activities with similar number of employees. The derived ratio is used to estimate the activity of nonresponded units.

The adjustments for underreporting to avoid tax and social payments are made using indirect methods, which can be summarized as follows:

  • The ratio between intermediate consumption by types of expenditures and gross output for different groups of units engaged in the same activity is calculated. The groups are defined according to the type of activity, ownership, and number of employees at the 3-digit level of NACE, Rev. 1. The investigation of the deviation in input-output ratios among the units within the chosen homogeneous group shows the degree of overreporting in intermediate consumption and underreporting in the value of gross output.

  • Additional adjustment is made on the basis of comparison of structural indicators in the public and private sectors, and between small and large enterprises. Typically for all small units the ratio between intermediate consumption and gross output is extremely distorted. It can be twice or three times larger than for public sector units or larger private enterprises in the same branch. For many small units, both outputs and inputs are unreliable. Also, for some businesses the share of externally supplied services can be several times higher than for other units belonging to the same activity group. This is because they are recording the expenses of their own staff as external services to avoid social security payments.

  • The evaluation of the retail trade sales in some specific economic activities—using the demand side data (HBS)—identifies some underestimation of sales. For example, the HBS shows higher expenditures on food than is recorded in retail sales.

The main exhaustiveness correction made with regard to missing income is based on LFS data. In order to avoid tax and social contribution payments, employers generally tend to employ some workers without any labor contract. According to the result of LFS, their number is about 100,000, with an increase in numbers in June and September, owing to the seasonal character of some activities.

3.3 Assessment and validation of source data

3.3.1 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 planning

Information about sampling errors is available. The NAD routinely reviews the nonresponse adjustments made by Basic Statistics Divisions. Information is available about imputation methods for nonresponse in the surveys. High-value transactions are confirmed with respondents.

The accuracy of budgetary data, international trade, price statistics, and other secondary sources used to compile national accounts is routinely assessed by national accounts experts.

Appropriate measures are taken to validate the source data and to make them consistent with the national accounts. The source data are analyzed to correct for underreporting/misreporting, in particular to check for temporal consistency and consistency with other related data sources. The consistency of the source data with the national accounts is checked, in particular concerning the definitions, valuation, reference periods, and classifications. Source data are analyzed to ensure that any revisions are reasonable.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Main intermediate data are validated against other information where applicable

The data compiled from the main sources used in the national accounts statistics are checked with other primary/secondary sources whenever possible.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated

A systematic procedure exists to routinely assess the potential discrepancies in intermediate data. Appropriate adjustments are made to remove any discrepancies.

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

Supply and use tables are used to investigate discrepancies and make the statistical outputs consistent. Data are validated against other estimates whenever possible. In particular, the figures for informal activity are checked against those produced by other organizations. In fact, some of the methods used by these organizations are taken up by the NAD.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used to inform statistical processes

The national accounts data are subject to routine revisions before being considered as final. The process of revisions results in the elaboration of different quantitative evaluations of one and the same variable. The direction and magnitude of revisions between preliminary and final data are investigated. Short explanations about the reasons and type of revisions accompany the release of national accounts.

The results of investigations and adjustments made in the statistical processes are taken into account in compiling the data for the subsequent periods.

4. Serviceability

4.1 Relevance

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

Users are represented on a number of groups responsible for establishing and maintaining the statistical program, e.g., the High Statistical Council. The NAD also arranges periodic meetings with groups of users to monitor their needs. Marketing surveys to identify user needs are also undertaken on a regular basis.

National accounts experts participate in international statistical meetings organized by Eurostat, OECD, UNECE, IMF, and other international institutions. Representatives of the NSI take part in the interinstitutional working groups established for ad hoc and accession related purposes.

4.2 Timeliness and periodicity

4.2.1 Timeliness follows dissemination standards

The GDP estimates for the first three quarters of the year are disseminated within 80 days after the end of the relevant quarter, exceeding the SDDS requirement of 90 days. However, the figures for the fourth quarter do not become available until 115 days after the end of the quarter, which does not meet either GDDS or SDDS requirements.

The first annual estimates of GDP and its subcategories become available 115 days after the end of the year, calculated as a sum of the preliminary quarterly accounts for the year. Revised estimates are published 18 months after the accounting period, based on annual information from statistical surveys, bookkeeping records, and administrative data sources.

The final estimates become available, after compilation of supply and use tables, 36 months after the end of the year in question.

4.2.2 Periodicity follows dissemination standards

The Bulgarian national accounts estimates are disseminated with quarterly periodicity as prescribed under the SDDS.

4.3 Consistency

4.3.1 Statistics are consistent within the dataset

The annual national accounts are internally consistent. The GDP estimates by activity and expenditure components are made consistent in a detailed balancing process on the basis of the supply and use framework. The estimates at current prices are consistent with the matching series at constant prices and the deflators. The quarterly GDP estimates add up to the annual estimates. Concepts, definition, and classification for compiling quarterly GDP estimates are the same as those used to compile annual estimates.

The only exception to the above is the quarterly GDP by the output and expenditure approaches and, consequently, the preliminary annual data based on them. These data are not completely balanced, and the officially published GDP estimates for these periods are based on the production approach. Then an additional residual item is included within the table for expenditure figures. This statistical discrepancy between the estimates is not large and is stable over time—less than 1 percent.

4.3.2 Statistics are consistent or reconcilable over a reasonable period of time

Consistent national accounts time series according to the ESA 95 are available from 1991 onward. Detailed methodological notes identifying and explaining the introduction of the ESA 95 were published in the NSI/OECD joint publication “National Accounts for Bulgaria–Sources and Methods.” Any breaks in the series are explained to the users—for example, changes in the classification of economic activities in 1996.

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

The national accounts statistics are in general consistent with balance of payments and government finance statistics. Currently, the BOP is the main source of statistical data for the transactions with the rest of the world, and its current account is used as a “mirror” account in the Bulgarian national accounts. Also, the net lending/borrowing of the country is reconciled with the BOP current external balance. Regarding the quarterly BOP data, the consistency between the two sets of macroeconomic statistics only exists for a short time as the BNB makes monthly revisions to its data. The net lending/net borrowing of the general government subsectors recorded in the nonfinancial national accounts is identical to the government debt and deficit, as presented by the MOF since 1998 (according to ESA 95 and Council Regulation 3605/93 of EC). However, the other transactions for general government have to be adjusted to conform to the requirements of the ESA 95.

4.4 Revision policy and practice

4.4.1 Revisions follow a regular, well-established, and transparent schedule

The first annual estimate of GDP and its subcategories is derived from preliminary quarterly accounts for the current year and become available 115 days after the end of the year. Then, independent annual national accounts estimates are prepared on the basis of more exhaustive annual information derived from statistical surveys, bookkeeping records, and administrative data sources. These new annual figures involve the first revision of the preliminary national accounts data and become available 18 months after the end of the year. The second revision of national accounts data is based on the compiled and balanced supply and use tables, regularly prepared three years after the period in question.

The first release of quarterly national accounts is within 80 days of the end of the reference quarter. The quarterly data are fully reconciled with each annual revision. In the process of elaboration of the accounts for the current quarter, some quantitative changes could result in revisions to the previous quarters of the current year. The purpose of these within-year revisions is to improve the consistency and comparability of the quarterly data. Such revisions also maintain consistency between the quarterly and annual accounts.

As well as these regular revisions, there are occasional one-off changes to the data. These occur as a result of changes in surveys, changes of estimation procedures, incorporation of information from nonannual statistics such as censuses, etc., or as result of changes in concepts, definitions, or classifications. One major cause of such revisions is continuing work on harmonization of the Bulgarian national accounts practice with the ESA 95. Users of national accounts data are given a clear view of these revisions through the press conferences and in the publications.

4.4.2 Preliminary data are clearly identified

Preliminary annual estimates are identified in statistical releases. As the quarterly data are subject to continuous revision, it would be redundant to identify preliminary data. However, the data published on the website have a clear indication of the exact date of their compilation.

4.4.3 Studies and analyses of revisions are made public

Users are informed of the causes of revisions to the statistical series. Revisions are explained in the publication of the statistical series.

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)

National accounts data are published clearly—charts and tables are disseminated with the figures to facilitate their analysis. Depending on the type of publication—yearbook or specialized national accounts publication—datasets are published with various levels of detail. Estimates are disseminated at a detailed level and as time series. An annual analysis of current-period developments is included in the NSI publication “Socioeconomic Development of Bulgaria.” The release of quarterly GDP estimates is supplemented with short comments and methodological notes.

5.1.2 Dissemination media and formats are adequate

GDP data are first published via a press release, which is then followed by the release of a more comprehensive publication. Recently released data and longer time series can be accessed through the NSI Internet site.

The national accounts data are available in the following hard copies:

  • 1. “Statistical Journal” (monthly edition), NSI (Bulgarian/English)—Quarterly GDP;

  • 2. “Main Macroeconomic Indicators,” NSI (Bulgarian/English);

  • 3. “Statistical Yearbook of the Republic of Bulgaria,” NSI (Bulgarian/English);

  • 4. “Statistical Reference Book,” NSI (Bulgarian); and

  • 5. “Statistical Reference Book of the Republic of Bulgaria,” NSI (English).

5.1.3 Statistics are released on a preannounced schedule

The statistical series, including the national accounts, are released according to a preannounced schedule. In 2001 a general assessment of the national statistical system took place in the NSI. In terms of the dissemination of statistical information, it was recommended to include in the calendar for presenting the results from the statistical surveys, conducted by the NSI, not only the week, but also the date for presenting the results. The calendar for 2002 was revised accordingly.

The calendar for presenting the results from the statistical surveys is announced after the adoption of the NPSS by the Council of Ministers (at the beginning of the current year). Therefore, the calendar for 2003 is not ready yet, which means that there is nothing on the website. In the interim period, the NSI supplies details of forthcoming events by fax to those users who have registered for such information.

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

The NSI provides data on an embargo basis to the media at a press conference. One hour after the beginning of the press conference, the data are available on the Internet.

5.1.5 Nonpublished (but nonconfidential) subaggregates are made available upon request

The NSI, upon the request and under signed bilateral contracts with some government agencies, provides more detailed data than is published. Such data can also be supplied to other users on request and, possibly, on payment of a fee. However, this facility is not publicized.

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—including information on concepts, definitions, classifications, data sources, and statistical techniques—are published on the Internet, but only for quarterly GDP and in Bulgarian. The national accounts chapters in the “Statistical Yearbook” and the “Statistical Reference Book” also contain metadata. The specialized national account publication “Main Macroeconomic Indicators” contains more detailed methodological notes in Bulgarian and English.

The GDDS summary methodology is regularly reviewed and updated.

NAD has been involved in a Eurostat activity to produce a detailed description of the sources and methods for the annual GDP estimates (similar to the work undertaken with the EU member states within the GNP committee), but this description is only used internally.

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

The description of the methodology included in each publication is designed to meet the needs of the users of that publication.

5.3 Assistance to users

5.3.1 Contact person for each subject field is publicized

Prompt and knowledgeable service and support is available to users of statistics. National accounts experts provide assistance and answer inquires of national and international users depending on their level of responsibility and competence. In line with NSI policy, publications provide a standard contact in the Information Division of the NSI, and only in Bulgarian.

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

A catalog of publications, documents, and other services to users is available and is updated each year. It is published in hardcopy and on the NSI website. The prices of the statistical products and services are clearly disclosed, and assistance is provided for placing orders.

Table 1.

Bulgaria: Data Quality Assessment Framework—Summary Presentation of National Accounts

(Compiling Agency: National Statistical Institute)

article image

II. Price Statistics (Consumer Price Index)

0. Prerequisites of quality

0.1 Legal and institutional environment

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

The Law on Statistics published in the Official Gazette No. 57 of June 25, 1999, amended in No. 42 of April 27, 2001, governs the activities of the National Statistical Institute (NSI) as well as other government statistical bodies. Chapter 3 of the Law on Statistics establishes the National Statistical Council (NSC), a body attached to the NSI president and responsible for the annual National Program for Statistical Surveys (NPSS). This program covers all significant statistical work undertaken by government agencies, including the consumer price index (CPI), and is approved by the Council of Ministers. The NSI disseminates data under Chapter 5 of this Law, which requires statistical bodies to inform the public about the results from statistical surveys through mass media, own publications, and other means.

The Bulgarian Annual Statistical Plan incorporates the NPSS of the NSI and other surveys that are of interinstitutional significance. For each specific statistical survey included in the NPSS, the following items must be stated: the purpose, contents, and coverage of the survey; the methods for conducting the survey, information sources, and forms of data collection; the body or bodies conducting the survey, obligations for data provision by the respondents, and the budget for the survey.

0.1.2 Data sharing and coordination among data producing agencies are adequate

In pursuance of the Law on Statistics the NSI is responsible for coordinating all national statistical activity. The main incentives to encourage coordination among agencies include the sharing of information, participation in joint statistical activities, access to statistical expertise, and influence on the setting of priorities. The tools to achieve coordination are as follows:

  • the NPSS statistical infrastructure—classifications, registers, definitions, indicators;

  • councils, which improve relations with users, data suppliers, and other bodies of statistics, that is the National Statistical Council, High Statistical Council, Methodological Council, Editing Council, Methodology and Training Council, and IT Council; and

  • agreements with bodies of statistics and other institutions, protocols, rules, methods, and ordinance.

The Law on Statistics specifies the obligations of legal entities to submit information in response to surveys, covered in the NPSS and conducted on an obligatory basis.

The NPSS covers statistical information required for observation and compilation of the Bulgarian economic, demographic, social, environmental, and other activities, taking into account the availability of sources, respondents burden, and cost-effectiveness principles. The NPSS includes a plan for the NSI statistical surveys and surveys with interinstitutional significance. The NPSS covers permanently conducted statistical surveys of established public significance, other statistical surveys for the particular year, and particular surveys and surveys related to the international obligations undertaken by the Republic of Bulgaria. Exhaustive censuses are conducted according to the specific laws enacted for that purpose. The NPSS contains detailed characteristics of the surveys in terms of:

  • theme of the survey;

  • contents and object of the survey;

  • method of conducting the survey;

  • sources and form of data collection;

  • body/bodies conducting the survey;

  • obligation for data provision; and

  • budget of the survey.

The funds needed for implementing the NPSS are provided from the State Budget. These funds are allocated to the bodies undertaking the surveys. There is no information available about the coordination of these funds, and the efficiency of their use, for monitoring. The NPSS is published in the Official Gazette.

Annual operating plans for statistical activity have been developed in conformity with NPSS. These plans contain detailed information about the allocation of tasks between different departments and divisions, including aspects of data exchange.

In accordance with the Law on Statistics there are two principal councils, i.e., the High Statistical Council and the National Statistical Council. The High Statistical Council, a consultative body to the NSI president, provides opinions and recommendation on the NSI plan for statistical surveys, drafts of legislative instruments in the field of statistics, national statistical standards, national classifications, nomenclatures, and methods. This council, consisting of 35 members, held two meetings in 2002, at which the following issues were discussed: the NPSS for 2003, the medium-term program, and updating of the statistical classifications.

The National Statistical Council is the body responsible for drafting and implementing the NPSS; it is attached to the NSI president and consists of 18 members. The roles of the National Statistical Council are to adopt a long-term strategy for the development of statistical activity, draft the NPSS, and approve a list of the structural units in the government administration, which are bodies of statistics, as well as rules on their activity.

In addition, the Methodological Council is a body responsible for implementing the principles of statistical activities incorporated in the Law on Statistics—impartiality, reliability, effectiveness, adequacy, publicity, confidentiality, and concrete issues of statistical activities such as development of methodology and planning of statistical surveys, statistical data collection, processing, analysis and storage, as well as dissemination of statistical information. This Council consists of 15 members—of which five are from the NSI and the rest from scientific, research institutions, NGOs, and other government agencies.

In general, the councils are useful and manageable mechanisms that meet regularly and reach decisions that the statistical bodies and represented parties can put into practice. To improve coordination and information between state bodies, agreements have been reached on data exchange between NSI and the main ministries and other state institutions beginning in 2000. The agreements specify the obligations of the NSI and the other institutions in providing statistical information, as well as other forms of cooperation in the field of statistics. Two annexes are included in the agreements in which statistical surveys, indicators, periodicity, terms, and information carriers are described. Annex 1 specifies NSI’s responsibilities, and Annex 2 indicates those of the other institutions. Terms of data provision are consistent with the requirements of the ARC for dissemination of the results of the statistical surveys.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

Confidentiality of reporting is guaranteed under Article 25 of the Law, which states that individual and personal data received may be used only for statistical purposes. Article 22 requires that all respondents to statistical surveys are informed of their rights and obligations, including the guarantee of data confidentiality. Data on individual reporters can only be provided to third parties if written permission has been received. In addition, Article 26 of the Law states that statistical bodies and their employees cannot make public or provide (1) individual statistical data; (2) statistical data that may be combined in a way that leads to the identification of an individual person; and (3) statistical data that summarize information for less than three units or in which the relative part of one unit is more than 85 percent of the total volume. Article 55 states that any person not complying with these provisions will be liable to a fine of up to 6,000 leva. One accidental breach of the confidentiality rules is known to have taken place, and appropriate disciplinary action was taken. All computers and databases are password controlled to restrict access to authorized staff. Destruction of unwanted survey forms is tightly controlled. Formal procedures are in place to authorize such destruction, and this is undertaken by specially licensed contractors.

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

Article 20 of the Law requires all organizations and individuals to submit data covered by the NPSS. Article 52 establishes a penalty of up to 6,000 leva for noncompliance with reporting requirements (including misreporting). If someone refuses to supply data, attempts are made to encourage participation. However, continuing refusal can result in prosecution, and such cases are publicized.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

At the end of 2002 the NSI had 1,689 staff. Of this total, 1,609 had permanent contracts and 89 had temporary contracts. In comparison with 2001, the permanent staff had decreased by 190 employees.

Six staff work on the CPI in the head office. These resources are considered to be inadequate given their responsibility for such a wide range of methodological, processing, compilation and dissemination functions. In the regional offices, price collectors also have responsibility for other collections and spend only part of their time on CPI work. These staff resources are also inadequate. A lack of financial resources has meant that there has not been any face-to-face training of price collectors over the last four years.

Low pay is a significant problem for recruiting and retaining well-qualified staff. The NSI is classified as a category 3 institution rather than category 2, such as the MOF. Category 2 would allow the NSI to improve staff pay. Following a Eurostat recommendation, a proposal to change its status was made at the end of 2001. Most agencies supported this application, except for the MOF and the Council of Ministers. Both of these bodies said there were insufficient funds available at the time. The NSI issued another proposal at the end of 2002 but has not yet received a response.

In 2002, the Council of Ministers decided that the NSI would have to vacate its current office building in Sofia because the office space was required for other purposes. Surprisingly, and disturbingly, suitable alternative accommodation for all of the NSI’s staff was not supplied. As of January 2003, some of the NSI staff are still in the old office, while attempts are made to find alternative accommodation. This situation has a serious impact on staff morale. It also demonstrates a serious lack of interest in statistics throughout the higher reaches of government.

0.2.2 Measures to ensure efficient use of resources are implemented

Funds from the central budget are assigned to the NSI each year on the basis of the State Budget of the Republic of Bulgaria. Wages and salaries and social security contributions for the personnel, together with current and capital costs, are covered in the expenditure part of the budget. The main financial problem facing the NSI relates to its operating costs. It is extremely difficult to cover the most pressing expenditures, related to conducting statistical surveys and studies, as well as overheads such as electricity, telecommunications, postal, and other external services. This calls for tight controls on expenditure, so that the NSI can run all of the surveys and studies included in the NPSS and disseminate the necessary statistical information.

The first set of main performance indicators was developed in 2000, according to the specific objectives Statistics about Statistics. The new performance indicators system, following the philosophy of the Statistical Capacity Building Indicators (PARIS 21 Task Team), covers seven dimensions:

  • effectiveness and usefulness;

  • resources—staff, budget, infrastructure;

  • transparency;

  • timeliness;

  • accuracy and relevance;

  • respondent burden; and

  • international cooperation.

For each of these indicators the approach is to review the data available (including identification of data gaps) and its usefulness as indicators of quality. This information is then used to produce time series in order to be able to see trends over time. This process is still ongoing.

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

The NSI, under Article 7(1) of the Law on Statistics, coordinates the state statistical activity by means of collaboration with the other bodies of statistics in drafting and implementing the NPSS and ensures methodological uniformity in the surveys.

The NSI, as the central body processing and disseminating statistical information, recognizes the importance of the quality of statistical data. It adopts the quality dimensions of the statistical information in accordance with the definition of quality accepted by Eurostat. The topic of quality of statistical information is contained in the Strategy for Development of Statistics and in the Priority Tasks for the period 2001–2005.

Using the General Data Dissemination System (GDDS) as a frame for development of statistics contributes to the improvement of the production and dissemination of the statistical information in accordance with the basic parameters of the system: that is, coverage, periodicity and timeliness of data, quality of data, integrity of data, and public access. Regular revisions are made to the national metadata for the GDDS, which are included in the Dissemination Standards Bulletin Board (DSBB) of the IMF. Work towards adoption of the Special Data Dissemination Standard (SDDS) of the IMF is continuing.

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

The statistical divisions of the NSI continually address the issue of the quality of statistical data received from statistical surveys. Since 2000, the Statistical Methods and Standards Division has been engaged in studying and elaborating issues of statistical data quality. For instance:

  • the theme of quality has been included and given priority in the NPSS since 2000;

  • quality assessment is connected with the development of national statistical practice on control of the quality of data and its harmonization according to the European Statistical System (ESS) requirements; and

  • focusing on quality is directed at all stages of the production and dissemination of statistical data. The main publication of NSI—the 2001 Bulgarian Statistical Yearbook—is an example of this approach. The dissemination of the methodologies of the statistical surveys (in every chapter of the yearbook) is one of the requirements on data quality. Also covered is the increasing use of administrative sources of statistical data to reduce respondent’s burden. It is also pointed out which categories of data are covered in the metadata in the DSBB.

User satisfaction on data quality has been researched. This has involved periodic marketing research on user satisfaction, surveys on respondent burden, and the attitude of businesses towards statistical confidentiality.

The NSI has adopted the principles of the Quality Declaration of the ESS and developed by the relevant Eurostat Leadership of Group on Quality. Recommendations on quality, which will be gradually applied in statistical practice for different surveys, will be considered in line with the accepted criteria established by the ESS. The reports prepared are in compliance with the European Commission and Council Regulations for the respective surveys.

0.3.3 Processes are in place to deal with quality considerations, including trade-offs within quality, and to guide planning for existing and emerging needs

The NSI holds regular meetings with users that consider the trade-offs between timeliness and quality. There are also surveys designed to get a wider perspective of user needs. The NPSS process considers the balance between supplying statistics as early as possible and ensuring that the disseminated data are of good quality. There is continuing pressure from Eurostat to improve the timeliness of data. However, moves in this direction are made only if they can be achieved by more efficient procedures rather than by reduced quality of the statistics.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

There is no external interference in the compilation and dissemination of NSI statistics. The 1999 Law on Statistics establishes the autonomy of the NSI. The Law requires that “statistical activity shall be carried out in compliance with the principles of the rule of law, impartiality, reliability, efficiency, adequacy, publicity, and confidentiality.” The Law also establishes a National Statistical Council to draft the NPSS, which, along with the rules for dissemination of statistical information products and services, ensure that objective principles are applied in the collection, processing, and dissemination of official statistics.

The NSI strives to adhere to objective international methodological recommendations in the preparation of its economic statistics.

Professionalism is actively encouraged. There are strong links between the NSI and the universities and research institutes in Bulgaria, which are mutually beneficial. Opportunities are taken for NSI staff to participate in international statistical training programs (e.g., IMF training courses), and this practice should be extended (e.g., in the fields of prices and national accounts). Between the beginning of 1999 and the end of 2002, experts have participated in two seminars held by the Joint Vienna Institute (two persons for 14 days each).

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

The NSI makes choices about sources of data and statistical techniques based solely on professional statistical considerations and in the light of international standards. Data sources are soundly based on household and business surveys. There is no evidence of outside interference.

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

Key data are presented and explained at regular monthly press conferences. One of the objectives is to prevent any misunderstandings or misinterpretations by users. In all cases of incorrect interpretation of data, the NSI responds publicly to provide the correct interpretation.

1.2 Transparency

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

Users have full access to information about the conditions under which statistical data are collected, processed, and disseminated as well as about release timetables, confidentiality of individual data, and other specific points. This information is contained in basic documents such as the Law on Statistics, rules for dissemination of statistical information products and services, calendar for presentation of the results from statistical surveys, and the catalog of NSI statistical publications, which are reproduced both on the NSI website and in its printed and electronic publications.

The fundamental principles of official statistics, methodologies, classifications, nomenclatures, mission statements, etc., are regularly released in the NSI Statistics Magazine in both printed and electronic form, and on the website.

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

Under new dissemination rules, CPI data are provided to relevant government ministers (under embargo conditions) for briefing purposes up to one hour prior to the embargo time. This new rule is only contained on the NSI website, in Bulgarian. It is not yet reflected in the IMF GDDS NSI metadata. The NSI plans to take immediate action to present the new rule in English on the website and to update the GDDS metadata.

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

The NSI is clearly identified on all its printed and electronic products through the standard inclusion of the NSI logo and a box containing publishing information and other insignia. In the case of joint publications with another institution, inclusion of the organization’s name and logo is obligatory, and the source of the information is stated in the methodological notes and tables. A basic requirement when giving permission for NSI statistics to be reproduced is that the NSI must be acknowledged as the source of the information.

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

The NSI informs users about the current status, and all significant changes, in relation to methodology, sources and data collection techniques for all the surveys in the NPSS. In-depth information is provided to the academic community and practicing statisticians through the institute’s Statistics Magazine. Similar information is also presented on the NSI website.

1.3 Ethical standards

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

In addition to itemizing specific responsibilities, the job descriptions for each NSI staff member contain a list of the laws which he/she needs to be familiar with and apply on the job i.e., the Law on Statistics, the Law for the Civil Servant, the Law for the Administration, and the Code of Behavior of the Civil Servant. While no special training course is provided to new staff on these laws, before they sign their employment contracts, they are required to sign a declaration not to release individual data which they access during the course of their everyday work.

2. Methodological Soundness

From the beginning of 1999, the National CPI has been a chained Laspeyres index with the weights changed annually. At the same time, the international classification “Classification of Individual Consumption by Purpose” (COICOP) was introduced for the CPI calculation. Since 2000, the geometric mean formula has been applied at the elementary aggregate level.

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 concepts and definitions of the National CPI generally follow the recommendations of the European System of Accounts, 1995 (ESA 95) and EU regulations. As a measure of inflation, the CPI relates to household final monetary consumption expenditure (HFMCE) and excludes imputed expenditures, especially imputed rentals of owner-occupied dwellings and own-account production of market goods for own final consumption. While these exclusions depart from ESA 95, they are in agreement with the specific EU regulations for price statistics.

2.2 Scope

The NSI compiles the National CPI that is regarded as the official measure of inflation in Bulgaria. Since 1997 a preliminary Harmonized Index of Consumer Prices (HICP) has been produced for Eurostat. The National CPI and the preliminary HICP are derived from the same database and at present are identical. No regional indices are produced. The NSI produces another index that relates to a small basket consisting of 100 goods and services that have particular social importance for low-income households in the country (the bottom 20 percent) and aims to support socioeconomic analysis. The CPI described below is the National CPI of Bulgaria.

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

The CPI covers the expenditures of all resident, noninstitutional households within the country. The weights are derived from the monthly results of a continuous Household Budget Survey (HBS). The CPI refers to market purchases of consumer goods and services made by resident households. However, the following expenditures are excluded from the index: imputed rents for owner occupied housing, dwelling purchases, major maintenance and repair of dwellings, life insurance services, pension and obligatory health insurance, gambling, self-production of goods and services, and illegally sold goods and services. Financial services are not yet covered, owing to their insignificant relative share (less than 0.1 percent of total expenditure). Note that the exclusion of both imputed dwelling rents and dwelling purchases means that housing is not covered in the CPI.

2.3 Classification/sectorization

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

ESA 95 is used to classify institutional units and transactions. COICOP is followed in the aggregation of household expenditures on individual goods and services.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks

Consumer goods and services included in the basket are valued at market prices, including VAT. That is, the valuation basis is purchasers’ prices, inclusive of trade and transport margins. Sales, or discounted, prices are included in the index consistent with the EU Regulation regarding the treatment of price reductions in the HICP. As far as practicable, detailed product specifications are defined including all price determining characteristics, for transactions at market prices.

2.4.2 Recording is done on an accrual basis

Prices of goods and services are recorded at price collectors’ visits to the selected outlets (stores, shops, restaurants, cafes, etc.). The price collection period is from the 5th to the 25th of the reference month. Prices of goods are recorded in the period in which they are purchased, and prices of services are recorded in the period in which their consumption commences (e.g., for education and package holidays). This treatment is in accordance with EC Regulation No 2601/2000 regarding the timing of entering purchaser prices into the HICP.

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

Used cars and other secondhand durables are not included in the CPI, so this is not applicable.

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

Comprehensive population and business registers maintained by the NSI are available for sampling frames. Each year the population register is updated with the annual demographic administrative data, and every ten years it is revised on the basis of the results from the population census. The HBS uses the population census data and the population register to prepare its sampling frame. However, the NSI business register cannot be used to select outlet samples for the CPI price collection because it is not sufficiently up-to-date for the volatile retail industry. Therefore, the approach to the sampling of outlets is judgmental and relies on the selection of outlets to be made by price collectors, and confirmed by the director of the relevant Regional Statistical Office (RSO). Price collectors are given instructions on how to select the outlets, i.e., they must have a large volume of retail sales and supply a variety of goods.

The main source of information for constructing CPI weights is the HBS. The HBS is continuous, with monthly and annual results regularly produced. The selected households are observed for a period of one year. The survey is based on 2-stage probability sample, which at the first stage covers enumeration districts according to their size and at the second stage, six households per enumeration district. The stratification of the sampling units is done firstly by region (28) and type of region (regional center, urban or rural), and secondly by the size of the household. The sample size for 2002 is 3,000 households (6,000 households were sampled in 2000 and 2001). The sampling frame refers to all private households and excludes people living in institutions such as nursing homes for the elderly or ill, prisons, military institutions, etc. The survey covers the whole country.

Participation in the survey is voluntary. The initial response rate of households, when first sampled, was 60.1 percent for 2001. However, all nonrespondent households are replaced with other households from the same strata. Standard errors are calculated for various survey variables. Data collection is done through an open diary and questionnaires. The recording period for the items in each diary is 15 days (i.e., within each month two diaries are completed by each household). For each 15-day period the responding households record in the diary (on a daily basis) all household expenditure, income and receipts, products produced by the household, as well as the products and services supplied from other sources in kind. The questionnaires provide data about socioeconomic and demographic variables relating to household members, foodstuffs available in households on the last day of each month, farm animals, etc. There are also questions that refer to the previous month.

The consumer basket consists of 512 goods and services. Each month, price collectors register the prices of 482 goods and services by personal visits to the outlets or service providers. The prices of the remaining 30 goods and services are centrally collected. The weight of the locally collected prices is 84.9 percent of the total CPI weight, and the weight of the centrally collected prices is 15.1 percent. In compliance with EU regulations, goods and services are included in the index if their weight accounts for more than 0.1 percent of total household expenditure. The number and types of outlets are selected in a way that assures the optimum number of collected prices, sufficient to represent national prices for any group of goods and services in the consumer basket. The main types of retail trade are covered, including the largest outlets such as supermarkets, general and specialized stores, cooperative markets, etc.

The price collection period is from the 5th to the 25th of the reference month. Judgmental sampling techniques are used for the selection of outlets, which must have a large volume of retail sales and supply a variety of goods, representative of the relevant elementary aggregates in the consumer basket. The principle of proportionality is followed. Judgmental sampling techniques are also applied to items and product varieties. Prices are collected each month from 4,589 outlets located in all 27 district centers (including the capital city); as a result, 31,151 price quotations are obtained each month. These 27 centers account for over 65 percent of national retail sales. The NSI has developed special software for data processing at both regional and national level. The collected price data are sent from the regional offices to the national office via an internal data exchange system (based on the Internet) by the end of the reference month.

Additional research and regular consultations with price collectors are undertaken in order to provide information about the weights and prices of new products that are to be included in the CPI. The press, research papers and Internet websites are also monitored for information on price developments.

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

The HBS, national accounts, and CPI data are closely coordinated, and there is broad consistency of definitions, scope, and classifications. This applies also to the time of recording, reference periods, and valuation of CPI estimates. For a few expenditure items (e.g., energy and telecommunications) details are obtained centrally from the respective providers.

3.1.3 Source data are timely

The expenditure data provided by the HBS are available at the end of the previous year and are used in CPI calculation for the current year. The collection of expenditure on a monthly basis enables the construction of annual weights for the CPI covering all household expenditure for a full calendar year, including seasonal expenditure recorded in only a few months of the year. It also facilitates an annual change of CPI weights that reflect the most recent pattern of consumption. Consumer price data are available on a timely basis.

3.2 Statistical techniques

3.2.1 Data compilation employs sound statistical techniques

The HBS provides detailed expenditure data for CPI weighting according to COICOP at the 2, 3, and 4-digit levels. In addition, HBS data are available for a national extension of COICOP to a 5th and a 6th digit which supports weighting down to the elementary aggregate level.

The goods and services purchased are valued at actual prices paid by the households regardless of the method of payment.

Since 2000, the geometric mean formula has been used to calculate the elementary aggregate indices. Higher level indices are then calculated by applying the Laspeyres chain formula, with annual links.

From 2000, prices for temporarily missing products (up to two months) are imputed at the outlet level on the basis of a matched sample of average prices (i.e., the short-term approach). If a product continues to be missing, a similar substitute is introduced.

Quality adjustment procedures are applied at the national level in a small number of selected cases only. The main approach is the overlap method in which the price difference between the old and new products in the overlap period is used as a proxy for the quality difference.

Until 2000, the long-term relative approach for imputing missing prices of fresh products was used. Commencing in 2001, the methodology has been improved through the introduction of the short-term relative approach.

Products that become permanently unavailable are replaced. New products are introduced into the sample at the time of the annual sample update when they gain market share, i.e., more than 0.1 percent of total expenditure.

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

The scope of the CPI relates to household final monetary consumption expenditure, and it is used in disaggregated form to deflate the corresponding component of the national accounts without data adjustments and transformations.

3.3 Assessment and validation of source data

3.3.1 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 planning

HBS sampling errors are acceptable and are made available. Nonresponse is not an issue as replacement households are introduced into the survey on a scientific basis.

Sampling errors are not available for the price survey because scientific sampling methods are not used for outlet and product selection. Nonresponse is not a problem as personal price collection is used.

Administrative prices and tariffs are crosschecked in the market. Other price data are checked and edited by special software and manual procedures at both regional and national office levels with rejected prices treated as missing observations. Price collectors are instructed to report if there are significant or unusual price changes and to provide relevant explanations. As the Bulgarian CPI is a chained Laspeyres index, the weighting source data are analyzed and validated each year.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Main intermediate data are validated against other information where applicable

CPI experts selectively make comparisons with relevant components or aggregates from other statistical surveys such as the PPI, export, and import price indices (for example petrol, pharmaceutical products, etc.).

During 2002, CPI price data for water supply were compared and checked with the corresponding price data from the annual survey of water supply and sewerage collection conducted by the NSI.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated

Each unusual (i.e., large or inconsistent) movement in components of the index is carefully investigated with the assistance of the price collectors and the suppliers of centrally collected prices.

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

The overall CPI is calculated as the weighted arithmetic mean of the group indices derived from the elementary aggregates weighted with the corresponding statistical weights, and therefore there is no internal inconsistency in the aggregation procedure. It is compiled only on a national basis.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used to inform statistical processes

As explained in 4.4.1 and 4.4.2 below, in line with most countries’ practices, the introduction of new weights and samples into the CPI each year does not result in revisions to the index numbers because they are not recompiled retrospectively on the basis of the new weights and samples. However, the initially released index numbers for January of each year are preliminary because they are based on incomplete HBS data (i.e., for January to November of the previous calendar year) and are revised at the time of the release of the February indices after the full year’s HBS data are available. Index numbers for all other months of the year are final when first released. Although the January revisions are minor, the differences are analyzed and explained in the publication.

4. Serviceability

4.1 Relevance

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

The survey of consumer prices is part of the NPSS, which is adopted by the Council of Ministries. The most important issues related to the CPI methodology are presented and discussed at the NSI Methodological Council which consists of users in Ministries, the National Bank and the Academy of Science. For example, the new revised CPI series for the period 1995–2000 were presented to the Methodological Council and it was agreed that the new series should be published with an explanation of the methodology. Also, as necessary, working groups are established in order to discuss and receive opinions presented by users (representatives of the Syndicates in Bulgaria, Ministries, etc.) about particular issues. Very good relations have been established with CPI users: the media, Ministries, the National Bank, the Agency for Economic Analysis and Forecasting, etc. NSI price statistics staff regularly participate in meetings and seminars organized by Eurostat, the UN, and the ILO.

4.2 Timeliness and periodicity

4.2.1 Timeliness follows dissemination standards

The CPI is normally released in the week containing the 10th of the month following the reference month and therefore exceeds SDDS requirements.

4.2.2 Periodicity follows dissemination standards

The periodicity of the CPI is monthly and complies with the SDDS.

4.3 Consistency

4.3.1 Statistics are consistent within the dataset

The NSI does not produce regional CPIs. The item/expenditure classification is used in the calculation of the National CPI. The overall CPI is invariant to the hierarchical level used to calculate the overall index.

4.3.2 Statistics are consistent or reconcilable over a reasonable period of time

The CPI is an annual chained Laspeyres index. The monthly indices for each calendar year are chained to those of the previous year in order to maintain index series for longer time periods.

The CPI Division revised the index series for the period 1995–2000 to apply the COICOP classification and introduce new methodology incorporating IMF and Eurostat recommendations. The new series have consistent coverage and methodology that meet the standards of the HICP. The revised series have been provided to Eurostat, the OECD, the IMF, and the ILO and have been published to replace the old series for this period.

The old CPI series for the period 1990–94 (May 1990=100) have been re-referenced to 1995 in order to provide historical time series. Users were advised that there is a break in the series between December 1994 and January 1995, informed of the reasons for the changes, and provided with methodological notes.

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

Following ESA 95, CPI data are broadly consistent with national accounts deflators for household consumption subject to the exclusions of imputed expenditures from the index as its scope relates to monetary expenditure.

4.4 Revision policy and practice

4.4.1 Revisions follow a regular, well-established, and transparent schedule

The NSI’s annual calendar for data releases is published on its Internet site and is available to all users. It provides information about the release dates for both the preliminary and final CPI (for January of each year). Users are also well informed, mainly through regular press conferences and NSI website postings, about the regular updates to the samples of outlets and basket of goods and services and the annual weight updates. Note that the index series are not routinely recompiled retrospectively, and revised, to reflect these new samples and weights; rather the new series are linked onto the old series each year to form a continuous time series.

Index numbers for the months of February to December of each year are final when first released and are not revised.

4.4.2 Preliminary data are clearly identified

Each year the January index is first released as a preliminary index (calculated with weights from January–November of the previous year). The final January index is released together with the February index and replaces the preliminary one. Users are informed that the initially published data for January are preliminary and subject to revision; this is clearly identified in the publications. Both preliminary and final January CPIs are disseminated with the same level of detail.

4.4.3 Studies and analyses of revisions are made public

Regarding the preliminary and final January CPIs, although the difference between them is minor, they are analyzed and explained in the publication.

In relation to the major revision for the period 1995–2000, a detailed analysis was undertaken and the results were presented to the Methodological Council. Users were notified what changes had been made, that it represented a marked improvement over the old series, and that it now complied with Eurostat requirements.

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)

At regular monthly press conferences, national CPI data are presented for the overall CPI, the 12 major COICOP headings, and the four main groups (food, nonfood, services, and catering) and along with analyses of the recent price developments. The same data are also posted on the NSI website and are sent to the Regional Statistical Offices for dissemination.

CPI back-data for the overall index and recent data at various levels of disaggregation are presented in tables and charts in the NSI’s monthly and annual publications. In addition, the NSI provides printed and electronic publication catalogs.

5.1.2 Dissemination media and formats are adequate

The CPI data are first released at the regular press conferences and via the Bulgarian Information Agency. An hour later, the information is posted on the NSI website. Subsequently, the data are published in the Statistical Journal and annually in the “Statistical Yearbook” and the “Statistical Reference Book”. The NSI Statistical Services division provides a wide range of services based on ad-hoc requests, for a fee.

5.1.3 Statistics are released on a preannounced schedule

The CPI data are released according to a preannounced schedule presented in the NSI’s annual calendar of data releases. The calendar contains the release dates and the embargo time. The calendar is available in hard copy as well as on the website.

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

The release of the CPI is tightly controlled. It is embargoed until 11.30 a.m. on the preannounced day of release. The data are provided under embargo conditions for briefing purposes to:

  • (i) media representatives at a press conference commencing at 11.00 a.m., and

  • (ii) government ministers up to one hour prior to the embargo time.

5.1.5 Nonpublished (but nonconfidential) subaggregates are made available upon request

The NSI Information Services Division can provide unpublished specialized tabulations for a fee. There are specific terms and conditions for dissemination of statistical products and services, including a price list, that are contained on the NSI website.

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 publication “Methodology for Statistical Observation of the Consumer Prices” is available to users on request. Annually updated CPI methodology is presented on the NSI website and in the NSI Statistical Journal. The GDDS summary methodologies are annually reviewed and updated. The last update was in July 2002.

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

The NSI publishes documentation at various levels of detail according to users’ needs. Methodological notes are published in the “Statistical Yearbook” and also the “Statistical Reference Book”.

5.3 Assistance to users

5.3.1 Contact person for each subject field is publicized

Contact persons are published in the Statistical Journal. The CPI press releases list the name of the contact person, including the telephone number and email address of the NSI press center.

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

A catalog of publications is published and updated each year. The prices of the statistical products and services are clearly disclosed and assistance is provided in placing orders.

Table 2.

Bulgaria: Data Quality Assessment Framework—Summary Presentation of Statistics (Consumer Price Index)

(Compiling Agency: National Statistical Institute)

article image

III. Price Statistics (Producer Price Index)

0. Prerequisites of quality

0.1 Legal and institutional environment

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

The Law on Statistics published in the Official Gazette, No. 57 of June 25, 1999, amended in No. 42 of April 27, 2001, governs the activities of the National Statistical Institute (NSI) as well as other government statistical bodies. Chapter 3 of the Law on Statistics establishes the National Statistical Council (NSC), a body attached to the NSI president and responsible for the annual National Programme for Statistical Surveys (NPSS). This program covers all significant statistical work undertaken by government agencies, including the producer price index (PPI), and is approved by the Council of Ministers. The NSI disseminates data under Chapter 5 of the Law, which requires statistical bodies to inform the public about the results from statistical surveys through mass media, own publications, and other means.

The Bulgarian Annual Statistical Plan incorporates the NPSS of the NSI and other surveys that are of interinstitutional significance. For each specific statistical survey included in the NPSS, the following items must be stated: the purpose, contents, and coverage of the survey; the methods for conducting the survey, information sources, and forms of data collection; the body or bodies conducting the survey, obligations for data provision by the respondents, and the budget for the survey.

0.1.2 Data sharing and coordination among data producing agencies are adequate

In pursuance of the Law on Statistics the NSI is responsible for coordinating all national statistical activity. The main incentives to encourage coordination among agencies include the sharing of information, participation in joint statistical activities, access to statistical expertise, and influence on the setting of priorities. The tools to achieve coordination are as follows:

  • the NPSS statistical infrastructure—classifications, registers, definitions, indicators;

  • councils, which improve relations with users, data suppliers, and other bodies of statistics, that is the National Statistical Council, High Statistical Council, Methodological Council, Editing Council, Methodology and Training Council, and IT Council; and

  • agreements with bodies of statistics and other institutions, protocols, rules, methods, and ordinance.

The Law on Statistics specifies the obligations of legal entities to submit information in response to surveys, covered in the NPSS and conducted on an obligatory basis.

The NPSS covers statistical information required for observation and compilation of the Bulgarian economic, demographic, social, and environmental activities and other activities, taking into account the availability of sources, respondents burden, and cost-effectiveness principles. The NPSS includes a plan for the NSI statistical surveys and surveys with interinstitutional significance. The NPSS covers permanently conducted statistical surveys of established public significance, other statistical surveys for the particular year, and particular surveys and surveys related to the international obligations undertaken by the Republic of Bulgaria. Exhaustive censuses are conducted according to the specific laws enacted for that purpose. The NPSS contains detailed characteristics of the surveys in terms of:

  • theme of the survey;

  • contents and object of the survey;

  • method of conducting the survey;

  • sources and form of data collection;

  • body/bodies conducting the survey;

  • obligation for data provision; and

  • budget of the survey.

The funds needed for implementing of the NPSS are provided from the State Budget. These funds are allocated to the bodies undertaking the surveys. There is no information available about the coordination of these funds, and the efficiency of their use, for monitoring. The NPSS is published in the Official Gazette.

Annual operating plans for statistical activity have been developed in conformity with NPSS. These plans contain detailed information about the allocation of tasks between different departments and divisions, including aspects of data exchange.

In accordance with the Law on Statistics there are two principal councils, i.e., the High Statistical Council and the National Statistical Council. The High Statistical Council, a consultative body to the NSI president, provides opinions and recommendation on the NSI plan for statistical surveys, drafts of legislative instruments in the field of statistics, national statistical standards, national classifications, nomenclatures and methods. This council, consisting of 35 members, held two meetings in 2002, at which the following issues were discussed: the NPSS for 2003, the medium-term program, and updating of the statistical classifications.

The National Statistical Council is the body responsible for drafting and implementing the NPSS; it is attached to the NSI president and consists of 18 members. The roles of the National Statistical Council are to adopt a long-term strategy for the development of statistical activity, draft the NPSS, and approve a list of the structural units in the government administration, which are bodies of statistics, as well as rules on their activity.

The Methodological Council is a body responsible for implementing the principles of statistical activities incorporated in the Law on Statistics—impartiality, reliability, effectiveness, adequacy, publicity, confidentiality, and concrete issues of statistical activities such as development of methodology and planning of statistical surveys, statistical data collection, processing, analysis and storage, as well as dissemination of statistical information. The Council consists of 15 members—of which five are from the NSI and the rest from scientific, research institutions, NGOs, and other government agencies.

In general, the councils are useful and manageable mechanisms that meet regularly and reach decisions that the statistical bodies and represented parties can put into practice. To improve coordination and information between state bodies, agreements have been reached on data exchange between NSI and the main ministries and other state institutions beginning in 2000. The agreements specify the obligations of NSI and the other institutions in providing statistical information, as well as other forms of cooperation in the field of statistics. Two annexes are included in the agreements in which statistical surveys, indicators, periodicity, terms, and information carriers are described. Annex 1 specifies NSI’s responsibilities, and Annex 2 indicates those of the other institutions. Terms of data provision are consistent with the requirements of the ARC for dissemination of the results of the statistical surveys.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

Confidentiality of reporting is guaranteed under Article 25 of the Law, which states that individual and personal data received may be used only for statistical purposes. Article 22 requires that all respondents to statistical surveys are informed of their rights and obligations, including the guarantee of data confidentiality. Data on individual reporters can only be provided to third parties if written permission has been received. In addition, Article 26 of the Law states that statistical bodies and their employees cannot make public or provide (1) individual statistical data; (2) statistical data that may be combined in a way that leads to the identification of an individual person; and (3) statistical data that summarize information for less than three units or in which the relative part of one unit is more than 85 percent of the total volume. Article 55 states that any person not complying with these provisions will be liable to a fine of up to 6,000 leva. One accidental breach of the confidentiality rules is known to have taken place, and appropriate disciplinary action was taken. All computers and databases are password controlled to restrict access to authorized staff. Destruction of unwanted survey forms is tightly controlled. Formal procedures are in place to authorize such destruction, and this is undertaken by specially licensed contractors.

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

Article 20 of the Law requires all organizations and individuals to submit data covered by the NPSS. Article 52 establishes a penalty of up to 6,000 leva for noncompliance with reporting requirements (including misreporting). If someone refuses to supply data, attempts are made to encourage participation. However, continuing refusal can result in prosecution, and such cases are publicized.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

At the end of 2002 the NSI had 1,689 staff. Of this total, 1,609 had permanent contracts and 89 had temporary contracts. In comparison with 2001, the permanent staff had decreased by 190 employees.

Two to three staff work on the PPI in the head office. These resources are considered to be inadequate given their responsibility for such a wide range of methodological, processing, compilation and dissemination functions. In the regional offices, PPI staff also have responsibility for other collections and only spend part of their time on prices work. These staff resources are generally reasonable, though workloads are cyclical and there can be delays in the implementation of new panels of samples during peak periods. A lack of financial resources has meant that there has not been any face-to-face training of price collectors over the last four years.

Low pay is a significant problem for recruiting and retaining well-qualified staff. The NSI is classified as a category 3 institution rather than category 2, such as the MOF. Category 2 would allow the NSI to improve staff pay. So, following a Eurostat recommendation, a proposal to change its status was made at the end of 2001. Most agencies supported this application, except for the MOF and the Council of Ministers. Both of these bodies said there were insufficient funds available at the time. The NSI issued another proposal at the end of 2002 but has not yet received a response.

In 2002, the Council of Ministers decided that the NSI would have to vacate its current office building in Sofia because the office space was required for other purposes. Surprisingly, and disturbingly, suitable alternative accommodation for all of the NSI’s staff was not supplied. So, as of January 2003, some of the NSI staff are still in the old office, while attempts are made to find alternative accommodation. This situation has a serious impact on staff morale. It also demonstrates a serious lack of interest in statistics throughout the higher reaches of government.

0.2.2 Measures to ensure efficient use of resources are implemented

Funds from the central budget are assigned to the NSI each year on the basis of the State Budget of the Republic of Bulgaria. Wages and salaries and social security contributions for the personnel, together with current and capital costs, are covered in the expenditure part of the budget. The main financial problem facing the NSI relates to its operating costs. It is extremely difficult to cover the most pressing expenditures, related to conducting statistical surveys and studies, as well as overheads such as electricity, telecommunications, postal, and other external services. This calls for tight controls on expenditure, so that the NSI can run all of the surveys and studies included in the NPSS and disseminate the necessary statistical information.

The first set of main performance indicators was developed in 2000, according to the specific objectives Statistics about Statistics. The new performance indicators system, following the philosophy of the Statistical Capacity Building Indicators (PARIS 21 Task Team), covers seven dimensions:

  • effectiveness and usefulness;

  • resources—staff, budget, infrastructure;

  • transparency;

  • timeliness;

  • accuracy and relevance;

  • respondent burden; and

  • international cooperation.

For each of these indicators the approach is to review the data available (including identification of data gaps) and its usefulness as indicators of quality. This information is then used to produce time series in order to see trends over time. This process is still ongoing.

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

The NSI, under Article 7(1) of the Law on Statistics, coordinates the state statistical activity by means of collaboration with the other bodies of statistics in drafting and implementing the NPSS and ensures methodological uniformity in the surveys.

The NSI, as the central body processing and disseminating statistical information, recognizes the importance of the quality of statistical data. It adopts the quality dimensions of the statistical information in accordance with the definition of quality accepted by Eurostat. The topic of quality of statistical information is contained in the Strategy for Development of Statistics and in the Priority Tasks for the period 2001–2005.

Using the General Data Dissemination System (GDDS) as a frame for development of statistics contributes to the improvement of the production and dissemination of the statistical information in accordance with the basic parameters of the system: that is, coverage, periodicity, and timeliness of data, quality of data, integrity of data and public access. Regular revisions are made to the national metadata for the GDDS, which are included in the Dissemination Standards Bulletin Board (DSBB) of the IMF. Work towards adoption of the Special Data Dissemination Standard (SDDS) of the IMF is continuing.

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

The statistical divisions of the NSI continually address the issue of the quality of statistical data received from statistical surveys. Since 2000, the Statistical Methods and Standards Division has been engaged in studying and elaborating issues of statistical data quality. For instance:

  • the theme of quality has been included and given priority in the NPSS since 2000;

  • quality assessment is connected with the development of national statistical practice on control of the quality of data and its harmonization according to the European Statistical System (ESS) requirements; and

  • focusing on quality is directed at all stages of the production and dissemination of statistical data. The main publication of NSI—the 2001 Bulgarian Statistical Yearbook—is an example of this approach. The dissemination of the methodologies of the statistical surveys (in every chapter of the yearbook) is one of the requirements on data quality. Also covered is the increasing use of administrative sources of statistical data to reduce respondent burden. It is also pointed out which categories of data are covered in the metadata in the DSBB.

User satisfaction on data quality has been researched. This has involved periodic market research on user satisfaction, surveys on respondent burden, and the attitude of businesses towards statistical confidentiality.

The NSI has adopted the principles of the Quality Declaration of the ESS and the relevant Eurostat Leadership Group. Recommendations on quality, which will be gradually applied in statistical practice for different surveys, will be considered in line with the accepted criteria established by the ESS. The reports prepared are in compliance with the European Commission and Council Regulations for the respective surveys.

0.3.3 Processes are in place to deal with quality considerations, including trade-offs within quality, and to guide planning for existing and emerging needs

The NSI holds regular meetings with users that consider the trade-offs between timeliness and quality. There are also surveys designed to get a wider perspective of user needs. The NPSS process considers the balance between supplying statistics as early as possible and ensuring that the disseminated data are of good quality. There is continuing pressure from Eurostat to improve the timeliness of data. However, moves in this direction are made only if they can be achieved by more efficient procedures rather than by reduced quality of the statistics.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

There is no external interference in the compilation and dissemination of NSI statistics. The 1999 Law on Statistics establishes the autonomy of the NSI. The Law requires that “statistical activity shall be carried out in compliance with the principles of the rule of law, impartiality, reliability, efficiency, adequacy, publicity and confidentiality.” The Law also establishes a National Statistical Council to draft the NPSS, which, along with the Rules for dissemination of statistical information products and services, ensure that objective principles are applied in the collection, processing, and dissemination of official statistics.

The NSI strives to adhere to objective international methodological recommendations in the preparation of its economic statistics.

Professionalism is actively encouraged. There are strong links between the NSI and the universities and research institutes in Bulgaria, which are mutually beneficial. Opportunities are taken for NSI staff to participate in international statistical training programs (e.g., IMF training courses), and this practice should be extended (e.g., in the fields of prices and national accounts). Between the beginning of 1999 and the end of 2002, experts have participated in two seminars held by the Joint Vienna Institute (two persons for 14 days each).

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

The NSI makes choices about sources of data and statistical techniques based solely on professional statistical considerations and in the light of international standards. Data sources are soundly based on household and business surveys. There is no evidence of outside interference.

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

Key data are presented and explained at regular monthly press conferences. One of the objectives is to prevent any misunderstandings or misinterpretations by users. In all cases of incorrect interpretation of data, the NSI responds publicly to provide the correct interpretation.

1.2 Transparency

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

Users have full access to information about the conditions under which statistical data are collected, processed, and disseminated as well as about release timetables, confidentiality of individual data, and other specific points. This information is contained in basic documents such as the Law on Statistics, rules for dissemination of statistical information products and services, calendar for presentation of the results from statistical surveys, and the catalog of NSI statistical publications, which are reproduced both on the NSI website and in its printed and electronic publications.

The fundamental principles of official statistics, methodologies, classifications, nomenclatures, mission statements, etc., are regularly released in the NSI Statistics Magazine in both printed and electronic form, and on the website.

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

There is no internal government access to the PPI prior to its release embargo time.

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

The NSI is clearly identified on all its printed and electronic products through the standard inclusion of the NSI logo and a box containing publishing information and other insignia. In the case of joint publications with another institution, inclusion of the organization’s name and logo is obligatory, and the source of the information is stated in the methodological notes and tables. A basic requirement when giving permission for NSI statistics to be reproduced is that the NSI must be acknowledged as the source of the information.

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

The NSI informs users about the current status, and all significant changes, in relation to methodology, sources and data collection techniques for all the surveys in the NPSS. In-depth information is provided to the academic community and practicing statisticians through the institute’s Statistics Magazine. Similar information is also presented on the NSI website.

1.3 Ethical standards

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

In addition to itemizing specific responsibilities, the job descriptions for each NSI staff member contain a list of the laws which he/she need to be familiar with and apply on the job i.e., the Law on Statistics, the Law for the Civil Servant, the Law for the Administration, and the Code of Behavior of the Civil Servant. While no special training course is provided to new staff on these laws, before they sign their employment contracts, they are required to sign a declaration not to release individual data which they access during the course of their everyday work.

2. Methodological Soundness

At the beginning of 1999 the National Statistical Institute (NSI) commenced the introduction of a new producer price index (PPI) methodology to accord with Eurostat requirements. A work program was adopted for the progressive implementation of transaction pricing to replace the use of unit values. Currently 75 groups (3rd digit level) of NACE Rev. 1 are included in the new price survey with these groups representing about 83 percent of the sales of industrial enterprises. The current base year for the PPI is 1995, but the NSI is in the process of updating the PPI to a 2000 base year.

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

PPIs for the domestic market are calculated following EU Regulation 1165/98, which established data requirements in relation to coverage, periodicity, and timeliness. NSI also uses the 1998 PPI Eurostat Manual on Business Statistics and, more recently, the new Eurostat manual Methodology of Short-term Business Statistics (interpretations and guidelines).

PRODCOM concepts and definitions are used in the selection of the individual products for monthly pricing. The PPI relates to output prices, with separate indices of final use also calculated. Input price indices and stage of processing indices are not yet produced.

The PPI Manual is followed in relation to the definition of individual product specifications, that is:

  • (i) the prices quoted represent prices achieved for a significant proportion of sales of industrial enterprises;

  • (ii) the products are quoted on the same basis each month, (i.e., the same product and terms of sale);

  • (iii) the product and the terms of sale are described in sufficient detail to ensure consistency each month; and

  • (iv) prices are ex-factory (ex-mine, etc.), selling prices (i.e., basic prices which exclude trade and transport margins and taxes on products).

2.2 Scope

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

The PPI covers the mining and manufacturing industries, the production and distribution of electricity and steam, and natural gas and water supply. Following EU recommendations, uranium mining, nuclear fuels, weapons and ammunitions, and shipbuilding are excluded because of difficulties in obtaining reliable data for these activities. Price indices for agricultural products are produced and released separately but are not incorporated in the PPI. Some transport prices indices are also calculated but they are based on unit values.

All resident market enterprises are in scope for the PPI. The index relates to sales on the domestic market, and exports are not included.

The coverage of the PPI differs from the 1993 SNA ideal in that it excludes changes in stocks of finished products, own-account production of market goods for own final consumption, output of market goods for own-account fixed capital formation, illegal market goods or services sold to willing buyers, production of selected services for own final consumption, and production of services for own account fixed capital formation. These exclusions are common for national PPIs and are accepted under EU regulations.

2.3 Classification/sectorization

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

The 1993 SNA is followed in the classification of institutional units and transactions. The National Classification of Economic Activities (NCEA), which complies with NACE Rev. 1, is used for the classification of statistical units to the 4-digit level. The classification of products is based on a combination of NACE, the Classification of Products by Activity (CPA) and PRODCOM down to the 10-digit level (the last two digits have been added for national purposes). A further three digits have been added (making a total of 13 digits) to provide for the unique classification of detailed product specifications for pricing.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks

Weights and prices are valued at sales prices (ex-factory, ex-mine, etc., selling prices), excluding VAT and other taxes. That is, the valuation basis is basic prices. Price data are collected for products produced and sold on the domestic market. The prices are actual transaction prices and include discounts, rebates, and other reductions. Excise duties and VAT are excluded.

Each specific product is precisely described when first selected and is carefully identified each month when its price is reported.

2.4.2 Recording is done on an accrual basis

Sales data used for weighting are recorded at the time the products are sold. Price data for the specific products surveyed relate to the time of the sales contract. Transaction prices are recorded as at the 15th day of each month, where possible. If no price information is available for that particular day, the reported price refers to the day when the contract or sale is concluded. PPI data are recorded when products are sold, not when they are produced.

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

As the PPI is a gross sector index, and there is no Stage of Production index, transactions between enterprises are recorded on a gross basis and no netting procedures are required.

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 annual PRODCOM industry census is used for the selection of PRODCOM groups, and the NSI Register of statistical units, EKPOU 96, is used for the selection of enterprises. The main purpose of the Bulgarian business register (EKPOU 96) is to serve as a sampling frame for statistical surveys and for the extrapolation of their results. The Law of Statistics (article 30) provides the legal basis for establishing and maintaining the register. The main source of information for updating the register for births and deaths of legal units is the administrative register, BULSTAT, that is also maintained by the NSI. By law, all economic subjects are obliged to register in BULSTAT in order to legitimately conduct economic activity. The business register (EKPOU 96) relates to all economic activities covered by NACE, Rev. 1.

All units, irrespective of their size in terms of number of employees or turnover, are included on the business register. Units are classified according to their principal activity at the class level (4-digit) in accordance to the National Classification of Economic Activities (NCEA), which complies with NACE, Rev. 1. The business register is updated on the basis of two sources: BULSTAT and the Structural Business Statistics (SBS) survey. Updates from BULSTAT are received monthly, and annually from the SBS survey.

The value of sales in the PPI base year is extracted from the annual PRODCOM survey for index product weighting purposes. PRODCOM covers all industrial enterprises, as well as nonindustrial enterprises (with an annual turnover of more than 75,000 leva), which have significant industrial activity.

The source for activity weights is the mandatory Annual Structural Business Statistics (SBS) survey which is conducted in accordance with the requirements of EU Regulation 58/97. The SBS survey is a census of all economic activities with the exception of NACE classes 65.11, 65.21, 65.22, 65.23, and 66.02.

The SBS survey, as well as the PRODCOM survey, are included in the NPSS and, according to article 20 of the Law on Statistics, all units (legal persons and natural persons) carrying out economic activity are obliged to participate and to provide the Bodies of Statistics with the data required. The Law of Statistics includes provisions relating to fines, which can be imposed for refusals, nonresponses, and delays.

The response rate varies significantly depending upon the size of the enterprise and its activity. The final overall average response rate in terms of turnover is about 95 percent. For the industrial sector, the response rate is close to 100 percent. Efforts are made to maximize response rates. First, a reminder letter is sent stating that the Law on Statistics requires that those that do not respond are required to pay a fine. Depending upon the relative importance of a nonrespondent in terms of their size or representativeness of their industry, the mail survey may be supplemented by a telephone interview and/or visit to collect the data. Data for nonrespondents may be estimated by a variety of methods including secondary sources. Estimates for some nonrespondents are based on relationships from the corresponding NACE activity and size category. Missing data items are estimated by using the values from similar units.

To improve quality, the results are carefully analyzed, first at the enterprise level. Internal consistency of the indicators within the questionnaire is checked by reference to ratios between variables. The consistency of results relative to previous years’ results is also checked. Second, the results are checked with data from other surveys conducted by the NSI (PRODCOM Survey, SBS survey, short-term surveys, and International Trade).

The PPI survey, based on the new transaction pricing methodology, covers 83 percent of industrial sales; the remaining 17 percent is still represented by the old unit value methodology, but it is planned to be converted to the new methodology during 2003. For the new survey, a three-stage sampling process is used based on PRODCOM groups, then reporting units and finally specific products, or transactions. Probability proportional to size techniques are used for selection to ensure sufficient coverage is obtained. These techniques involve identifying strata that will be selected with certainty (i.e., probability=1) with the remaining strata selected with probabilities based on their relative contribution to sales.

Questionnaires are sent to about 2,000 enterprises. Approximately 7,800 specification prices are collected for about 850 product groups.

After an enterprise is selected, a qualified employee from a Regional Statistical Office (RSO) visits the enterprise to select a sample of products (transactions) that are most representative of changes in the prices of the PRODCOM group(s) selected for disaggregation. Every product selected for monthly pricing is specified in detail including characteristics of the transaction. Purposive sampling techniques are used to select products that have high sales volumes and that have been produced for a reasonable period of time.

The producer price index aims to take account of changes in quality that can be caused by a change in either the physical characteristics of a product or a change in the transaction characteristics.

A monthly mail questionnaire is used for the price collection. A new computer program for data entry and collection control was recently introduced.

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

The source data used for weights and prices under the new methodology reasonably approximate the definition, scope, classification, and valuation and timing requirements of the PPI. The NSI plans to commence compiling an export price index. Field investigations have revealed that the 50 largest exporting enterprises account for 65 percent of the total value of industrial exports.

3.1.3 Source data are timely

The price collection is completed in a timely manner to conform to the release schedule. The mail questionnaires are sent out before the end of the reference month by the RSOs. The deadline for response is the 20th of the reporting month. RSOs check the data, and on the 25th of the reporting month data files are sent to Central Statistical Office. Data collection is normally closed two days before publication.

The weighting and sampling source data are also available on a timely basis, which support a five-year index reweighting program and the maintenance of samples.

3.2 Statistical techniques

3.2.1 Data compilation employs sound statistical techniques

The Industrial Producer Price Index measures the monthly change in prices of industrial products produced by Bulgarian enterprises and sold on the domestic market. Indices are calculated at the 10-digit PRODCOM level. Also, aggregated indices are produced at the group level (third digit), division level (second digit), and at section and subsection levels of NACE Rev. 1. Indices are calculated for Main Industrial Groupings (MIGS), following EU Regulation 586/2001 concerning short-term statistics.

The PPIs are base-weighted (Laspeyres) indices. That is, they are weighted according to sales in the base year, which is currently 1995. PPIs are calculated at successive levels of aggregation with indices for each level of aggregation based on the weighted arithmetic mean of indices from the level below. The weights of the lower level indices are sourced from the PRODCOM survey, and weights for PPIs at higher levels of aggregation are based on the value of sales on the domestic market in the base period drawn from SBS survey results.

The weight reference period will be updated every five years according to the EU Regulation 1165/98. The weights are currently over seven years old (based on 1995 data), and the NSI is in the process of changing the base year to year 2000 = 100. When new weights are introduced, the new index will be linked to the old index using internationally accepted techniques.

In relation to quality changes of the surveyed products, efforts are made to ensure that only ‘pure’ price changes are recorded in the PPI. Quality changes can be caused by a change in either the physical characteristics of a product or a change in the characteristics of the transaction. Currently, four different methods are used for quality corrections in the new PPI system, depending upon the particular circumstances. This is an area where more training and resources would improve results.

Prices for temporarily missing products are handled by one of three methods: (1) asking the enterprise to estimate the missing price, (2) imputing the price based on the change in price of similar products, or (3) repeating the last reported price.

In the case of seasonal products, when selecting the product samples, it is ensured that there are some price data available for each month of the year. This is achieved through working closely with respondents. When the price for a particular seasonal specification is not available, an imputation is made using the short-term price change from a matched sample of specifications from the same product group or enterprise.

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

The PPI covers mining, manufacturing, electricity, water, and gas and has very high coverage of the traditional industrial goods producing industries. Transaction pricing techniques are employed for the majority of the industry output (some 83 percent of the sales of the industrial enterprises) with the remainder measured by the old methodology using unit value indices, pending conversion to the new methodology over the next year.

The other price indices compiled by the NSI relating to agricultural products (using transaction pricing methodology) and the transport sector (based on unit values) provide a significant increase in economic coverage.

3.3 Assessment and validation of source data

3.3.1 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 planning

In relation to the weighting data sources, the Structural Business Statistics (SBS) and PRODCOM surveys have no sampling error impact because they are censuses. The principal activity of the enterprises is observed and updated in the SBS survey so there is no significant misclassification impact. There are small nonresponse rate errors, which vary depending on the size of the enterprise and its economic activity. Rigorous data editing concentrates on the largest enterprises, though effort is put into checking the internal consistency of data for all enterprises by analyzing ratios between variables. The coherence of results with previous years’ results is also checked.

Sampling and nonsampling errors are not available for the producer price survey because scientific sampling is not used. The response rate is nearly 100 percent. Missing prices are imputed using the techniques described in 3.2.1 above. Unusual movements in the PPI are investigated. Large price changes (more than +/- 15 percent) are detected by computer editing and then queried with respondents and amended if necessary. RSO staff responsible for price collection contact the enterprises by phone to check the ongoing representativeness of the product specifications and to obtain replacement specifications if necessary.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Main intermediate data are validated against other information where applicable

The NSI experts responsible for the compilation of the index regularly compare component PPI data by product, or by different levels of aggregation, with relevant components or aggregates from other surveys such as the CPI, export and import price indices.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated

Unusual movements in the index arising from large movements in particular sectors or from particular respondents are investigated during the process of compiling the PPI. Unusual movements are not normally explained in descriptive documentation; however, they are explained to users as necessary.

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

PPI indices at higher levels of aggregation are calculated directly as aggregations of product indices so there are no discrepancies in aggregates.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used to inform statistical processes

The PPI data are normally final when first released, and revisions are rare and insignificant. As such, there are no routine analyses of revisions. The revisions may be caused by reporting or other errors discovered for previous months. Revision studies are completed following the 5-yearly PPI revision, when new weights, industries, and products are introduced.

4. Serviceability

4.1 Relevance

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

Meetings are periodically convened with policymakers and other statistical users to review existing statistical surveys and the resultant data and to identify unmet demand. The PPI survey is included in the NPSS. NPSS is discussed by the High Statistical Council, which includes the main data users and aims to improve information provision. The Methodological Council, which includes representatives of the other statistical bodies, academic circles, and other statistical data users, discusses the methodologies of new or changed statistical surveys. Professional committees of experts from other statistical bodies, institutions, etc., including trade unions, have been established with the aim of improving coordination and cooperation between institutions in relation to statistical data exchange.

The new methodology for PPI calculation is published in the NSI’s Statistics magazine. Producer price statistics experts regularly participate in meetings and seminars organized by Eurostat.

4.2 Timeliness and periodicity

4.2.1 Timeliness follows dissemination standards

Monthly producer price indices are available 40 days after the end of the reference month, which does not meet the SDDS standard of one month. Publication of the monthly indices was discontinued in 2001 pending the implementation of the new methodology. Users may obtain the PPI data and special tables from the Information Services Division of the NSI, upon request. As the new methodology has been largely implemented, the NSI proposes to shortly recommence publishing the PPI each month and release it on the NSI website.

4.2.2 Periodicity follows dissemination standards

Producer price indices are calculated monthly and conform to the SDDS.

4.3 Consistency

4.3.1 Statistics are consistent within the dataset

PPI by product and activity, and at different levels of aggregation, are consistent.

4.3.2 Statistics are consistent or reconcilable over a reasonable period of time

In 1998, a new system was introduced for the calculation of the producer price index using the Laspeyres formula, to replace the Paasche formula, and with 1995 as the base year. Indices have been recalculated back to January 1995 on a consistent basis. Users were informed of the new methodology.

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

The PPI conforms with EU requirements and is broadly in line with deflators of industrial production and comparable components of the CPI.

4.4 Revision policy and practice

4.4.1 Revisions follow a regular, well-established, and transparent schedule

The PPI data are normally final when first released, and routine revisions owing to misreporting, etc. are rare and very minor. Any changes to methodology, etc. resulting in retrospective recompilation of the series, as happened in 1998, are announced to users in advance, and documentation of the changes is provided with the revised data series.

4.4.2 Preliminary data are clearly identified

Users are informed that data may be revised. Revised data are disseminated with the same level of detail as the previously published data.

4.4.3 Studies and analyses of revisions are made public

All differences between original and revised PPI data are analyzed. Revised PPI data are presented in the same aggregates as the preliminary data. Users are informed of causes of revisions to the price series.

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)

Publication of the monthly indices was discontinued in 2001 pending the implementation of new methodology. The data are not currently presented at the regular press conference nor released on the website. Users may obtain the PPI data and special tables from the Information Services Division of the NSI, upon request. As the new methodology has been largely implemented, the NSI proposes to recommence publishing the PPI each month, commencing in March 2003, and release it on the website.

5.1.2 Dissemination media and formats are adequate

See 5.1.1 above.

5.1.3 Statistics are released on a preannounced schedule

The producer price indices are available for users under a regular timetable. However, because they are not currently published, they are not listed in the annual calendar of data releases at present.

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

The PPI is available, upon request. As it is not published or announced in the calendar of data releases, in practice access to all users is not simultaneous.

5.1.5 Nonpublished (but nonconfidential) subaggregates are made available upon request

With the permission of the President of the NSI, special tables and more detailed data can be obtained by users from the Information Services Division (subject to confidentiality rules). There are specific terms and conditions for the dissemination of statistical products and services, including a price list, that are contained on the NSI website.

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 new “Methodology of the Producer Price Index on the Domestic Market” was published in the NSI Statistics Magazine in 2000. The GDDS summary methodologies are regularly reviewed and updated—the last update was in July 2002.

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

The new “Methodology of the Producer Price Index on the Domestic Market” was published in the NSI’s Statistics magazine. After publication of the PPI recommences, the methodology will be also published. The methodology describes the different levels of detail at which the PPIs are calculated and are available. Methodological notes are published in the “Statistical Yearbook” and in the “Statistical Reference Book.”

5.3 Assistance to users

5.3.1 Contact person for each subject field is publicized

The PPI is currently not published, but the NSI Information Services Division provides users with contact details where expert advice can be obtained.

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

The catalog for NSI publications is available and updated every year. However, the PPI is not listed because it is not currently published.

Table 3.

Bulgaria: Data Quality Assessment Framework—Summary Presentation of Statistics (Producer Price Index)

(Compiling Agency: National Statistical Institute)

article image

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

Under the Organic Budget Law (July 25, 1996) the Ministry of Finance (MOF) has the responsibility to collect, process and disseminate fiscal statistics on the budgetary and extra-budgetary central government and local governments. The Law on the Preparation and Implementation of the State Budget (published in the State Gazette, additional issue No. 91, 1960; amendments and additions, No. 99, 1963, No. 36, 1979, No. 27, 1992 and No. 104, 1995) was repealed with the 1996 Organic Budget Law. The Fiscal Statistics and Accounting Division, Budget and Treasury Directorate in the MOF is dedicated to the compilation of fiscal statistics. Compilation of statistics on government debt is governed by the State Debt Law (October 1, 2002), which notes (Article 34) that the Bulgarian National Bank (BNB) shall register government and government guaranteed debt, but (Article 38) that the MOF maintains the official register of State Debt and State Guarantees and that it shall publish information on the Consolidated State and State Guaranteed Debt monthly in an official bulletin and on the Internet.

0.1.2 Data sharing and coordination among data producing agencies are adequate

There is no standing statistical committee to share information across data producing agencies; data sharing arrangements are informal and seem to resolve all data issues satisfactorily. The NSI has prepared and maintains a complete list of public units that comprise the different subsectors of the general government. This list is routinely updated and shared with the MOF and BNB for use in classifying their data consistently with the national accounts definition of general government. The government finance statistics (GFS) derivation links are built into the so-called “budget classification” system (classification codes that serve as a chart of accounts for cash data collection), used for budgeting and accounting purposes. These data are used, in turn, as direct inputs into the national accounts. Weekly data from the BNB are compared with MOF sources, and detailed quarterly banking data for government agencies are confirmed by the BNB and audited by the National Audit Office (NAO). Debt data are taken directly from the BNB database. Budgetary central government, extra-budgetary agencies, and local governments are required to report fiscal data to the MOF.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

Confidentiality of individual responses is not generally an issue. Fiscal data are not confidential, but must be officially approved to be disseminated to the public.

Tax authorities have the obligation to keep confidential any circumstances that have become known to them in the discharge of their duties (Art. 60, Taxation of the Income of Natural Persons 1999 as amended). Violators will be fined 500–1,500 Bulgarian leva (US$250-$750), if not subject to heavier punishment, including penal decrees by the Chief Tax Director applying the Code of Tax Procedure (Articles 64-65). Specifically the Code of Tax Procedure states in Article 12 on confidentiality: “Article 12. (1) The tax administration officials and the recruited contractual experts, as well as public enforcement authorities and experts, shall have the obligation to keep confidential and not to use for other purposes but discharge of their official duties, all circumstances or facts, which have become known to them and have been determined as confidential under this Code, including for a five year period of limitation following the termination of their employment/temporary assignment contracts. (2) The confidential information may be disclosed on the basis of a court ruling, as well as at the written request of another tax authority following the procedure, determined by the General Tax Administration Directorate in connection with executing his/her powers under this Code.”

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

Under the Organic Budget Law (July 25, 1996) the MOF has the authority to require all units that comprise the budgetary and extra-budgetary central government and local governments to report fiscal data according to a format (template) determined by the MOF. Reporting instructions issued by the MOF are designed to feed into budget execution monitoring and are coordinated with the NAO. Relevant paragraphs of the Organic Budget Law are quoted below:

Organic Budget Law (1996)
Section 5
Article 37.

(1) The Accounting Chamber shall monitor state budget implementation in accordance with the Law on the Accounting Chamber.

(2) The National Assembly, the Council of Ministers, the Minister of Finance, the leaders of the ministries and other departments, and agencies of the tax and customs administrations shall monitor state budget implementation within their jurisdictions.

Article 40.

(1) Primary budget credit managers shall compile periodic and annual reports on the cash implementation of their budgets using the complete budget classification, incorporating the reports of secondary budget credit managers.

(2) The Ministry of Finance shall compile a report on state budget implementation using the complete budget classification on the basis of reports on the central budget, the judicial authorities, and primary budget credit managers certified by the Accounting Chamber.

(3) The Ministry of Finance, in coordination with the Accounting Chamber, shall set forth procedures and deadlines for compiling and submitting the reports referred to in Paragraphs 1 and 2.

Article 41.

(1) The annual report on state budget implementation on indicators approved for the year in question shall be reviewed by the Council of Ministers and submitted to the National Assembly.

(2) The commission responsible for the budget shall in coordination with other standing commissions of the National Assembly, review the report on state budget implementation and make findings concerning it.

(3) The Accounting Chamber shall submit to the National Assembly a paper on the state budget implementation report.

(4) The National Assembly shall resolve to approve the state budget implementation report no later than the end of the next fiscal year.

Section 7
Article 49.

(1) Reports on extrabudgetary accounts and funds shall be prepared and submitted together with those on the implementation of corresponding budgets and budgetary accounts.

(2) Reports on extrabudgetary accounts and funds shall be certified by the Accounting Chamber using certification procedures for reports on the implementation of state authority budgets.

(3) Reports referred to in Paragraph 1 shall be submitted by the Minister of Finance to the Council of Ministers, which shall review and submit them to the National Assembly together with the state budget report.

Article 50.

Funds of others shall be collected, used and reported by state authorities and budget organizations using procedures set forth by the Minister of Finance and the Governor of the Bulgarian National Bank.

Article 51.

Extrabudgetary funds shall be monitored by the authorities, and using the procedures, established for the state budget.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

Resources appear adequate for current GFS compilation demands. The Fiscal Statistics and Accounting Division in the MOF Budget and Treasury Directorate comprises eight professional staff to develop Bulgaria’s GFS methodology and compile the general government statistics (including the development of underlying MOF accounting and reporting rules for source data units). Retention of staff has remained stable in recent years. The Budget and Treasury Directorate overall contains approximately 40 staff to support these statistical efforts and to use the data for fiscal policy purposes. In addition, the Government Debt Directorate comprises 20 staff to compile domestic and foreign debt statistics for the central government. Computer resources are adequate for data manipulation and storage; computer center services are outsourced. Building facilities are adequate and are being refurbished. Two staff in the Fiscal Statistics and Accounting Division have been trained in the Government Finance Statistics Manual 2001 (GFSM 2001) or European Debt and Deficit methodologies. However, despite these adequate facilities for current fiscal statistics, more resources would speed the implementation and development of the GFSM 2001 and EU fiscal data reporting requirements. Thus far none of the Debt Directorate staff have been nominated to training courses on the new methodology.

0.2.2 Measures to ensure efficient use of resources are implemented

While the costs associated with compiling fiscal statistics are not measured, the authorities claim that greater efficiency in the use of available resources has been demonstrated in recent years. In addition to continuing to produce currently available fiscal statistics, the staff are simultaneously preparing for the implementation of a new set of fiscal data as part of the EU accession arrangements. In particular, the staff have developed a new chart of accounts, begun to develop a new Financial Information Management System (FIMS), and are preparing changes in instructions for reporting agencies, methodological notes, and new accounting guidelines. All of these developments would support future efforts to implement the GFSM 2001.

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

There is a recognition throughout the MOF that quality builds trust and is the cornerstone of statistical work. Data quality processes are official, developed by MOF staff, and instructions are signed by the Minister of Finance.

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

Instructions to reporting agencies, actions to deal with missing reports, verifications, crosschecks, and audits are among the processes to monitor the quality of fiscal data collection, processing, and dissemination. For example, the “budget classification” chart of accounts is imposed by the MOF on reporting agencies. The Organic Budget Law requires data to be reported and audited by the NAO. International statistical and accounting standards are taken into account in monitoring, and planning improvements to statistical quality. The new Accountancy Law (2001) requires the MOF to develop requirements and guidelines for public sector units that are consistent with international accounting standards by the beginning of 2005. Arrangements have been put in place on the Web in January 2003 to obtain feedback from users of fiscal data. On the MOF website, users’ feedback is solicited through feedback@minfin.government.bg, as well as through an innovative online chat room advertising open communication with the Minister of Finance.

0.3.3 Processes are in place to deal with quality considerations, including tradeoffs within quality, and to guide planning for existing and emerging needs

Monthly data collected from reporting units is less detailed than quarterly reports, reflecting the recognition of the trade-off between timeliness and the quality of detail. The more detailed quarterly reports are scrutinized more carefully against other sources of data (especially banking data) and are audited by the NAO within 60 days. Quality is taken into account in statistical planning, especially as part of the plans to develop fiscal statistics that will meet the requirements for notification to the EU and the migration to the GFSM 2001.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

The Fiscal Statistics and Accounting Division of the MOF operates professionally. Part of the efforts to promote professionalism have included publication of articles on fiscal statistics methods (and the development of balance sheets), participation in seminars and training, as well as the provision of training to Bulgarian reporting units on the implementation of the new chart of accounts. Moreover, the quarterly data are audited by the NAO, which is independent from the executive branch of government. The National Assembly elects the NAO’s chairman and members.

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

Consistent with the provisions in the Organic Budget Law, fiscal statisticians choose the most appropriate sources and methods for compiling data without pressure. Consistent with the GFSM 2001 and ESA 95, the fiscal compilers have classified privatization proceeds and the net acquisition of financial assets with public corporations (“net lending minus repayments” in the GFSM 1986 system) below-the-line.

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

The Fiscal Statistics and Accounting Division provides expert advice on technical (not policy) aspects of the fiscal data. They occasionally clarify any misinterpretation of the data; for example, explanatory material has been provided to users (such as rating agencies).

1.2 Transparency

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

The Organic Budget Law notes the terms and conditions under which fiscal statistics are collected, processed, and disseminated, as well as the requirements for audits, and Council of Ministers and National Assembly approval. It is available to the public and appears on the MOF website.

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

There is no internal government access to the fiscal statistics prior to their release, except as part of the publication process.

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

All MOF statistical publications and data disseminated on the Web clearly indicate that the data are compiled by the MOF.

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

Notes on changes to methodology accompany the data on the MOF website to inform users of changes in methodology, source data, and statistical techniques. Institutional tables appear on the website to highlight for users structural changes in the economy that affect the coverage of the government data.

1.3 Ethical standards

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

As noted in the Law for the Civil Servant (as amended SG 45/02), Article 25(1) “The civil servant shall be obliged to protect the classified information representing state or official secret that have become known to him at or on the occasion of fulfillment of his official duties.” The compilation and dissemination of fiscal statistics also are governed by the Law on Access to Public Information (July 7, 2000) and the Law on Protection of Classified Information (April 30, 2002). MOF staff are well aware of the rules concerning classified information.

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 (GFSM 1986 guidelines)

The Bulgarian GFS concepts and definitions are based on the recommendations of the IMF’s A Manual on Government Finance Statistics, 1986 (GFSM 1986). The Bulgarian authorities have not fully developed an explicit action plan to migrate to the GFSM 2001. However, they have taken steps to implement the statistical methods for the general government statistics, based on the European System of Accounts, 1995, consistent with the requirements of EU accession, and in preparation for reporting according to the EU Excessive Deficit Procedure under the Stability and Growth Pact (Eurostat notification). In particular, two key steps have been the development and planned implementation of a new unified chart of accounts and preliminary actions to develop a new Financial Management Information System (FMIS) for cash and accrual recording. The new harmonized chart of accounts has not yet been linked to the EU notification table or GFSM 2001 codes; however, it has taken into account the main recommendations of the 1999 technical assistance from the IMF and has been supported by other international organizations. These developments, particularly those taken to prepare for EU accession, have positioned Bulgaria well for future migration to the GFSM 2001.

2.2 Scope

2.2.1 The scope is broadly consistent with internationally accepted standards, guidelines, or good practices (GFSM 1986 guidelines)

Data are available on all cash transactions within the scope of GFS following the GFSM 1986. There are two distinct presentations of the fiscal statistics for the entire general government prepared by the Fiscal Statistics and Accounting Division in the MOF Directorate of Budget and Treasury: a national presentation, and an international presentation (labelled, “GFS 1986 presentation”). While both use the GFSM 1986 guidelines and practices, the national presentation of general government data uses a framework that more closely follows the structure of the Bulgarian budget. Thus the Bulgarian fiscal data on the basis of the “national budget classification” really refers to general government data. The international “GFS 1986 presentation” uses the standard international framework and includes a presentation of expenditure by function. The national presentation does not show expenditure by function, even though such quarterly data are compiled for internal purposes.

The national presentation explicitly shows the coverage of the general government data and shows the transactions among the units eliminated in consolidation. The national presentation is organized as follows, beginning with the units of central government (and using the Bulgarian terminology as translated into English). The “Republican Budget” includes 1) the “Central Budget,” which mainly includes interest payments and transfers to other parts of the general government; 2) line ministries (33, also called first-line spending units) and agencies (second-line or sub-ministerial units); 3) and the NAO. The “State Budget” adds to the “Republican Budget” budgets of the judiciary authorities. The rest of central government includes, in addition to the “State Budget,” 1) the budgets of social insurance schemes (the Pension Fund, the Industrial Injuries and Professional Illness Fund, the Common Illness and Maternity Fund, the Health Insurance Fund, and the Fund for Professional Qualification and Unemployment (FPQU)4); 2) universities, the Bulgarian Academy of Science, Bulgarian National TV, and Bulgarian National Radio (all aggregated together); and 3) extra-budgetary funds and accounts. In the annual data extra-budgetary units of social security schemes and local governments are included with the subsector to which they correspond; however, in the monthly data all extra-budgetary activity is reported as a separate subsector of the general government. To complete the general government, local governments are added (263 municipalities and 5,078 communes); the municipalities report on the activities of the communes. Coverage is identical for the international “GFS 1986 presentation” of Bulgarian fiscal data.

Data are presented in both the national presentation and the international “GFS 1986 presentation” for central government, local governments, and for the consolidated general government. There are no state, regional, or provincial governments. No data are collected on the rest of the public sector for publication, in particular nonfinancial public corporations. Thus, data are not compiled for the consolidated nonfinancial public sector.

Depending on the year under consideration, certain funds may be included as extra-budgetary or as social security funds.5 In addition, rapid structural changes to this transition economy have resulted from privatizations (for example, hospitals and medical clinics were transformed from budget units to non-financial corporations by the end of 2000) or the absorption of public corporations into the general government. As a result, significant changes in coverage for the general government data have occurred over time. These structural changes are reflected correctly in the data and institutional tables for each year of data list the changes in coverage.

The MOF Government Debt Directorate compiles and disseminates monthly consolidated central government gross debt (domestic and foreign by instrument), including the debt of social insurance schemes, in its Government Debt Management bulletin. At the same time, it provides data on central government debt guarantees. The sources of the central government deficit (presently constituting only domestic securities and foreign loans) are shown explicitly in the annual MOF publication Annual Review of the Government Debt. The MOF has strengthened the methodology used to record debt data through the “Computerized Information System for Analyses, Forecast and Management of the Government and Government Guaranteed Debt” program that began in 1999. The Fiscal Services Department in the BNB compiles and disseminates central government domestic debt by holder in its Government Securities Market monthly bulletin (also available on the BNB website). A major difference in coverage between the two data sets concerns transactions with the IMF. The BNB considers purchases from the IMF as a liability of the BNB, which is on-lent to the Government through domestic financing. This is domestic financing in the fiscal data. In contrast, the MOF Debt Directorate reports these transactions with the IMF as external debt. Data on the stock of gross debt for local governments are not available; data collection began in 2003. However, the national presentation of annual fiscal data shows separately local governments’ domestic and external financing. As of the end of 2002, preliminary data has shown that local government debt was negligible, representing less than 0.2 percent of total central government debt.

Preliminary data submitted to the MOF and the NAO are replaced with the audited data. However, in general all data are submitted on a timely basis; therefore there are no partial data that would need to be replaced with data based on full coverage.

2.3 Classification/sectorization

2.3.1 Classification/sectorization systems used are broadly consistent with internationally accepted standards, guidelines, or good practices (GFSM 1986 guidelines)

In general, revenue, expenditure, and financing transactions are classified using the methodology set out in the GFSM 1986. However, according to Section 1 Article 2(7) of the Organic Budget Law (1996): “Revenues and expenditures shall be classified according to the uniform budget classification, which shall be approved by the Minister of Finance.” Thus, this budget classification serves as the framework for the national presentation of fiscal data. In contrast, the international “GFS 1986 presentation” follows the GFSM 1986 structure more closely (see 3.1.2).

There are some exceptions to the GFSM 1986 classifications in the national presentation. Privatization receipts are included under financing rather than under net lending minus repayments. Beginning in 2002, general government acquisition of financial assets and lending to nonfinancial corporations also are classified as financing instead of net lending minus repayments. These departures from the GFSM 1986 are in line with the GFSM 2001.

The international “GFS 1986 presentation” data include a summary table, revenue and grants, expenditure by economic classifications and by functions, and financing (domestic and foreign). Further details on domestic and external financing flows are not disseminated. There are specialized publications on central government debt stocks (see 2.2.1). Data are provided on the central government gross debt, classified by residency (foreign and domestic) of the counterparty to the liabilities, as well as by instrument.

2.4 Basis for recording

2.4.1 Prices are used to value flows and stocks reflect actual cash payments (GFSM 1986 guidelines)

Flows are valued on the basis of the amount of cash received (cash deposits) and cash payments made (cash withdrawals). However, the debt data are not valued consistently. Domestic debt is valued on an issue price (cash received) basis, that is at face value less discounts plus premia, which is not consistent with the GFSM 1986 recommendation to value debt at par (full redemption) value. Foreign debt is valued at full face value, regardless of discounts or premia, that is on the basis of cash due at maturity, consistent with the GFSM 1986.

Amounts denominated in foreign currency are converted to national currency using the midpoint exchange rate between the buying and selling market rates on the date of the transaction. The exchange rates are provided by the BNB.

2.4.2 Recording is done on a cash basis (GFSM 1986 guidelines)

Flows are valued on the basis of the amount of cash received (cash deposits) and cash payments made (cash withdrawals), consistent with the GFSM 1986 guidelines. Foreign and domestic central government debt stocks are recorded at issue price (face value less discounts or plus premiums), that is based on cash received. Currently, little information is collected on noncash supplementary data, especially on arrears and nothing on arrears is disseminated. Noncash domestic debt is included (associated with the Law on Settlement of Nonperforming Credits negotiated prior to end-1990, so called ZUNK bonds).

The planned new FMIS will allow complete recording of budgetary transactions on an accrual as well as a cash basis. The new FMIS, using SAP R3 software, has been successfully piloted in two MOF units. It contains three modules: budget, accounting, and financial reporting. The FMIS is targeted to apply initially to 39 budgetary units (all 33 front line ministries and six MOF second line directorates) and automatically collect monthly, quarterly, and annual budget and accounting data. On current plans, full implementation of the new FMIS will take several years, but it will support implementation of the GFSM 2001 and the data requirements for accession to the EU.

In parallel with the FMIS, new national accounting standards are to be developed to fulfill the requirements of the Accountancy Law (Nov. 16, 2001). This requires the MOF to develop accounting standards and documentation rules for all general government units by January 1, 2005, consistent with international accounting standards (the International Public Sector Accounting Standards approved by the International Federation of Accountants).

2.4.3 Grossing/netting procedures are broadly consistent with internationally accepted standards, guidelines, or good practices (GFSM 1986 guidelines)

All transactions are shown on a gross basis, except for tax refunds and financing transactions, which are shown net. This is consistent with the GFSM 1986. Fiscal data on the subsectors of the general government are recorded on a gross basis and intersectoral flows are eliminated in consolidation.

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

Current administrative systems provide data covering the full range of economic stocks and flows for central and local government units. All units are required to report as specified in the Organic Budget Law, the Social Security Code, and the Law on Health Insurance. Reporting units develop their own internal systems based on 10 different kinds of software packages in use in the Bulgarian government sector. Data are available to allow consolidation for both levels of government, as well as subsectors within the central government (social security and extra-budgetary units) and general government. Surveys are not needed because all information is provided through administrative sources.

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

All general government units report according to the national “budget classification,” which serves as a harmonizing chart of account to collect and compile fiscal data. A bridge table has been created from the “budget classification” to the “national presentation” of fiscal data, and from those data to the international “GFS 1986 presentation.” This provides automatic derivation of GFS items from budget items. Accounting standards are delineated in the Accounting Act (2001), the Excise Tax Act and Value-Added Tax Act. Currently data are reported on a cash basis. Data are reported by all government units each month; however, the data are reported on a cumulative, rather than a discrete, basis. Monthly data are derived as a residual, preventing distinction between revisions to previous months’ data and flows that occurred during the latest month.

3.1.3 Source data are timely

Most central government units report their data 12 days after the end of the reference month, but local governments and some second-line central government units report data within 15–20 days. This allows monthly budget execution reports on the operations and financing of general government (and both levels of government) to be prepared approximately 20–25 days after the reference month. The MOF also receives more detailed data quarterly from reporting units within one month of the end of the reference quarter. The BNB collects information on the cash balances of all general government units and provides this information to the MOF twice per month. Flash reports on cash execution for the general government are compiled two times per month (on the 17th and at the end of the month following the reference month). The BNB also provides cash balance data to the MOF quarterly. Central government debt data are submitted by the 5th of the month after the reference month, and the data are finalized by the 10th of the month.

3.2 Statistical techniques

3.2.1 Data compilation employs sound statistical techniques

All data come from administrative and accounting sources that are required under the Organic Budget Law to report on a timely basis; therefore, there is no need to estimate missing data and no surveys are employed. The MOF contacts late reporters by phone to obtain those submissions.

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

Suspect submissions or discrepancies concerning data to be eliminated in consolidation are investigated and resolved, frequently through telephone contacts.

3.3 Assessment and validation of source data

3.3.1 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 planning

All data are provided through administrative records and are complete (see 3.2). Systematized assessments of late reporting, suspect submissions, and discrepancies are assessed periodically. Nevertheless, persistent difficulties may be addressed through regular follow-up with the reporting unit, or on-the-job training.

There is an appropriate balance between accuracy and timeliness because unaudited monthly data are reported within one month of the reference month but, in contrast, quarterly data are audited by the NAO (see 3.4.3) and inaccuracies are addressed within 2½ months of the reference quarter. The data are reported on a cumulative, rather than a discrete basis (see 3.1.2). Monthly data are derived as a residual, preventing distinction between revisions to previous months’ data and flows that occurred during the latest month. Therefore, it is difficult to discover and correct material differences in data except for the latest month. These quarterly corrections typically are small and are not disseminated to the public.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Main intermediate data are validated against other information where applicable

Fiscal data reported by the BNB, NSI, and the Agency for Economic Analysis and Forecasting are fully consistent with the MOF data. In particular, financing data are checked and validated with BNB data on government deposits, the issue of domestic securities, and sources of external financing (bonds and loans).

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated

The MOF fiscal data are cross-checked against all other intermediate data, such as information from tax and customs administrations (see also 3.4.1 and 3.4.3).

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

Any cross-checks with intermediate data that give rise to discrepancies are resolved by the Fiscal Statistics and Accounting Division of the MOF with the providing agency, usually via phone or email, but occasionally through formal letters. The BNB submits data on government accounts at commercial banks and at the BNB to the MOF twice a month and quarterly. Significant discrepancies are identified. The quarterly data are reviewed and validated by the NAO. The MOF and the NAO investigate any discrepancies on a quarterly basis.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used to inform statistical processes

Data are reported cumulatively, so that typically errors and omissions are captured in the latest month’s data. There are no revisions to the published data. Revision studies are not conducted, and corrections made during the compilation process are not systematic or thoroughly documented.

As part of the compilation process, quarterly data are audited by the NAO, but this does not lead to systematic revisions of the cumulative data for the quarter. Except for the audit report, material corrections are not fully documented (that is the main aspects of the source data leading to the corrections, the reasons for the corrections, and the direction and magnitude of the corrections are not always recorded).

The process to review and assess corrections is not formal, documented, or systematic to determine the reliability of data. However, audits and periodic feedback with source data compilers lead to improvements in the compilation process through on-the-job training and contacts with MOF fiscal data compilers.

4. Serviceability

4.1 Relevance

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

The international “GFS 1986 presentation” of Bulgarian fiscal data are derived from the national presentation, and both are used in the process of determining and evaluating fiscal policy. The GFS data are used primarily by the international community, including international financial institutions (such as the IMF and World Bank) and international rating agencies. The national presentation of the fiscal data is used primarily in Bulgaria. Monthly fiscal data are available in sufficient detail to allow users to assess the degree to which government is achieving its stated social and fiscal policies. However, greater detail could be provided to the public on the financing of government and on the types of expenditure by function (currently only aggregates are available for the latter in the international “GFS 1986 presentation”). While the methodological basis and classification structure of the fiscal data are adequate for users’ analytical needs, migration to the methods and classifications in the GFSM 2001 and the EU Deficit and Debt Guide would enhance the policy relevance and the analytical information for users. Feedback from users and potential users on the relevance of existing statistics is not sought on a regular basis; however, the recently renewed website http://www.minfin.government.bg/en/ contains email addresses and phone contacts for such feedback.

4.2 Timeliness and periodicity

4.2.1 Timeliness follows dissemination standards

In principle, central government operations data are published approximately within one month after the reference period (reference month). The publication of these data on the Web began in January 2003, so experience is limited with meeting the SDDS timeliness requirements on the MOF website. Prior to January 2003, no monthly data were available. The monthly MOF budget bulletin is disseminated about one month after the reference month, but the level of detail and analytical usefulness of the tables are limited, especially for the consolidated central government operations and financing.

Monthly central government debt stocks are available three weeks after the reference month. This exceeds the SDDS requirement that quarterly central government debt data be disseminated within one quarter after the reference period.

Annual general government operations data are normally presented to the NAO on the 24th of January, and the NAO issues its report in mid-February. Preliminary annual data normally would be published in March following the reference year. Again, the publication of these data on the Web began in January 2003, so experience is limited with meeting the SDDS timeliness requirements using this medium. Prior to January 2003, the MOF website only contained annual data (as of December 2002 the latest data were for 2000).

4.2.2 Periodicity follows dissemination standards

Monthly central and general government operations data (the national presentation) are published monthly in the MOF budget bulletin. Beginning in January 2003, the MOF website has monthly and annual central and general government data following the national presentation, and monthly, quarterly, and annual data following the international “GFS 1986 presentation.”

Monthly central government debt stock data are disseminated monthly, and annual data are disseminated annually. Monthly debt stocks simultaneously provide the information to create a quarterly series. This exceeds the SDDS requirements.

4.3 Consistency

4.3.1 Statistics are consistent within the dataset

Fiscal statistics are fully internally consistent within the database: deficits/surpluses equal financing (except the subsector of general government in the “GFS 1986 presentation” which is reported on a consolidated basis); major aggregates equal the sum of their components; expenditure by function equals expenditure by economic classification (and interest expenditure is identical in both sets of tables); transfers paid equal transfers received (and are checked carefully between levels of government and between subsectors of central government); domestic financing is reconcilable with changes in domestic debt (differences are mainly due to zero coupon bonds measured at the issue price (cash received), revaluations, and noncash “ZUNK” bonds); and foreign financing is reconcilable with foreign debt. Financing differs from the debt statistics, due to the misclassification of IMF purchases in the debt statistics (as foreign rather than domestic lending from the BNB, see 4.3.3).

4.3.2 Statistics are consistent or reconcilable over a reasonable period of time

Monthly data are fully consistent with quarterly and annual data. Monthly data are reported on a cumulative basis. Therefore, subannual data do not represent activity that took place during the month (or quarter) because errors and revisions are captured in the latest month’s (quarter’s) data reported.

As a transition economy, Bulgaria has experienced significant structural changes. Therefore, the coverage of the data over time has been affected. Units that were part of the government may have shifted to the public or private corporations sectors. Significant privatizations have affected the government’s accounts and, therefore, made it more difficult for users to compare fiscal data over time. The authorities note that it would be impractical to make adjustments to the statistics; however, institutional tables for 1992–2001 and other metadata on the website signal these changes in coverage across time to users.

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

In general the fiscal data are not formally reconciled with the national accounts, monetary, or balance of payments statistics. However, in principle, it should be possible to produce such reconciliations, and the data validation and cross-checks support the notion that the data are fully consistent. General government sector data in the national accounts are based on the data reported by the MOF and are internally consistent. The BNB provides information twice a month and quarterly on all general government units’ deposits in the BNB and the commercial banks, and it also supplies information on its activities as an agent of the government to issue domestic securities (Treasury bills and bonds). This information is used to compare the domestic financing reported by individual reporting units. External financing is extracted from the BNB database using Reuters software. Differences across data sets frequently result from structural changes in the economy that have not been made fully consistent, but cross-checks resolve these differences.

A major difference in coverage between the fiscal and monetary data sets concerns transactions with the IMF. The BNB considers purchases from the IMF as a liability of the BNB, which is on-lent to the government through domestic financing. In contrast, the MOF erroneously records these transactions with the IMF as external debts. However, an adjustment is made so that external financing excludes these transactions and domestic financing reflects the on-lending from the BNB. Another difference stems from valuation of government debt, reflected at market prices by the BNB and at cash value (issue price for domestic debt and maturity value for external debt) by the MOF.

4.4 Revision policy and practice

4.4.1 Revisions follow a regular, well-established, and transparent schedule

Revision policy and practice are not publicly disseminated. Quarterly data are submitted by reporting units in more detail than monthly reports. These quarterly data are vetted thoroughly using cross-checks with BNB data and audit procedures. Thus data for the last month of the quarter are preliminary when reported and disseminated within roughly one month; however this is not done. There are no revisions made on a regular, well-established, and transparent schedule concerning government domestic debt statistics. Foreign debt statistics are revised to take late submissions into account. None of these practices are made known to the public.

4.4.2 Preliminary data are clearly identified

Preliminary data are not separately distinguished. Foreign debt revisions are not clearly labeled in the publications. In general, preliminary data are sufficiently close to final data in most cases so that they can be used with confidence for policy determination and analysis.

4.4.3 Studies and analyses of revisions are made public

There are no revision studies or analyses to be made public.

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 refreshed website (as of January 2003) contains the international “GFS 1986 presentation” as well as the national presentation of fiscal data and recent time series; it also contains a hyper-link to older data (http://www2.minfin.government.bg). Most of the fiscal data are presented in a way that allows major aggregates and balancing items to be identified and related to detailed underlying data. The two major exceptions concern details on financing and on expenditure by function. The national presentation is used mainly to monitor budget implementation and to formulate future fiscal policy.

5.1.2 Dissemination media and formats are adequate

The MOF disseminates two monthly publications (the monthly budget bulletin in hardcopy only and Government Debt Management in hardcopy and on the MOF website) and an annual publication in hardcopy and on the MOF website (Annual Review of the Government Debt). The BNB (Monthly Bulletin) and the Agency for Economic Analysis and Forecasting (monthly bulleting and semi-annual survey) reproduce some of these data. The BNB also disseminates hardcopies of the monthly Government Securities Market, which contains domestic debt stock information. As of January 2003, the MOF website contains more timely fiscal data, in both the national presentation and the international “GFS 1986 presentation.” Maintenance of this website with current data remains to be tested. The BNB website contains the fiscal data disseminated in its hardcopy publications.

5.1.3 Statistics are released on a preannounced schedule

There are no advanced release calendars for fiscal data. The annual data using the national presentation are prepared in March and are typically approved by the Council of Ministers for submission to the National Assembly by end-June. Organic Budget Law Article 41(4) states “The National Assembly shall resolve to approve the state budget implementation report no later than the end of the next fiscal year.” However, this was not done for the data concerning 2001. Therefore, to meet SDDS requirements the authorities would need to disseminate the annual data approved by the Council of Ministers. The data could be revised following National Assembly approval, if needed. The monthly bulletins are usually released about one month after the reference period. There is no explicit calendar of release dates for fiscal data.

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

Statistics are made available to all users at the same time through hardcopy and web dissemination formats. No prior release or preferential access to the data occurs.

5.1.5 Nonpublished (but nonconfidential) subaggregates are made available upon request

Nonpublished, but nonconfidential, information is provided via email or hardcopy without charge. The availability of nonpublished but nonconfidential data and the terms and conditions under which they are made available has not been publicized in the GFS publications. However, the contact section of the MOF website notes that users can obtain additional information about the budget (including historical fiscal data) through this contact via email or by phone.

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

Broad statements on concepts, scope, classifications, basis of recording, and differences from international standards are annotated (on the website and GDDS only). An elaborate document on concepts, sources, and methods has not been compiled. Bridge tables showing the links between “budget classification” (chart of accounts) and the fiscal presentations are made available on request. Although rare, universities and international organizations have requested and received this information.

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

Only broad statements on concepts, sources, and methods are available. More detailed metadata have not been requested by users and have not been developed. The less detailed metadata accompany the data (on the Web) to orient users concerning the basic framework under which the fiscal statistics have been produced.

5.3 Assistance to users

5.3.1 Contact person for each subject field is publicized

The GDDS and MOF websites list contact persons for fiscal data (and other budget questions). Users can write to a designated MOF official or directly to budget@minfin.government.bg. Although not widely announced, prompt and knowledgeable service and support is available from the Fiscal Statistics and Accounting Division, Budget and Treasury Directorate, and the Government Debt Directorate. However, in most instances, users would be directed to contact the MOF Information and Public Relations Directorate to coordinate queries (at feedback@minfin.government.bg or by phone in Sofia, Bulgaria on (02) 98.59.20.24). No officer in the MOF has been designated to provide a rapid response to the misuse of fiscal data. However, the senior staff in the units listed above would attempt to correct errors by users, when noticed.

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

The MOF does not have a separate catalog of publications. However, the monthly budget bulletin can be obtained via subscription and the relevant information is available on the MOF website, including the subscription price and contact address (http://www.minfin.government.bg/en/act/index.php?lang=2&catid=426&depth=1). In addition, MOF website browsers can subscribe to the MOF news items online.

Table 4.

Bulgaria: Data Quality Assessment Framework—Summary Presentation of Government Finance Statistics

(Compiling Agency: Ministry of Finance)

article image

V. Monetary 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

Monetary statistics are collected, compiled, and disseminated by the Bulgarian National Bank (BNB), the central bank of the Republic of Bulgaria. The BNB collects data from all operating commercial banks and also from banks in liquidation. The legal basis for data collection is given in Article 4(1) of the Law on the Bulgarian National Bank (1997) and Article 55 of the Law on Banks (1997). The Law on BNB states that the BNB can demand that the banks submit any documentation and information it needs in connection with the performance of its functions.

The Rules of Procedure on the Functions and Responsibilities of the Structural Units and the Powers of Senior Executives of the Bulgarian National Bank state that the Monetary and Banking Statistics Division (MBSD), which is a unit of the BNB’s Statistics Directorate, is responsible for compiling monetary statistics. The MBSD is also responsible for the development of methodology conforming with the international standards provided by the IMF and the European Central Bank (ECB).

The Statistics Directorate is identified in the list of structural units of the Bulgarian administrative institutions that are designated as statistical bodies and that constitute the national statistical compilation system (Articles 3(2), 14(4) of the Law on Statistics (1999)). This list is defined by the National Statistical Council (NSC), in which the director of the Statistics Directorate represents the BNB (Article 14(2)(4) of the Law on Statistics). The MBSD’s mandate to compile and disseminate monetary statistics for Bulgaria is also specified in the Rules for the Activity of the Statistics Directorate approved by the NSC (Article 14(4) of the Law on Statistics).

0.1.2 Data sharing and coordination among data producing agencies are adequate

The BNB is the sole institution involved in compiling monetary statistics. It collects all requisite source data needed for this purpose. The MBSD disseminates monetary statistics to domestic and foreign users. Under current practice, the Statistics Directorate regularly provides to the MOF, the NSI, and other government institutions specific nonconfidential data collected by the BNB that those institutions need for their own statistical compilation and verification programs. In particular, detailed records on positions of government spending units with the banks are routinely transmitted to the MOF.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

The Law on Statistics (Chapter 6 on the BNB’s Statistics Directorate) and the Law on BNB (Articles 4(2) and 23(1)(2)) ensure the confidentiality of individual data and require that data be disseminated in a manner that prevents identification of data on individual transactions and reporting institutions. The BNB staffs access to data is only for work-related needs. Maintaining confidentiality is an obligation of all BNB employees, and in this respect MBSD staff follow these rules. BNB data and data collected from the commercial banks are used by MBSD staff only for statistical purposes.

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

Article 4(1) of the Law on BNB (see section 0.1.1) and Article 55 of the Law on Banks provide the legal basis for statistical reporting and data collection. Article 55 of the Law on Banks specifies that banks shall submit to the central bank reports in a format, content, periodicity, and timeliness established by the central bank. Further, Articles 61 and 62 of the Law on BNB specify significant penalties imposed for its breach, including for failure to report to the BNB documents and information it may demand from the banks under Article 4(1) of the Law on BNB.6 These legal instruments form a part of the general legal framework for data collection and reporting obligations.

Commercial banks are also encouraged to report to the central bank since they benefit from the availability of comprehensive monetary statistics needed for analytical and other purposes. Comprehensive and detailed instructions are provided to the banks to facilitate their data compilation and reporting. Additional support, as needed, is available at the BNB through the designated contacts at the MBSD and the Banking Supervision Department.

At present, the list of financial institutions required to report for the monetary statistics compilation program includes only the banks. This current legal mandate for statistical reporting is adequate. When other (nonbank) financial institutions qualify for inclusion in the coverage of the broad money survey, the reporting mandate will need to be extended so as to include them in the reporting population.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

At the beginning of 2003, the number of MBSD staff was increased from 15 to 17, and a new three-section structure replaced the previous two-section one. The three sections within MBSD cover methodology, monetary statistics, and interest rate statistics.

The current MBSD resources are sufficient for the division’s statistical program. Nonetheless, the staff resources are overburdened by the growing needs to harmonize compilation of monetary statistics in line with the ECB requirements. The number of MBSD staff cannot grow at the expense of other divisions within the Statistics Directorate because their staff needs are also increasing, owing to work on harmonization with the ECB’s and EUROSTAT’s data requirements.

Raising qualifications of the MBSD staff is part of the BNB’s general human resource management policy. Division staff attend specialized courses offered by other central banks and international institutions. Internal competition for attendance is coordinated by the BNB’s Human Resources Directorate. The EU PHARE-financed project between the BNB and Banque de France includes technical assistance in compiling money and banking statistics.

Computing resources at present cover the needs but will have to grow in line with the growth in the MBSD activities. Computing resources are provided by the BNB’s Informatics and Communications Directorate (ICD) according to an annual allocation plan as part of the BNB’s annual budget.

Financial resources for MBSD are allocated from the Statistics Directorate’s financial resources, which are a component of the BNB’s annual budget approved by the Parliament.

0.2.2 Measures to ensure efficient use of resources are implemented

The plan for the use of staff, financial, and computing resources is drafted by the director of the Statistics Directorate with the input from the head of MBSD (and other divisions within Statistics Directorate). The annual plan, with a quarterly breakdown, is approved by the Deputy Governor responsible for the Banking Department (the Statistics Directorate is a unit of that department).

The BNB Budget Committee, which is a standing executive management body established by a decision of the BNB Managing Board, is responsible for the preparation, monitoring, and reporting on the execution of the BNB’s budget. The use of resources by the MBSD is monitored and reported by the director of the Statistics Directorate following the rules set up by the Budget Committee.

The efficient use of computing resources is monitored by the Automation Executive Committee, an auxiliary body under the BNB Managing Board (among this committee’s tasks are the development of strategy for automating banks’ main operations).

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

Raising quality of statistics as a public good is of primary importance for the Statistics Directorate. The quality of monetary statistics is monitored by the MBSD head in discussions with the staff in the methodological section of the division. MBSD’s aims at present are to ensure conformity with the international standards such as ESA 95, the ECB’s framework for monetary data, and also compatibility with the IMF’s Monetary and Financial Statistics Manual (MFSM).

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

The quality of data, which are received electronically from the commercial banks, is checked by the ICD using automated procedures. The quality checks performed by the MBSD staff include checks of individual and aggregated data series focusing on their intertemporal behavior and logical coherence. Monetary statistics for dissemination are checked by the MBSD head and the director of the Statistics Directorate.

A regular contact with users is maintained. An email address of the MBSD is provided to users to ask questions, express opinions, and give recommendations. Users’ questions are answered within three working days. Informal surveys of users are also carried out.

0.3.3 Processes are in place to deal with quality considerations, including tradeoffs within quality, and to guide planning for existing and emerging needs

The focus at present is on quality data collection and compilation. A special inter-organizational group of experts from the MBSD and commercial banks has been set up to discuss methodological issues related to data quality. Documents related to the project are posted on the BNB website, and an email address mbstatistics@bnb.bg is provided for user comments.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

The statutory provisions under which monetary statistic are compiled by the BNB are adequate to ensure its independence in performing these functions. Statutory provisions in the Law on BNB and Law on Statistics provide the legal basis for institutional independence and professionalism. Professionalism of MBSD staff is underpinned by the BNB recruitment procedures. The selection and promotion process evaluates exclusively the merit and the potential of all candidates and current employees. An informal mentoring program for new staff is in place.

MBSD staff attend meetings organized by international statistical agencies, including the IMF and the ECB, for consultations and to foster the development of national statistical practices. MBSD staff participate in the coordinating meetings in the area of monetary statistics held by the ECB with the central banks of the accession countries. Additional training is offered in Sofia at the International Banking Institute, which is operated by the BNB and the commercial banks.

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

The choices of source data and statistical techniques for monetary statistics are determined solely by statistical considerations conforming with the principles in ESA 95, ECB’s framework for monetary data, and MFSM.

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

Supported by its institutional independence, the BNB seeks to build trust in its statistical activity by commenting publicly, in mass media or in any other fora, on the perceived misinterpretations or misuse of monetary statistics. To facilitate the use and interpretation of the disseminated statistics, detailed methodological notes have been provided to all users of monetary statistics on the BNB’s website, as well as in its publications.

1.2 Transparency

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

Documents detailing the methodology for collecting, processing, and disseminating monetary statistics are posted on the BNB website and are included in its publications. The various legal instruments underpinning BNB’s statistical activity are also posted on its website.

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

There is no internal government access to BNB’s monetary statistics prior to their release.

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

All BNB publications, as well as its website, clearly identify the BNB as the data-producing agency for its outputs. Data obtained from other agencies (e.g., MOF, NSI) are clearly identified. After dissemination through the website or its publications, the BNB-produced statistics can be quoted or reproduced without further permission; however, due acknowledgement is requested.

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

The BNB’s policy in statistical dissemination is that major methodological changes are announced in advance. New data first appear in the weekly publication and then in the monthly bulletins. Dissemination of data and announcements of methodological changes are first made through the BNB’s website and later in the hard-copy monthly publications. Following a major revision in early 2002, monetary statistics were disseminated in two formats—the old and the new—for a period of three months.

1.3 Ethical standards

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

Guidelines for the BNB staff behavior are specified in the BNB’s Rules of Procedures. BNB staff have access to this document on an internal BNB website. New employees are briefed on these rules and responsibilities by the Human Resources Directorate. Further, in accordance with the Law on BNB (Article 23(1)), when taking office, the BNB employees sign a declaration of oath to observe this law, maintain data confidentiality, and dutifully perform the functions entrusted in connection with the BNB’s activities.

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 concepts and definitions used by the BNB for compiling monetary statistics conform with the principles presented in the IMF’s Monetary and Financial Statistics Manual (MFSM).7 The monthly monetary survey compiled by the BNB (the Monetary Survey table)8 is equivalent to the MFSM’s depository corporations/broad money survey. It presents measures of liquidity and its counterparts, with comprehensive detail on depository corporations’ positions by economic sector and financial instrument. As a general principle, banks’ positions in leva and foreign currency on all financial instruments are always shown.

The measures of liquidity are broad money and its components, progressing from narrow money (M1), to narrow money plus quasi-money (M2), and finally to broad money (M3), each disaggregated by the constituent financial instruments.9 Narrow money M1 is composed of currency outside banks and demand (transferable) deposits in leva; quasi-money comprises time, savings, and all foreign currency deposits of the nonbank, noncentral government sectors.10 The M2 aggregate comprises M1 and quasi-money. By its national definition, the broad money aggregate M3, in addition to narrow money and quasi-money, includes restricted deposits11 and other funds mobilized by banks from the resident nonbank, nongovernment sectors, namely debt securities issued and loans (“deposits”).12

The counterparts to measures of liquidity are net claims on nonresidents (foreign assets (net)) and domestic assets net of central government deposits. Foreign assets (net) are disaggregated by financial instrument. Domestic assets have three major components: domestic claims net of central government deposits, capital and reserves shown as a negative claim, and other items (net). Domestic claims net of central government deposits are first disaggregated by sector of debtor and then by financial instrument. Capital and reserves consist of funds contributed by owners, reserves, and the financial result account. Other items net comprises net interbank balances, fixed assets, and unclassified assets minus unclassified liabilities.

The monetary survey is derived by consolidating accounts of the central bank and other depository corporations (ODCs). The monthly BNB survey (the BNB Analytical Reporting table) and the ODC survey (the Commercial Banks’ Analytical Reporting table) contain a comprehensive level of sectoral and instrument detail at least as extensive as in the broad money survey. In addition to the sectoral surveys (in the MFSM-sense), the BNB compiles the monthly sectoral balance sheets (the Monthly Sectoral Survey tables) for the BNB and the ODCs. The sectoral balance sheets are in a matrix format that displays positions with the various resident and nonresident economic sectors by the underlying financial instrument.

To provide a fuller view of banking developments, the BNB also prepares and publishes a monthly analytical survey for the nonoperating banks (banks in liquidation) and their sectoral balance sheet. The survey and the sectoral balance sheet are in the same detailed format as used for the operating commercial banks.

2.2 Scope

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

The scope of the depository corporations/broad money survey (the Monetary Survey table) is consistent with the MFSM principles. The survey presents positions of all resident depository corporations vis-à-vis other sectors of the economy and with nonresidents. The depository corporations sector covers the BNB and all resident ODCs, which in Bulgaria consist of all commercial banks licensed by the BNB. Licensed commercial banks comprise banks with domestic headquarters and branches of foreign banks.

As of end-December 2002, there were four banks in liquidation (which are referred to as the nonoperating banks) whose total assets were approximately 0.2 percent of the commercial banks’ total assets. These four banks, whose banking licenses have been revoked, are excluded from the coverage of the broad money survey (this practice accords with the ECB’s framework for monetary data under which they are not considered monetary financial institutions, and would be excluded from the coverage of the money creating sector).13 However, to provide a fuller picture of banking developments, a separate survey of nonoperating banks (the Analytical Reporting of Nonoperating Banks table) with a comprehensive sectoral and instrument detail is compiled and published in the BNB’s monthly Monetary Statistics Bulletin and the general BNB Monthly Bulletin.

There are at present no other (nonbank) financial institutions (viz., money market funds) that would meet the criteria for ODCs in terms of issuing liabilities that are a close substitute for deposits (and that would also meet the ECB’s criteria for other monetary financial institutions).

2.3 Classification/sectorization

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

The sectoral and instrument classification is broadly consistent with the MFSM principles. The distinction between domestic and external transactors is based on the definition of residents in the Foreign Exchange Law (1999). This law specifies that residents are (a) physical persons permanently residing in Bulgaria or having a permit to stay in the country for more than six months; (b) legal persons registered in Bulgaria; and (c) physical persons (and their families) staying abroad on a mission of the Bulgarian state or Bulgarian enterprises. All others are considered nonresidents. This definition broadly conforms with the residency principle based on center of economic interest discussed in the Balance of Payments Manual, fifth edition (BPM5).14

It should be noted that legal persons are considered residents upon their registration. The NSI maintains a register of legal persons (the BULSTAT) and is responsible for their sectoral classification in conformity with the ESA 95 standards, which are generally consistent with the MFSM principles. The physical persons’ dwelling address, the BULSTAT, and the legal registry documents of legal persons are the basis for sectoral classification of clients by the banks. When a client opens a bank account (either as a debtor or a creditor), the account number includes codes identifying his economic sector, as well as the financial instrument underlying the account.

The following resident sectors and subsectors are separately identified15 (with two exceptions mentioned in the footnotes): the financial sector S.12 (composed of the central bank (BNB) S.121, other depository corporations S.122 (the licensed commercial banks), and nonbank financial institutions/other financial corporations (comprising other financial intermediaries S.123, financial auxiliaries S.124, and insurance corporations and pension funds S.125);16 general government S.13 (composed of budgetary and extrabudgetary central government S.1311, budgetary and extrabudgetary local governments S.1313, and the social security funds17 S.1314); nonfinancial corporations S.11 (composed of public (S.11001) and private (S.11002 and S.11003) nonfinancial corporations); and households S.14.18

The following financial instruments are recognized: monetary gold, SDRs, currency, deposits (comprising demand, time, savings, and restricted deposits), credits (“loans”), securities other than shares, shares and other equity, and debt securities issued (as defined by ECB regulations). Accounts payable/receivable are not separately identified and, as per ECB regulations, are included in other assets/liabilities. As a general principle, positions on each financial instrument are always disaggregated by national/foreign currency position.

In addition to the aforementioned financial instruments, credit aggregates present data on “other” position, which, with one exception, are mostly relatively small.19 Much of these data capture certain historical positions prior to the accounting reform in June 1995 that are not easily amenable to classification by financial instrument. The data also include accrued interest on these positions.

A supplementary table20 on the BNB’s foreign assets and liabilities contain data on Bulgaria’s financial relationship with the IMF, including Reserve Position in the Fund and the Use of Fund Credit (Liabilities to the IMF, Credits). However, shares and other equity component of Other Foreign Assets aggregate21 also includes data on the quota paid in leva. Further, Liabilities to the IMF, Other aggregate includes positions on the IMF No.1 account. Inclusion of these positions may be redundant, and on the gross level would amount to double-counting. Therefore, their compilation deviates from the recommended practice. (The BNB will implement the requisite revisions for the December 2002 issue of the Monetary Statistics Bulletin, which will be published at the end of January 2003.)

The repurchase agreements are treated as collateralized loans rather than as outright sales of securities. Starting in January 2003, liability positions arising from repurchase agreement operations are being separately identified. There are no gold swaps, but in that event, they too would be treated as collateralized loans. Nonnegotiable instruments that possess all other properties of securities other than shares are classified as credits (loans), as recommended in MFSM.

Prior to January 2003, banks’ positions on financial derivatives were maintained offbalance sheets (these positions are reportedly relatively small). Starting in January 2003, in line with the International Accounting Standards (IAS), positions on financial derivatives are being included in the balance sheets. In monetary statistics, in line with the ECB regulations, financial derivatives are being classified in other assets/liabilities. Additionally, starting in January 2004, these data will be disaggregated by currency and sector of the counterparty as recommended in MFSM.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks

The Bulgarian Accountancy Law22 (November 2001) prescribes that all financial assets and liabilities be reported at market price. Whenever market price is unavailable, a measure of a fair value has to be calculated and used in the valuation. Valuation at market price or a fair value is applied to all financial instruments except currency and coins, deposits, loans, and capital and reserves, which are maintained at book value. These valuation practices conform with the MFSM recommendations.

The book value of loans and deposits denominated in foreign currency, is converted into leva using the official exchange rate23 (along with the relevant daily cross rates) prevailing on the balance sheet date. The official exchange rate is also used (along with the relevant daily cross rates) for the conversion of all other foreign-currency-denominated positions into BGN for the end-period balance sheet. The valuation adjustment is included in the Financial Result account, which is part of the Capital and Reserves aggregate (the valuation adjustment is not separately identified). The nonfinancial assets are carried in the balance sheets at their acquisition cost rather than at market value.

Monetary gold is not necessarily valued at market prices. Article 28(6) of the Law on BNB prescribes that monetary gold be valued at DM 500 per troy ounce and, since January 2002, at BGN 500 per troy ounce or the market price whichever is lower. Commodity gold is valued at market prices.

2.4.2 Recording is done on an accrual basis

Article 4(1) of the Bulgarian Accountancy Law prescribes recognition of the accruing interest. In monetary statistics, accruing interest is incorporated in the outstanding position on the underlying instrument, as recommended by MFSM. Accrual of interest in the balance sheets stops on instruments that are 90 days past due in service. The detailed data, by instrument and sector, on accrued interest and interest arrears are also presented in the memorandum items tables in the BNB bulletins.

Financial transactions are recorded on the settlement date (in line with the ECB’s guidelines), which may marginally differ from the change of ownership date, as recommended in MFSM.

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

As recommended in MFSM, source data are collected on a gross basis. In particular, claims on a client or a group of clients are not netted against the liabilities to that client or a group of clients. Gross data on assets are collected, and the various provisions (including the loan loss provisions) are classified in the Other liabilities aggregate.

Within the central bank, the ODC, and the broad money surveys, the stock data on financial assets and liabilities are aggregated into major analytic categories: on the asset side, by claims on the debtor sectors, and by the various measures of liquidity on the liability side. The underlying component data, by economic sector and financial instrument, are also shown.

The ODC survey (the Commercial Banks’ Analytical Reporting table) includes in other assets and in other liabilities, respectively, the gross claims on and gross liabilities to banks. These interbank positions can easily be netted in order to conform with the MFSM recommendation that, within a sectoral survey, such positions be consolidated. (This operation is straightforward because in the above-mentioned table the interbank positions, by financial instrument, are clearly and separately identified.)

Within the broad money survey, data on the financial relationship between the BNB and the commercial banks are consolidated, and the net amount is separately identified in other items net. Whenever net monetary aggregates are shown in the broad money survey (viz., net foreign assets or claims on general government net of central government deposits), the underlying gross data are also presented.

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 central bank and the commercial banks produce a complete set of balance sheets that are reported to the BNB for the statistical compilation program. At present, data on the detailed, 4-digit balance sheet accounts for the BNB and the commercial banks constitute the source data for compiling monetary statistics. The balance sheet data are comprehensive and contain sufficient information for classification by financial instrument, economic sector of counterparty, currency of denomination, and maturity. Information available at the BNB’s Banking Supervision Department and in other sources is used only for validation and crosschecking.

Currently the BNB is developing a new reporting system for the commercial banks which will be implemented in 2004. This reporting system will comprise the report forms and the methodological instructions for compiling data needed to fully implement the ECB’s framework for monetary data as well as for meeting the MFSM requirements.

Developments in the financial sector are closely monitored to assure that the data collection system remains comprehensive. Key available sources, including those in the Banking Supervision Department and the financial press, are monitored for relevant information. In particular, the evolution of the money market funds, which at present are relatively small, is followed so as to capture them in the coverage of monetary statistics, in due course.

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

Monetary statistics are derived from the BNB’s balance sheets and the balance sheets reported by the commercial banks. As noted in section 3.1.1, the currently available source data provide sufficient information for classification of the balance sheet accounts by financial instrument, economic sector of the counterparty, currency of denomination, and maturity. Therefore, there is no need for auxiliary data, approximations, or other adjustments.

3.1.3 Source data are timely

The accounting records of the central bank and the commercial banks are timely. The BNB’s end-week balance sheet is available on a Monday following the week’s end. The weekly commercial banks’ aggregated balance sheet is available on the second business day following the week’s end. The BNB’s and the aggregated commercial banks’ monthly balance sheets are available, respectively, on the first business day and the 15th calendar day following the month’s end. Should a delay in reporting occur, the BNB would immediately follow up by contacting the nonreporting bank. All data reporting is by electronic means, via a secure network.

3.2 Statistical techniques

3.2.1 Data compilation employs sound statistical techniques

The BNB receives data from the commercial banks via a VPN (Virtual Private Network) interbank electronic system. Each bank reports data on the same set of 4-digit balance sheet accounts. Therefore, the structure of reported data is uniform across all reporters. Aggregation of the reported balance sheet data is performed electronically and is robust.

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

The balance sheet data are not subject to statistical adjustments. Seasonal adjustment on certain monetary aggregates is performed for internal analytical purposes only, and these seasonally-adjusted monetary aggregates are not part of the officially published monetary statistics. (Following the current internal development stage, and if proved to be of analytical value, the seasonally-adjusted monetary aggregates may become part of the officially published statistics.)

3.3 Assessment and validation of source data

3.3.1 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 planning

No census or sample survey or secondary sources data are used for compiling monetary statistics. Balance sheet records (at the 4-digit level of accounts) constitute the source data for compiling monetary statistics. Procedures are in place for source data validation. These procedures are performed when the balance sheet data are first reported to the BNB. The data are subject to preliminary checks and an error-check log is automatically generated. Follow-up queries to resolve outstanding issues are undertaken by MBSD and the Banking Supervision Department. As part of the validation procedures, growth rates for all positions are calculated, including the interbank positions for each commercial bank. The interbank positions are also subject to verification across the banks. Once the reported source data have passed the validation procedures, they are made available for statistical compilation.

A new information technology system, ISMS (Information System for Monetary Statistics), is under development. It will provide the means to conduct additional verification checks. This system will also enable implementation of a coding system similar to GESMES and thus will facilitate data reporting in the GESMES format.

3.4 Assessment and validation of intermediate data and statistical outputs

3.4.1 Main intermediate data are validated against other information where applicable

Monetary statistics are derived from the balance sheet records and satisfy internal constraints imposed by the structure of the double-entry system of bookkeeping. Cross-checks are carried out against data reported to the Banking Supervision Department and on the interbank positions. Data on banks’ positions with nonresidents are regularly provided to the Statistics Directorate’s Balance of Payments and External Debt Division for use in compiling the balance of payments statistics (BOP). This practice engenders a consistency between monetary and BOP statistics as concerns source bank data. The detailed weekly balance sheet data on positions of the various general government units with the BNB are regularly provided to MOF for verification with the data provided to MOF by these units. Further, the detailed monthly data on positions of the various general government units with the banking system are also provided to MOF for verification. Consultations take place to discuss and resolve discrepancies, if they arise.

3.4.2 Statistical discrepancies in intermediate data are assessed and investigated

After balance sheet records have been validated they are ready for use as the basis for compiling monetary statistics (also see section 3.3.1). The MBSD identifies anomalous data behavior in the detailed monetary accounts (e.g., unusual growth patterns) and contacts the banks to investigate its sources.

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

Procedures are in place for consultations with the MOF to address data issues (as noted in section 3.4.1) that may arise between monetary statistics and certain fiscal data. In general, banks’ balance sheet records reported to the BNB are considered the authoritative source for use in the relevant BOP accounts and for validating the MOF’s fiscal data compiled on the basis of reporting to MOF by the various central government spending units.

3.5 Revision studies

3.5.1 Studies and analyses of revisions are carried out routinely and used to inform statistical processes

Monthly monetary statistics are based on the final balance sheet data of the BNB and the commercial banks. The revision studies are undertaken in the event of methodological changes that affect the compilation system. One such revision was implemented in March 2002 and involved excluding positions of nonoperating banks from the data starting in December 1997. An internal analysis of this change in the coverage of the ODC sector was carried out. Further, revisions to correct erroneous data if they occur tend to be of small magnitude and rarely affect major statistical aggregates. In light of this, no other revision studies are undertaken.

4. Serviceability

4.1 Relevance

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

Established procedures are in place for periodic consultations with the BNB’s Research and Treasury Directorates, MOF, NSI, and advisory groups that include representatives from the commercial banks. The MBSD staff regularly participate in the meetings and seminars for international and regional statistical community organized by, among others, the ECB and IMF. These meetings afford an opportunity to observe developments in the statistical methodologies, share national experiences, and discuss issues particular to the national statistical compilation environment and practices.

4.2 Timeliness and periodicity

4.2.1 Timeliness follows dissemination standards

The detailed monthly central bank survey and the broad money/depository corporations’ survey are disseminated within one month following the end of the reference month. In addition to the detailed monthly surveys, the BNB compiles weekly monetary statistics. The weekly main aggregates of the central bank and the broad money survey are disseminated within five days following the end of the reference week. The dissemination of the main aggregates for the central bank and the broad money survey within five days following the end of the reference week meets and exceeds the SDDS timeliness guideline for the analytical accounts of the central bank and the banking sector, respectively.

4.2.2 Periodicity follows dissemination standards

The SDDS monthly periodicity guideline for the analytical accounts of the central bank and the banking sector (the depository corporations sector) is observed. Additionally, the central bank’s and the broad money weekly main aggregates are also disseminated, which exceeds the SDDS periodicity guideline for these statistics (the weekly statistics are more aggregated relative to the monthly accounts).

4.3 Consistency

4.3.1 Statistics are consistent within the dataset

The BNB strives to ensure accuracy of information used in compiling monetary statistics through internal consistency checks, which include steps to reconcile the available stock data.

Although there are deviations, data on the BNB’s depository liabilities to commercial banks, as shown in the central bank survey, are broadly consistent with the data on the commercial banks’ depository claims on the BNB, as shown in the ODC survey. During the twelve months to November 2002, the deviations of the commercial banks’ positions are between 10–15 percent of the corresponding positions shown in the central bank survey (and are approximately 2 percent of reserve money).

As noted in section 2.2.1, the BNB’s claim on banks include claims it still has on nonoperating banks which are excluded from the coverage of the ODC survey. Accordingly, data on the BNB’s claims on banks shown in its survey are significantly larger than the data on the commercial banks’ liabilities to the BNB shown in the ODC survey. For data since March 2002, the difference in these positions can be explained with reference to the data in the nonoperating banks’ survey (the Analytical Reporting of Nonoperating Banks table), specifically the records of their liabilities to the BNB. Prior to March 2002, the deviation between the BNB’s positions, on the one hand, and the commercial banks’ and the nonoperating banks’ positions, on the other, remains large. These deviations are attributed to the lack of data reporting by banks that had already been liquidated through court proceedings. The BNB has, however, retained claims on the banks in liquidation (the so-called non-operating banks) as well as on the liquidated banks. All these claims are fully provisioned and, effective January 2003, the majority of these claims have been written off. Consequently, starting in January 2003, a close symmetry in these BNB-commercial banks’ cross-positions is expected.

4.3.2 Statistics are consistent or reconcilable over a reasonable period of time

When methodological changes are implemented, data are revised as far back as the available detailed information permits. A revision noted in 3.5.1, which was implemented in March 2002, was carried through back to 1997.

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

The BNB’s monetary statistics are broadly consistent with other statistical systems, such as balance of payments and government finance statistics. Changes in positions on net foreign assets in monetary statistics are broadly consistent with the corresponding measures shown in the BOP. Differences exist between the two sets of statistics, but they are attributable to differences in valuation (e.g., end-period prices and exchange rates in monetary statistics compared to transactions date prices and exchange rates in the balance of payments statistics). Differences between corresponding measures in monetary statistics and fiscal accounts are attributed to the differences in the valuation of securities (e.g., market prices plus the accrued interest in monetary statistics compared to the “cash” value used in the government finance data).

4.4 Revision policy and practice

4.4.1 Revisions follow a regular, well-established, and transparent schedule

Only final data are used for compiling monthly monetary statistics. Preliminary data are only used for compiling the weekly monetary statistics for the last week of the year. Revisions are made to correct erroneous data or upon implementation of changes in the accounting standards or the statistical compilation methodology. Except for revisions in the preliminary data covering the last week of the year, the other aforementioned events do not occur on a regular, well-established schedule, and hence they do not engender a regular revisions schedule.

4.4.2 Preliminary data are clearly identified

Historical data are revised to reflect any changes in the compilation methodology whenever historical data would be affected by the change (revisions are clearly marked by symbol R). Since weekly publication precedes the monthly publication, revisions are announced for the first time in the weekly publication. Preliminary data are clearly identified by symbol P.

4.4.3 Studies and analyses of revisions are made public

An internal analysis of the revision noted in section 3.5.1 was carried out. Revisions to correct erroneous data tend to be of small magnitude and rarely affect major statistical aggregates. In light of this, no other revision studies are undertaken and made public.

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)

Monetary statistics in the general BNB Monthly Bulletin and in the BNB’s monthly Monetary Statistics Bulletin are published in a clear manner. In the general BNB Monthly Bulletin, charts accompany data tables to facilitate the analysis. Monthly monetary statistics are presented in the time series format for the past 13 months, except for the sectoral balance sheets in a matrix format, which are presented for a specific month. Extensive supplementary tables focusing on particular aggregates accompany the surveys and present more detailed information (for example, a breakdown by instrument maturity or more detailed positions with a particular sector). Comprehensive tables of memorandum items are also included in the monthly publications. Development of seasonally adjusted monetary aggregates is underway. Analysis of macroeconomic developments is included in the BNB’s semiannual and annual reports. In the near future, an analysis of monetary developments will also be included on the BNB’s website, in the monthly Monetary Statistics Bulletin, and in the weekly publication (Weekly Statistical Data on Selected Balance Sheet Indicators).

The BNB’s Weekly Statistical Data on Selected Balance Sheet Indicators provides, in a clear format, the condensed data on the main credit and monetary aggregates and the interest rates, along with the abbreviated methodological notes.

5.1.2 Dissemination media and formats are adequate

The BNB’s Weekly Statistical Data on Selected Balance Sheet Indicators is published five business days after the end of the reporting period. The BNB’s monthly Monetary Statistics Bulletin is published not later than 30 days after the end of the reporting period. The general BNB Monthly Bulletin, which includes monetary, external, fiscal, and real sector data, is issued 45 days after the end of the reporting period. All these publications are released simultaneously to all subscribers. Further, these publications are posted and are accessible through the clearly organized BNB’s website at http://www.bnb.bg.

Monetary statistics are also included in the BNB’s annual and semiannual reports. The annual report for the reporting year is usually published at the beginning of May of the following year, and the semiannual report is published in November of the current year. These publications are available in hard copy and on the BNB’s website at http://www.bnb.bg.

Since June 1999, the BNB’s website at http://www.bnb.bg has become a key source of information. The weekly and monthly data, from 1995 and onward, are available for downloading free of charge at this site. The users can retrieve data by user-specified time period, data series, and periodicity.

5.1.3 Statistics are released on a preannounced schedule

Monetary statistics are released according to a preannounced schedule. The Publications section of the General Methodological Notes in the BNB’s monthly Monetary Statistics Bulletin contains a schedule for the release of weekly and monthly monetary statistics. Further, the Press Office page on the BNB’s website provides a schedule for the release of daily, weekly, biweekly, monthly, quarterly, semiannual, and annual data.

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

The data are released simultaneously to all interested users. Government authorities and agencies (viz., ministries, departments) are not given privileged or insider access (however, inherent to the data validation process, as noted in section 3.4, some agency-specific data are exchanged between the BNB and the specific agency).

5.1.5 Nonpublished (but nonconfidential) subaggregates are made available upon request

All nonconfidential monetary data are published; therefore, no additional data are provided.

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

Monetary statistics metadata pages, including concepts, scope, classifications, data sources and definitions, are available within the GDDS information pages on Bulgaria posted on the IMF’s electronic DSBB/GDDS bulletin board. These metadata are regularly reviewed and updated. Users have easy access to these metadata via a hyperlink from the BNB’s website.

Detailed methodological notes concerning general compilation methodology for monetary statistics (namely, coverage, sectoral and instrument classification, valuation, grossing and netting procedures) and specific methodological notes, table-by-table, are included in the general BNB Monthly Bulletin, the monthly Monetary Statistics Bulletin, in hard copy and on the BNB’s website (the weekly publication contains abbreviated notes). A comprehensive methodological document providing a detailed discussion of the BNB’s compilation system for monetary statistics is prepared for inclusion in the ECB’s Methodological Manual for the Accession Countries (the last update was provided for the December 2002 issue). The BNB’s website has a hyperlink to this document posted on the ECB’s website.24

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

In addition to the detailed surveys and subsidiary tables, the BNB’s monthly Monetary Statistics Bulletin contains, at the beginning, a one-page-condensed broad money survey (the Short Monetary Survey table). It also includes informative tabular summary presentations on the monetary base and its counterpart sources, money multipliers, and measurers of money supply.

5.3 Assistance to users

5.3.1 Contact person for each subject field is publicized

Prompt and knowledgeable support is available to users of monetary statistics. The email address (mbstatistics@bnbank.org) to reach MBSD contact persons with questions and provide comments, as well as the contacts’ telephone numbers, are posted on the BNB’s website. In 2003, a FAQ (frequently asked questions) document concerning monetary statistics will be published on the BNB’s website.

A dedicated interorganizational group of experts from MBSD and the commercial banks has been set up to discuss methodological issues related to data quality. Documents explaining the project to improve quality of monetary statistics and the implementation of major requirements set by the ECB from January 2004 onward are posted on the BNB’s website, and an e-mail address (mbstatistics@bnbank.org) is provided for users’ comments.

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

A catalog of publications, documents, and other services to users is available on the BNB’s website. The prices of the statistical products and services are clearly disclosed, and assistance is provided in placing the orders.

Table 5.

Bulgaria: Data Quality Assessment Framework—Summary Presentation of Monetary Statistics

(Compiling Agency: Bulgarian National Bank)

article image

VI. Balance of Payments 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

Since 1991, the Bulgaria National Bank (BNB) has had responsibility for collecting, compiling, and disseminating balance of payments, international investment position (IIP), and external debt statistics.

The legal authority for collecting balance of payments and international investment position data is provided by 1) the Foreign Exchange Law, an instrumental regulation for the collection of statistics, adopted in December 1999 and in force since January 1, 2000, 2) Article 42 of the Law on the Bulgarian National Bank, 3) three BNB regulations,25 adopted to implement the Foreign Exchange Law and govern the detailed requirements of statistical information, and 4) the Law on Statistics of June 25, 1999, which addresses the general rules applicable to confidentiality and disclosure of individual information.

However, the Foreign Exchange Law contains many important gaps, which imply weak coverage of critical information. In particular, it does not mandate reporting transactions on accounts opened with nonresident banks before it entered into force, and does not mandate the reporting of outstanding assets with nonresidents.

The BNB staff believe that the law does not provide an adequate framework to collect all necessary balance of payments statistics. They noted that within the limitations of the existing legal framework, supplementary BNB regulations were adopted to cover some of the coverage gaps of the law.26 However, those provisions are not sufficient to redress the coverage gaps in the law. The BNB staff have advocated amendments to the law that would close these information gaps. As at end-January 2003, the proposed revision was under consideration by the National Assembly.

Within the Banking Department of the BNB, the Balance of Payments and External Debt Division (BOPED) is responsible for compiling balance of payments, international investment position, and external debt statistics. At present, the BOPED is made up of the following three sections: Methodology and Compilation, Analyses and Forecasting of Foreign Trade and the Balance of Payments, and International Investment Position sections. The BOPED is also responsible for developing methodology in relation to the relevant international standards and guidelines produced by the IMF, Eurostat, and the ECB for balance of payments, IIP, and external debt statistics. The BOPED belongs to the Statistics Directorate, together with two other BNB divisions (the Monetary and Banking Statistics Division—MBSD, and the General Economic Statistics and Publications Division).

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

Data sharing and coordination are covered by formal or working arrangements; however, commitments of each agency need some further strengthening to eliminate delays in data provision by the producing agencies.

In June 2000, the BNB and the National Statistical Institute (NSI) implemented a formal agreement for exchanging data. The agreement includes regular cooperation on techniques for data compilation and scheduling of data provision for the external sector data (e.g., monthly merchandise trade data from the NSI, results of quarterly and annual FDI, and surveys of nonfinancial companies). Since mid-2002, the BNB and NSI have felt the timeliness of trade data, as disseminated by customs, to be a source of concern because trade data were transmitted by the customs with substantial delays and were undergoing revisions growing in magnitude.

In addition to this formal arrangement, data sharing includes transmission of data from different agencies, such as: 1) the Ministry of Finance for government current and capital operations with nonresidents and debt-related transactions; 2) the central depository agency for portfolio transactions of listed companies traded on the Bulgarian stock exchange; 3) the Notary office for transactions in real estate; and 4) the Ministry of the Interior for the number of travelers to and from Bulgaria.

0.1.3 Respondents’ data are to be kept confidential and used for statistical purposes only

The BNB is mandated to maintain the confidentiality of individual institutions or transactions. The Foreign Exchange Law includes a requirement for confidentiality: Article 8.3 provides that individual data received from respondents shall not be communicated to third parties; Article 8.4 provides that external communication must bear only on aggregated data, which prevents the identification of individual persons or entities. For its part, the Law on the BNB specifies in its articles 3.2 and 23.1 that individual data shall not be disclosed. In addition, the Article 4(2) provides that certain types of information shall not be released to third parties.27

The Law on Statistics also makes the case for protection of individual data. Chapter 6 provides that statistical information is provided exclusively for the purpose of statistical compilation (Article 25), and it provides the rule for dissemination of aggregate data (Article 26). Confidentiality provisions are enforced by different ranges of fines, disciplinary sanctions applicable to staff (Article 53 and 55), and physical protection of data records (Article 27).

The BOPED follows these rules and disseminates only aggregate information for balance of payments purposes, as evidenced by the dissemination of data on combined transactions for gas and oil sectors. In addition, allocation of tasks within the division makes it difficult for a given staff member to access information not explicitly assigned.

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

Statistical reporting is mandatory under the Foreign Exchange Law (Article 7) and the Law on the BNB (Article 42). Of the two, only the Foreign Exchange Law defines the penalties for noncompliance with the reporting requirements or misreporting and provides that no secrecy provision can be raised by any bank or nonfinancial entity to oppose requested information (Article 19).

However, since BNB’s Legal Department deems these provisions impractical to enforce, the penalties for noncompliance with the mandatory reporting are not implemented. In addition, interpretation of Foreign Exchange Law is hampered by an ambiguous definition of its mandate; Article 8 provides that “the BNB and the Ministry of Finance may request from every person information which is significant for the balance of payments.” This wording may prove misleading since it refers to information presented by name, which implies that the identity of the respondent should be shown, whereas the requirement was only meant as a general provision to access detailed data for the purpose of compiling statistics.

0.2 Resources

0.2.1 Staff, financial, and computing resources are commensurate with statistical programs

Established by the BNB Management Board, the BNB Budget Committee is a standing executive body in charge of managing financing aspects, including preparation, monitoring, and reporting of the BNB budget and annual allocation to the different departments. It allocates the financial resources directorate-by-directorate, while making no specific allocation by BNB activity, task, or product. For instance, in the case of the Statistics Directorate, common budgetary resources serve both BOPED and MBSD.

At present, BOPED suffers severe understaffing in the context of BNB’s growing awareness of data quality and the pressing requirements to comply with ECB and European Union guidelines. The BOPED is composed of eight staff (including a division chief), who compile balance of payments, international investment position, and external debt statistics. In addition to these functions, the BOPED is actively involved in a number of activities relating to comparison and harmonization with European Union entities and other initiatives (e.g., the Coordinated Portfolio Investment Survey (CPIS)).

The relative shortage of staff was addressed by the Fund’s 2000 Statistics Department technical assistance mission on balance of payments statistics, which recommended that additional economists be recruited to face the BOPED mounting workload. For its part, the EC noted in its 2001 “Regular Report on Bulgaria’s Progress Towards Accession” that “the BNB balance of payments department, however, needs substantial capacity building in order to ensure permanent and detailed monitoring of the capital inflow and outflow.”28 Nonetheless, training of BOPED staff is satisfactory, since half the staff received formal training, and all received on-the-job training upon joining the division. The BOPED’s staff have a high work ethic in the face of a heavy workload.

0.2.2 Measures to ensure efficient use of resources are implemented

The limited staff resources are efficiently used. The Director of the BNB’s Statistics Department, using the proposals of the heads of the BOPED and MSBD, drafts an annual plan with a quarterly breakdown for the use of staff, financial, and computing resources. It is approved by the Deputy Governor responsible for the Banking Department and then approved by the BNB Budget Committee. The annual budget is not revised during the fiscal year; thus, it does not allow for the inclusion of new priorities or emerging issues.

0.3 Quality awareness

0.3.1 Processes are in place to focus on quality

The BNB views its role of providing balance of payments statistics as primarily to satisfy the requirements of policymakers. While BNB managers acknowledge their emphasis on quality issues, as evidenced by the implementation of the BPM5 and relevant international guidelines, they also recognize that the qualitative comparison with agencies producing statistics in countries of the European Union points to the need for further strengthening of the BOPED procedures. In particular, the authorities indicated that the quality monitoring of balance of payments statistics as a public good is primarily assigned to the balance of payments division chief, while no other internal processes emulate specific assessment and monitoring of quality within the BNB. The authorities deem the current internal processes adequate to the present situation.

0.3.2 Processes are in place to monitor the quality of the collection, processing, and dissemination of statistics

The focus of internal processes is mostly limited to ensuring only the consistency of data prior to their dissemination. It excludes other processes that may have a significant impact on the quality of the sequence for data compilation and processing, e.g., source data checks, internal monitoring of intermediate results, and comparability between sources. Regarding checks on individual data, while BNB management accepts and encourages the principle of contacting respondents for verification of their data, the BOPED workload allows little interaction with respondents. Regarding checks of aggregate results, the formal and analytical checks routinely run by BOPED staff include formal correctness, logical coherence between main items and their subcomponents for aggregate results, and comparison between the behavior of different time series. Final results are ultimately validated by the division chief and by the director of the Statistics Directorate.

0.3.3 Processes are in place to deal with quality considerations, including trade-offs within quality, and to guide planning for existing and emerging needs

At the BNB-management level, a project for substantial improvement of the quality of balance of payments statistics is expected to start after more human and other resources are allocated to BOPED. As yet, a precise assessment is to be made of the trade-offs involved in the transition to detailed reporting and the more intensive use of direct reporting.

1. Integrity

1.1 Professionalism

1.1.1 Statistics are compiled on an impartial basis

Chapter 6 of the Law on Statistics stipulates that the collection of individual data is reserved only for statistical purposes unless the concerned party explicitly approves public dissemination of its data (Article 25).

Regarding the implementation of other internal measures, the BOPED work is guided solely by professional considerations, as indicated in the internal rules of procedures for staff. For instance, the methodological decisions on sources and methods, statistical adjustments, and dissemination are taken in line with the commitment of the BNB to issue statistics in accordance with the Law on the BNB.

The professionalism of the BOPED staff is supported by the recruitment procedures of the BNB. The BNB evaluates the merit and the potential of all candidates and current employees solely on professional criteria measured with objectivity.

BOPED staff indicated that comments in the press on the balance of payments statistics are closely monitored, and that the feedback received from users in the media provides evidence that balance of payments statistics are credited with high professionalism.

1.1.2 Choices of sources and statistical techniques are informed solely by statistical considerations

The choice of methods and statistical techniques is informed by the implementation of BPM5 and other relevant guidelines issued by the European institutions (Eurostat and the European Central Bank—ECB). Different types of sources are assessed against statistical needs when alternative sources are available, as evidenced when a recent survey on FDI was designed.

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

The BNB is entitled to react when commentators or media erroneously interpret the statistics. The BOPED is keen to maintain a good reputation through building up trust with the respondents and the users.

1.2 Transparency

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

The terms and conditions under which statistics are collected, processed, and disseminated are adequately made available to the public on the BNB Internet site.

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

Reference to prerelease access by the MOF and other government agencies is made only on the Dissemination Standards Bulletin Board (DSBB) Internet pages for Bulgaria; no other reference is made in BNB hard copy or electronic dissemination. Government agencies, primarily the MOF, receive balance of payments statistics 24 hours before the public dissemination of the information.

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

The identification of the statistical products is made clear in the press releases and in all subsequent documents that disseminate the data. As mentioned on page 2 of the BNB Monthly Bulletin, the BNB press office requests attribution when data are used or reproduced.

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

The BNB provides advance notice in the BNB Monthly Bulletin two months before it introduces changes in methodology. The main reasons for statistical changes are made clear to users.

1.3 Ethical standards

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

Guidelines for BNB staff behavior are specified in the BNB Rules of Procedures and are well known. BNB staff are provided access to this document through the BNB Intranet. These rules are also used for the orientation of new employees by the BNB’s Human Resources Directorate.

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

Balance of payments statistics are compiled in line with the concepts and definitions set out in the Balance of Payments Manual, fifth edition (BPM5). Regarding definitions, the critical departure from BPM5 concerns the simultaneous use of two distinct residency criteria in the collection of balance of payments source data.29 In addition, foreign exchange deposits of resident households with resident commercial banks are erroneously included in balance of payments statistics.30

2.2 Scope

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

In principle, Bulgaria’s balance of payments cover all transactions of residents vis-à-vis the rest of the world. However, a large number of items are seriously undercovered, chiefly shuttle trade and travel services. It was noted by the IMF’s balance of payments technical assistance report in 2000 that, while the magnitude of shuttle trade is deemed to require serious assessment, no resources have been allocated to this area and no subsequent adjustment has been made to trade data. In 2000 the BNB envisaged establishing a working group with NSI and other interested parties to assess shuttle trade and adjust trade records; however, so far no progress has been made. The authorities have acknowledged that deficiencies in trade data were critically impeding further harmonization work as requested for accession to the EU.

Regarding travel services, source data are mostly based on surveys conducted among travel agencies. In principle, the surveys are annual; however, there were such surveys only in 1999 and 2001. Data were collected from about 100 travel agencies, for a total coverage of an estimated 75 percent of the sector, for the 2001 survey. In addition, other sources provide qualitative information that is used for the purpose of travel. As yet, no supplementary information is collected from foreign currency bureaus on foreign exchange to estimate cash transactions by tourists; absent that information, the compilation techniques currently used suggest that the travel receipts are probably underestimated.

2.3 Classification/sectorization

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

The classification system used in Bulgaria’s monthly transactions data for institutional units and transactions is broadly consistent with the structure of BPM5, with the following important exceptions:

  • In the analytic presentation used for dissemination of balance of payments data, goods data do not distinguish between (1) general merchandise; (2) goods for processing; (3) repairs on goods; (4) goods procured in ports by carriers; and (5) nonmonetary gold. Instead, all these items are presented as credits or debits of goods. They should be disseminated using the standard components rather than the analytical presentation;

  • When a central depository agency reports financial flows, the ITRS does not separately identify financial commissions;

  • When the f.o.b. value is calculated on imports, insurance services on merchandise trade are not separated from merchandise value;

  • Financial derivatives, reported separately, are still included in portfolio investment. There are no plans to disseminate them separately;

  • Trade credits between affiliates are reported under trade credits in other investment and not under foreign direct investment (FDI) flows and stocks in the IIP because detailed information is missing. Consequently, income on trade credit is not separated from other investment income;

  • Other investment income excludes the interest share on financial leasing payments; and

  • Loans between affiliates are recorded under other investment–loans and not under direct investment–other capital.

2.4 Basis for recording

2.4.1 Market prices are used to value flows and stocks

Transactions are recorded in broad consistency with the BPM5 methodology. Balance of payments statistics are expressed in euro, leva, and U.S. dollars and are valued at market prices where they are available. Outstanding amounts in debt instruments are valued at nominal value while flows are valued at market prices. Exports and imports of goods are valued on an f.o.b. basis (imports are adjusted to exclude transport, but not insurance, however). Where transactions are derived from outstanding stocks of financial claims and liabilities at the beginning and the end of the recording period, adjustments are made to isolate actual transactions from valuation and other changes affecting the stock data.

Consistent with the BPM5 methodology, the BNB has begun adjusting changes in holdings of international reserves for valuation changes, owing to exchange rate changes (backward revision of data was made to 1998). The main exception to the market valuation principle is the recording of foreign direct investment stocks, using book value as reported by banks and other reporters. However, the quarterly FDI survey conducted by the NSI, and data reported by the privatization agency and by the Notary office, are valued either at market prices or at a proxy for market prices.31

For their part, currency conversion practices are broadly consistent with the BPM5 methodology. Transactions taken from the international transactions reporting system (ITRS)—mainly services, compensation of employees, and some private transfers—are recorded at the time of payment. Imports and exports from customs records are converted by the reporting banks from the transaction currency into the unit of account at the relevant average market exchange rate, unless the actual exchange rate of the transaction is reported.

Reporting instructions to banks prescribe the use of the average market rates in leva and in U.S. dollars to report monthly aggregated transactions in the presentation of the standard component. As financial flows are derived from changes in stocks for portfolio and other investment, data reported in foreign currencies are converted using the monthly average exchange rate, in conformity with BPM5 prescribed guidelines.

Also in conformity with BPM5 methodology, data obtained from external debt records—all of which report data in the original transaction currencies—are also converted using the average monthly exchange rates for the reporting period. The recording of arrears is valued at book prices and adjusted to take into account accumulated interest.

2.4.2 Recording is done on an accrual basis

Recording is mostly done on a cash basis, and no subsequent adjustment is made to achieve accrual recording.

In case of the ITRS, two reasons explain the large use of cash basis data: (1) the use of settlements-based reporting means that transactions are recorded at the time of banking intermediation. Most data are therefore on a cash basis, for example, income on debt for both direct and portfolio investment is reported when paid;32 and (2) given the legal arrangements for reporting introduced by the Foreign Exchange Law, the accrual principle is not mandated, although it is deemed by BOPED to be more frequent on liabilities than on assets. Assessing the respective parts of reports, using either the cash or the accrual basis in the ITRS, proves difficult to BOPED.

Regarding survey sources, data are mostly communicated in stocks (from which flows are derived) and are therefore mostly based on accrual.

Regarding other sources, customs data are based on the Special System of Trade. Effective January 1, 1999 the time of recording for customs data reflects the date on the Single Administrative Document. For administrative sources, there is no correction to adjust data to an accrual basis; for example, data on interest payments on government debt reported by the MOF are also reported on a cash basis. Private debt is to be reported in accrual; however, it is believed that the respondents do not carry it out appropriately.

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

Credits and debits are mandated from respondents in the reporting regulations in gross figures, for both the credits and the debits in the current account. Financial account statistics are reported net to the BNB for each category (e.g., credits and debits are not reported separately), except for the FDI where both credits and debits are reported. The international reserves are collected for assets only.

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

Bulgaria’s balance of payments statistical system relies on a comprehensive framework of data collection that includes the international transaction reporting system (ITRS), direct reporting, surveys, and the use of administrative data sources within and outside the BNB.

The ITRS is the primary data collection system used for compiling balance of payments statistics. It is based on aggregate monthly transactions reported following the standard component of the BPM5 both in euros and U.S. dollars. The ITRS was introduced in 1996 and covers reporting from 1) 35 commercial banks33 and, in application of the Foreign Exchange Law, 2) 21 enterprises that hold bank accounts with nonresident financial entities and report directly to the BOPED. ITRS data are mostly used to record data in the categories of services and nongovernment transfers. In the financial account, ITRS data are used for Bulgarian FDI abroad, portfolio investment assets and other investment assets, and liabilities for the commercial banks.

However, because the ITRS is viewed by the authorities as insufficient, they plan to migrate to transaction-by-transaction reporting before the accession to the EU in 2007. In addition, the authorities recognize that the present reporting by enterprises could be further improved by requesting more detailed information from reporting enterprises.

The ITRS information is automated through banks (electronic forms submitted by email) and largely based on manual recording for enterprise reports (hard-copy questionnaires). Both groups of reporters use BPM5 economic codes to classify transactions.

For the reporting format of the banks, the classification established in 1996 was reviewed in 2000. It mandates the reporting of all transactions above the 5000 leva threshold, both in U.S. dollars and leva, including reporting of financial derivatives. Nonfinancial companies report on less detailed forms, whose classification follows the main items of the standard components. In this latter case, nonfinancial companies report initial and end positions and flows quarterly and annually.

ITRS data, however, are not used for the compilation of the following items: (1) exports and imports data based on customs declarations, (2) the result of the annual tourism survey, used for estimating travel services, (3) transport services, which result from calculations based primarily on trade data, (4) the foreign direct investment surveys, managed by the NSI and the BNB, and (5) the result of the quarterly data from the register on the private debt, which provides information on specific items in the financial account.

Furthermore, the ITRS sources are hampered by several deficiencies, such as:

  • It is based on the Foreign Exchange Law, which is limited in scope to banks and to enterprises that opened accounts with nonresident financial institutions after January 2000; those accounts that were opened before that date are not covered;

  • When banks submit reporting forms, the central bank considers that the proportion of forms missing the economic code to be about 1 percent of the total current account.34 If the transaction is below the reporting threshold of 5,000 leva (approximately US$2,500), then the economic code is not required for individual operations. If it is above that amount, then the code is required, but the information may be missing. The BOPED classifies unallocated data in their totality under other investment, other liabilities. Consequently, the magnitude of data not classified and reported under other investment is significant for that item (50.6 percent of the total for other investment–other liabilities during the 11 cumulated months January through November 2002); and

  • Electronic submission of data is mostly limited to banks; hence, the amount of information that requires manual recording, from companies and most of the administrative data, is significant.

To facilitate more comprehensive coverage and classification, particularly in the financial accounts, two annual surveys are conducted.

The survey on foreign direct investment abroad is based on reports from 80 nonfinancial enterprises. The initial listing of target companies was provided in 1999 by the Ministry of Economy and covers companies that held FDI abroad before 1990. For foreign direct investment in Bulgaria, the NSI runs a survey whose results are communicated to the BNB; BOPED supplements this information with seven other sources.35

However, given these institutional arrangements, no FDI register is maintained and no sampling techniques are used by BOPED. Thus the authorities are aware that the present coverage of the FDI survey is incomplete. The revision of the Foreign Exchange Law is deemed a prerequisite to improve the coverage of the current FDI survey. Regarding the periodicity of the FDI survey, although initially designed as an annual survey, it was run only twice in 1999 and 2000 on the outstanding FDI stock for 1998 and 1999 and flows for 1999 and 2000. For 2003, there are plans to collect stock and flow data for 2001 and 2002 and to follow up on FDI operations announced in the press.

Regarding the travel survey, as indicated above (see indicator 2.2.1), the approach adopted is not systematic and allows for large undercoverage. The survey administration does not comprise any benchmarking or register follow-up. Other qualitative information is used to identify the purpose of the travel: the Border police provide monthly information on the number of travelers to and from Bulgaria (used to derive the number and the purpose of travel along six main categories), and the ITRS provides information from banks on travel motivated by education and health purposes.

For both surveys, no registers and no benchmarks are employed.

In addition to the ITRS and to surveys, BOPED uses other sources, including administrative data sent on hard-copy forms, which all require manual recording by BOPED. These sources include:

  • customs data for transaction in goods, which replace the ITRS data;

  • transfers by the EU of funds communicated by the MOF;

  • external debt-related data reported by the Ministry of Finance;

  • for the compilation of current transfers, data communicated by the Bulgarian Red Cross—the agency for foreign aid (received aid);

  • the Monetary Banking Statistics Unit in BNB, which reports data for other investment–other foreign exchange deposits–other sectors (household deposits with domestic banks) and BNB accounts abroad with other central banks;

  • the BNB Fiscal Services Department, which transmits data on nonresident holdings of government, securities issued under Regulation 5, and the ZUNKs;36

  • the BNB Information Technology Directorate, which reports foreign assets and assets and liabilities of the BNB and commercial banks;

  • for the purpose of completing the FDI transactions, the Notary office, which communicates data on the purchase of real estate property in Bulgaria by nonresidents, and the privatization agency, which provides information on the successful nonresident bidders of privatized enterprises;

  • the register on financial loans maintained by the BNB on private sector debt, which provides quarterly data, recorded under relevant items in the financial accounts;

  • portfolio investment transactions, reported by the Central Depository Agency when transactions are recorded between residents and nonresidents; and

  • the Insurance Supervision and Pension Insurance Directorates of the Ministry of Finance, which report respectively data on nonresident holdings in equity capital.

Developments in the external sector are monitored through the follow-up of the financial press information on major transactions, especially for the follow-up of foreign direct investment operations. The monitoring and drafting of research papers is severely constrained by the limited number of staff. Discussions are informally held with market participants.

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

Definitions, scope, and classifications used in source data are broadly in line with BPM5, and adjustment procedures, including the use of supplementary information, exist to ensure that source data conform to the coverage, classification, valuation, and timing conventions of BPM5. In particular:

  • Foreign direct investment: The BNB supplements the data received from the NSI with other sources for flows (see 3.1.1).

  • Trade statistics: Data received by the BNB comprise transactions in natural gas, electricity, military equipment, gold as commodity, barter trade, financial leases, direct investment in kind, fuel and food sold to foreign carriers, and foreign aid in kind (official and private). No other sources are used.

  • C.i.f./f.o.b. correction: The correction is made using a coefficient of 8 percent on imports, while no correction is made on exports, provided f.o.b. by the NSI. The adjusted amounts are subsequently