|
For any question on data and metadata, please contact: Eurostat user support |
|
|||
1.1. Contact organisation | Eurostat, the statistical office of the European Union |
||
1.2. Contact organisation unit | D1: Excessive deficit procedure, methodology and GFS |
||
1.5. Contact mail address | 2920 Luxembourg LUXEMBOURG |
|
|||
2.1. Metadata last certified | 24/04/2024 | ||
2.2. Metadata last posted | 24/04/2024 | ||
2.3. Metadata last update | 24/04/2024 |
|
|||
3.1. Data description | |||
The data correspond to quarterly non-financial accounts for the general government sector which are conceptually consistent with the corresponding annual data compiled on a national accounts (ESA 2010) basis. All data is at current prices. Both non-seasonally and seasonally and calendar adjusted data is collected. Data for the subsectors of general government (S.1311 central government, S.1312 state government, where applicable, S.1313 local government, S.1314 social security funds, where applicable) is also collected. Data coverage is best for general government non-seasonally adjusted data. |
|||
3.2. Classification system | |||
See European System of Accounts (ESA2010) |
|||
3.3. Coverage - sector | |||
General government sector (S.13), which according to the ESA2010 includes: |
|||
3.4. Statistical concepts and definitions | |||
The data correspond to quarterly non-financial accounts for the general government sector which are consistent with the corresponding annual data compiled on a national accounts (ESA2010) basis. The domain presents main aggregates (transactions and balancing items) for the general government sector. A number of countries additionally provide data for the subsectors of general government (central government, state government, local government, social security funds). A number of countries also provide data on taxes received by the institutions of the EU (S.212). The indicators are reported under Table 25 'quarterly non-financial accounts for general government' on the basis of a gentlement agreement with Member States in the Financial Accounts Working Group. For definitions of the transactions, see European system of accounts, 2010 edition (ESA2010). The following indicators are available: The following national accounts' indicators are collected: P.11 + P.12 + P.131 - Market output, output for own final use and payments for non-market output - Net lending (+) / Net borrowing (-) (B.9) is the difference between general government revenue and expenditure. Furthermore, due to a seasonal pattern of taxes, other revenues, and certain expenditure items the quarterly evolution is volatile and country specific. Therefore, users should be cautious in its analysis before making any extrapolation or drawing conclusions based on its quarterly evolution. Users should be aware that the amplitude of the seasonality and the nature of seasonality varies considerably across countries. Quarterly government finance statistics are reported to Eurostat in the form of non-seasonally adjusted (raw) figures. However, a certain number of the reported series contain seasonal patterns (partly explained by the link with the seasonality of economic activity and by the budgetary planning and accounting practices of national governments), which make it difficult to carry out a direct meaningful cross-country and time series analysis. The same is true for GDP, which reflects the seasonal pattern of all economic activities in the economy. To overcome this difficulty and thus to gain a better understanding of trends in addition to the non-seasonally adjusted data, seasonally adjusted data is presented for the EU Member States and EA-19 and selected countries. The seasonal adjustment for total revenue and total expenditure is done using an indirect procedure (at country level using Tramo-Seats in Demetra+).
Where available, National Statistical Institutes own estimates are used as input for the aggregates, which are supplied to Eurostat on a gentlemen's agreement basis. Some country level estimates as well as data for the EU aggregates are published on Eurobase. These estimates are supplemented by Eurostat's own estimates for those countries, which do not yet supply their own estimate. This data is labelled confidential and not published.
Net lending (+)/ net borrowing (-) is derived indirectly from the accounting identity:
As concerns GDP, no independent estimate is derived. |
|||
3.5. Statistical unit | |||
Institutional units and groupings of units as defined in ESA2010. |
|||
3.6. Statistical population | |||
Target population is the general government sector (S.13) as well as its subsectors. |
|||
3.7. Reference area | |||
EU and euro area aggregates, EU Member States and EFTA countries. |
|||
3.8. Coverage - Time | |||
The length of the time period varies across countries. |
|||
3.9. Base period | |||
Not applicable. |
|
|||
Data are expressed in millions of Euro, millions of national currency units and as percentages of GDP. For euro area countries, for reference periods prior to accession of the country to the euro area, data in national currency are expressed in euro-fixed, that is the former national currency divided by the irrevocable exchange rate. |
|
|||
The reference period is the quarter. |
|
|||
6.1. Institutional Mandate - legal acts and other agreements | |||
Gentlemen's agreement with Member States. |
|||
6.2. Institutional Mandate - data sharing | |||
Not available. |
|
|||
7.1. Confidentiality - policy | |||
Regulation (EC) No 223/2009 on European statistics (recital 24 and Article 20(4)) of 11 March 2009 (OJ L 87, p. 164), stipulates the need to establish common principles and guidelines ensuring the confidentiality of data used for the production of European statistics and the access to those confidential data with due account for technical developments and the requirements of users in a democratic society. |
|||
7.2. Confidentiality - data treatment | |||
Confidential data is flagged 'C'. Only authorised staff have access to the database. They have signed appropriate documentation on handling of confidential data. There are blocks which prevent the inadvertent extraction and publication of confidential data. |
|
|||
8.1. Release calendar | |||
The data are received at by t+3 months after the end of the reference quarter from reporting countries (e.g. European Union Member States and EFTA countries). They are then processed and released gradually up to around t+112 days after the end of the reference quarter. Quarterly government deficit is published in a euroindicator news release. The planned release dates are thus published on the Eurostat website (release calendar). Revisions received by Eurostat in between the main release dates are processed and published. |
|||
8.2. Release calendar access | |||
See 8.1. |
|||
8.3. Release policy - user access | |||
In line with the Community legal framework and the European Statistics Code of Practice Eurostat disseminates European statistics on Eurostat's website (see item 10 - 'Accessibility and clarity') respecting professional independence and in an objective, professional and transparent manner in which all users are treated equitably. The detailed arrangements are governed by the Eurostat protocol on impartial access to Eurostat data for users. Pre-release is granted to DG ECFIN and the ECB in line with existing MoU and SLA as well as the impartiality protocol. |
|
|||
The data are disseminated at least each quarter. |
|
|||
10.1. Dissemination format - News release | |||
Not available. |
|||
10.2. Dissemination format - Publications | |||
Additional data is published in Statistics Explained. |
|||
10.3. Dissemination format - online database | |||
Please consult data on-line or refer to contact details Data access path: |
|||
10.4. Dissemination format - microdata access | |||
Not applicable (aggregated data, any microdata received in the course of validation would be strictly confidential). |
|||
10.5. Dissemination format - other | |||
Not available |
|||
10.6. Documentation on methodology | |||
The methodological framework comprises: - European System of Accounts, 2010 edition (ESA2010) - Manual on Compilation of Taxes and Social Payments on a Quarterly Basis (2002) - Manual on quarterly non-financial accounts for general government - 2011 edition. However, it relates to data collection under ESA95. An update referring to ESA2010 methodology is planned. Manuals are available on Eurostat's website. |
|||
10.7. Quality management - documentation | |||
The European Commission (Eurostat) submitted in 2006 a report to the European Parliament and Council assessing the reliability of quarterly data delivered by Member States. This quality report was updated in 2008, on the basis of recent developments and improvements in data quality achieved since the publication of the first report in 2006. The updated quality report is available on the Eurostat website. |
|
|||
11.1. Quality assurance | |||
Not available. |
|||
11.2. Quality management - assessment | |||
The overall quality of the QNFAGG data is considered to be very good. All important aspects of quality were covered in the previous quality report released in 2006. This document served as an incentive for further improving the QNFAGG data quality. The quality of data improved, over the years following the publication of the first quality report, as shown in the update of quality report, released in July 2008. |
|
|||
12.1. Relevance - User Needs | |||
QNFAGG data users are classified as follows: - General users, general public who access to the data through the Eurostat Web Site (Eurobase) or through the Web Page of the Specific Section on Government Finance Statistics - National Statistical Institutes (NSIs) - The Directorate-General Statistics of the European Central Bank (ECB) - European Commission, Directorate-General of Economic and Financial Affairs (DG ECFIN) - Quarterly Sector Accounts (QSA) compilers within Eurostat, who use QNFAGG data as input for the compilation of QSA. - The units within Eurostat in charge of data for the excessive deficit procedure (EDP) |
|||
12.2. Relevance - User Satisfaction | |||
Assessed through regular meetings with key users. Permanent contacts are maintained with advanced institutional users (e.g. the European Central Bank, The Economic and Financial Affairs Directorate-General of the Commission). |
|||
12.3. Completeness | |||
QNFAGG data completeness is good, in line with a gentlemen's agreement with Member States. |
|
|||
13.1. Accuracy - overall | |||
Not available. |
|||
13.2. Sampling error | |||
Not available. |
|||
13.3. Non-sampling error | |||
Not available. |
|
|||
14.1. Timeliness | |||
Quarterly data are to be delivered three months after the end of the quarter to which the data relate. Quarterly data are released at around t+112 days after the reference quarter. |
|||
14.2. Punctuality | |||
Punctuality of data transmission is considered very good under the terms of the Gentlemen's agreement with Memeber States. For further details, see also "Quality report on QNFAGG - Update of July 2008", paragraph 1. |
|
|||
15.1. Comparability - geographical | |||
Comparability between countries is ensured by the implementation of ESA2010 rules. |
|||
15.2. Comparability - over time | |||
Many Member States have reported no breaks in their time series. |
|||
15.3. Coherence - cross domain | |||
The following consistency checks with other datasets are carried out:
For further details, see also paragraph 5 of the last update of the quality report mentioned in Section 11.2 |
|||
15.4. Coherence - internal | |||
Provisional data are checked with final data. |
|
|||
Not available. |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
17.1. Data revision - policy | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
To further specify the general Eurostat revision policy, the following revision policy has been established for government finance statistics. Revision policy is set at the level of national authorities. In general, the data are revised for the latest years according to change from preliminary to half-finalised and final data sources. The complete time series can be revised due to changes in the methodology or methods of data compilation, correction of errors or in case of major and benchmark revisions. Revisions are accepted at any time and following validation, data is the republished for the country and EU / euro area aggregates concerned. Revisions are broadly classified in 3 categories: - current revisions, occuring each quarter and mainly affecting the past quarters of the same year - major regular revisions taking place on a regular basis to incorporate results of changes in surveys and/or in estimation procedures, of new basic data sources, integrating the results of new censuses and/or of new estimation methods - major occasional revisions deriving from major methodological changes in national accounts, like changes in concepts and definitions and/or in the classifications used (examples are the adoption of a new accounting system - like in September 2014 the introduction of ESA2010 - or the use of a new nomenclature). |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
17.2. Data revision - practice | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Data revisions may occur at any time. Major changes in methodology are the result of legislation, and therefore announced in the Official Journal. However, some changes may be implemented beforehand on the basis of gentlemen's agreements. All reported errors (once validated) result in corrections of the disseminated data. Reported errors are corrected in the disseminated data as soon as the correct data have been validated. Data for specific countries may be published even if they are missing for other countries or flagged as provisional. They are replaced with final data once transmitted and validated. European aggregates are recalculated every time new data is published and are released simultaneously. Whenever new data are provided and validated, the already disseminated data are updated. In routine revisions, the length of the time series revised is country-specific and depends on the relevance of source data updates. . As part of routine revisions, temporal consistency (annual/quarterly) is usually established at coinciding transmission deadlines. While the revision calendar for government finance statistics is described by the scheduled releases indicated on the Eurostat website, revisions can occur at any time. Notable time series breaks caused by changes in data sources or incomplete application of a methodological change are flagged. Major revisions remove such breaks in series as far as feasible. .
National revision practices for quarterly non-financial accounts for general government:
|
|
|||
18.1. Source data | |||
Quarterly non-financial accounts for general government are primarily derived from administrative and other records of general government. Once more detailed annual information becomes available (not all units in general government may have detailed quarterly reports), the data is benchmarked on annual GFS data. |
|||
18.2. Frequency of data collection | |||
The data are collected on quarterly basis. |
|||
18.3. Data collection | |||
Reporting tables on quarterly non-financial accounts for general government are to be filled in by national authorities. Once data are compiled by national authorities in the reporting format, they are transmitted to Eurostat, using the SDMX/XML format. |
|||
18.4. Data validation | |||
Data are loaded into Eurostat Reference Database once validated by Eurostat. The validation process consists of arithmetic and quality checks as well as consistency with ESA2010 methodology. The main checks conducted are: |
|||
18.5. Data compilation | |||
Quarterly data should be based as much as possible on direct information from basic sources, shall be completed by coverage adjustments, if needed, and by conceptual adjustments in order to bring quarterly data in line with ESA2010 concepts. The quarterly data and the corresponding annual data transmitted under table 2 ‘Main aggregates of general government' of the ESA2010 transmission programme should be consistent. Data are transmitted in national currency. Eurostat converts into Euro using quarterly average exchange rates. EU and euro area series are formed by the aggregation of the country data. A Manual on Compilation of Taxes and Social Payments on a Quarterly Basis was first published in 2002. A Manual on Quarterly Non-Financial Accounts for General Government (2006) replaces the first manual providing methodological guidance and describing Member States' compilation practices for all ESA95 transactions. An updated Manual on Quarterly Non-Financial Accounts for General Government (2011 edition) was released in September 2011. |
|||
18.6. Adjustment | |||
Not available. |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
For more specific or general comments, please also refer to metadata files for gov_10a_main, gov_10a_taxag, gov_10a_exp and the latest EDP news releases for reservations or amendments to EDP data as well as general or specific notes.
NOTES ON SEASONAL AND CALENDAR ADJUSTMENT (updated on 22/04/2024) Quarterly government finance statistics are reported to Eurostat in the form of non-seasonally adjusted (raw) figures. However, a certain number of the reported series contain seasonal patterns (partly explained by the link with the seasonality of economic activity and by the budgetary planning and accounting practices of national governments), which make it difficult to carry out a direct meaningful cross-country and time series analysis. The same is true for GDP, which reflects the seasonal pattern of all economic activities in the economy. To overcome this difficulty and thus to gain a better understanding of trends in addition to the non-seasonally adjusted data, seasonally adjusted data is presented for the EU and EA and selected countries. The seasonal adjustment for total revenue and total expenditure is done using an indirect procedure (at country level using Tramo-Seats on JDemetra). Where available, National Statistical Institutes own estimates are used as input for the aggregates, which are supplied to Eurostat on a gentlemen's agreement basis. Most country level estimates as well as data for the EU aggregates are published on Eurobase. These estimates are supplemented by Eurostat's own estimates for those countries, which do not yet supply their own estimate. Such data is labelled confidential and not published. Net lending (+)/ net borrowing (-) is derived indirectly from the accounting identity: Net lending (+)/ net borrowing (-) = total revenue - total expenditure. As concerns GDP, no independent estimate is derived. EU AGGREGATES: Estimated indirectly at Eurostat on the basis of Member States' data a far as this is supplied nationally and complemented by Eurostat's own estimates, where no nationally supplied data is available. Tramo-Seats run on JDemetra is used in all cases. For the following countries, the estimates are produced by the respective National Statistical Institute, which all follow the “ESS guidelines on seasonal adjustment”: Belgium: The seasonally adjusted series are computed following an indirect approach. The components of the revenue and of the expenditure of the General Government are seasonally adjusted by means of "Tramo-Seats", taking into account the presence of possible outliers and calendar effects. The model of each component (>20) has been individually validated (no automatic modelling). The absence of residual seasonality after aggregation has been checked. The data are benchmarked on annual totals of the non-adjusted series. The annual benchmarking is computed on each component by means of a multiplicative Denton procedure. Bulgaria: Tramo-Seats on Demetra +. Total expenditure: no trading days effects, no Easter effect, log-transformation, ARIMA model [(0,1,1)(0,1,1)], outliers: AO[2007-IV], AO[2008-IV], TC[2014-IV], 1 pre-specified outlier TC[2020-II]. Total revenue: no trading days effects, no Easter effect, log-transformation, ARIMA model [(0,1,1)(0,1,1)], outliers: 1 pre-specified outlier TC[2020-II]. Czechia: Tramo-Seats on Demetra +. Total expenditure: No trading days effects, no Easter effect, ARIMA model [(0,1,1)(0,1,1)], outliers: AO[2003-I], AO[2003-III], AO[2012-IV], TC[2001-IV], LS [2016-I]. Total revenue: No trading days effects, no Easter effect, ARIMA model [(0,1,1)(0,1,1)], outliers: AO[2003-I]. P.51g: No trading days effects, no Easter effect, ARIMA model [(1,1,1)(0,1,1)], outliers: AO[2003-I], LS[2016-I]. Denmark: X13-ARIMA. Total expenditure: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,1,1) (0,1,1)], outliers: TC[2012-II], TC[2020-IV]. Total revenue: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,1,0)(0,1,1)], outliers: AO[2009-IV], LS[2019-I], LS[2015-I], AO[2009-I], LS[2020-IV]. Germany: X13-ARIMA. Total expenditure: Log-transformation, no trading day effects, ARIMA model [(0,1,1) (0,1,1)], outliers AO [1995-I, 1995-III, 2010-III, 2021-III, 2022-IV] LS [2020-II, 2020-III, 2022-I]. Total revenue: Log-transformation, no trading day effects, ARIMA model [(0,1,0) (0,1,1)], outliers LS [2020-II] AO [2021-II]. Estonia: Tramo-Seats on Demetra +. The seasonal adjustment of all time series is done with TRAMO-SEATS using JDemetra+ software. For TE and TR no calendar adjustment has been added as it does not have a notable impact on the results. According to ESS guidelines there is also no temporal consistency forced on the time series in order to provide a more purely seasonally adjusted time series for users. Ireland: JDemetra+ X13 RSA full method. Total revenue: ARIMA Model (2,1,1)(0,1,1), Outlier detected LS (I-2020), LS (II-2021) Total expenditure: ARIMA Model (2,1,1)(0,1,1), Outliers detected AO (I-2010), TC (III-2010), AO (IV-2010), AO (III-2011), LS (II-2020). General Government Surplus/Deficit using indirect approach. Spain: For P.3, OTE, OTR and B.9, the seasonally adjusted series are computed following an indirect approach. The components are seasonally adjusted using Tramo-Seats on JDemetra+ 2.2.2, taking into account the presence of possible outliers and calendar effects. The model of each component has been individually validated (no automatic modelling). The absence of residual seasonality after aggregation has been checked. The data are benchmarked on annual totals of the non-adjusted series. Following Eurostat recommendations and the practice followed in all the INE short-term statistics, an atypical type of impulse (AO) has been introduced into the series when it has been statistically significant in the data referring to the last quarter. When the data for subsequent quarters is available, we analyse whether this impulse should be modified by another type of intervention. For P.51g: The specifications can be found as part of the Standardised Methodological Report (https://www.ine.es/dynt3/metadatos/en/RespuestaDatos.html?oe=30026) France: Seasonally adjusted data is transmitted. Working day adjustment is also done when relevant. An indirect method is used. Seasonal adjustment is done using X-12-ARIMA. For more information, you can read INSEE's methodology (starting on page 21) at the following link (the document is available in English and French): https://www.insee.fr/en/information/2579410. In 2019Q1, non-seasonally adjusted data on taxes on income (D.51REC) decreases strongly due to a change in seasonality. For this reason, the evolution of the seasonally adjusted data differs significantly. Such changes in seasonality are technically complicated to model, hence the seasonally and seasonally and calendar adjusted data for 2019Q1 should be interpreted with caution. The changes in seasonality are primarily due to a new system in the collection of personal income tax (introduction of advance payments and retention at source) and the early repayment of a tax credit in January, introduced in 2019. Croatia: The direct approach and X11-Arima methods were used for seasonal adjustment. Detailed pre-treatment of the series, including graphical analysis, testing the significance of calendar (trading/working days, leap year, Easter effect, etc) was ran. The series were checked for outliers of different types. Outliers for which a clear interpretation exists are included as regressors in the model. The automatic test for log-transformation suggested the choice of the decomposition scheme. The model, filters, outliers and calendar regressors will be re-identified once a year and the respective parameters re-estimated every time new or revised data become available. Latvia: Tramo-Seats on JDemetra+ (version 2.0.0). No trading day effect, national calendar adjusted. Total expenditure: Log-transformation, ARIMA model [(0,1,1)(0,1,1)], 3 pre-specified outliers: LS[2006-IV], LS[2020-III], AO[2020-IV], ramp:[I-2009 - IV-2009]. Total revenue: Log-transformation, ARIMA model [(0,1,0)(0,1,1)], pre-specified outlier: AO[2006-IV]. Lithuania: Tramo-Seats on NbDemetra 2.2.2. Total expenditure: Log-transformation, no Easter effect, ARIMA[(1,0,0)(1,1,1)], outliers AO[2011-IV], AO[199-II], AO[2006-II], LS[2020-I], AO[2022-IV]. Total revenue: Log-transformation, no Easter effect, ARIMA[(0,1,1)(0,1,1)], no outliers. Gross Fixed Capital Formation: Log-transformation, no Easter effect, ARIMA[(0,1,1)(0,1,1)]. Luxembourg: All series are seasonally and calendar adjusted with automatic outlier detection and correction. No benchmarking or other adjustments are made. The method used is non-parametric X13 RSA5c with the Luxembourgish calendar. The software used is JDemetra+ (v2.2.2). Hungary: JDemetra+ TramoSeats method. Hungarian specific calendar is used. Working day, Easter and leap year effects are tested. Total revenue: Log-transformation, no trading day effects, no Easter effect, ARIMA model [(1,1,0)(0,1,1)], 1 predefined outlier: AO (2015-IV). Total expenditure: Log-transformation, no trading day effects, no Easter effect, ARIMA model [(0,1,1)(0,1,0)]. Malta: Total expenditure: Tramo-Seats on JDemetra+ 2.2.2, Series has been log-transformed, No trading days effects, No Easter effects, ARIMA model [(0,1,1)(0,1,1)] with a statistically significant mean, 2 pre-specified outliers: AO(IV-2003), TC(II-2020) Total revenue: Tramo-Seats on JDemetra+ 2.2.2, Series log transformation, No trading days effects, No Easter effects, ARIMA model [(0,1,1)(0,1,0)], 3 pre-specified outliers: AO(III - 2000), AO (IV - 2000), TC (I - 2020). Netherlands: X13-ARIMA on JDemetra+. Total revenue: Log-transformation, no trading day effects, no Easter effect, ARIMA model [(1,0,1)(1,1,0)], outliers; LS (I-2009), TC (II-2020). Total expenditure: Log-transformation, no trading day effects, no Easter effect, ARIMA model (0,1,0)(0,1,1)], outliers; AO (II-2009), AO (II-2020). Austria: Seasonal adjustment: Tramo-Seats on jDemetra+. Total expenditure: log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,0,0)(0,1,1)], outliers (specific pre-treatment): [2004-II], [2004-IV], [2009-IV], [2014-IV], [2015-III] , [2020-I], [2020-II], [2020-III], [2020-IV], [2021-I], [2021-II], [2021-III], [2021-IV], [2022-I], [2022-II], [2022-III], [2022-IV], [2023-I], [2023-II], [2023-III], [2023-IV]. Total revenue: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,0,0)(0,1,1)], outlier: LS[2009-II], LS[2016-I], AO[2020-II], AO[2020-III], AO[2020-IV], AO[2021-I], AO[2021-II], AO[2021-III] , AO[2021-IV], AO[2022-I], AO[2022-II], AO[2022-III], AO[2022-IV], AO[2023-I], AO[2023-II], AO[2023-III], AO[2023-IV]. Poland: Tramo-Seats on JDemetra +. Direct method used. Concurrent adjustment for Q1 each year, current adjustment Q2, Q3, Q4 (model revised once a year). Calendar effects adjustment used. Working days with leap year effect (2 regressors) and Easter effect tested for each series - only significant effects used in final specification. Automatic model selection with additional non-automatic verification of problematic cases. Total expenditure: P.2 - log transformation; mean effect, no calendar effect, ARIMA model [(0,0,0)(1,1,0)], outliers: AO(Q3_2010), AO(Q3_2022), AO(Q4_2022), AO(Q1_2023), AO(Q4_2023); P.5 - log transformation; no calendar effect, ARIMA model [(1,0,0)(0,1,0)], outliers: TC(Q1_2016), AO(Q1_2022), AO(Q4_2023); D.1 - log transformation; no calendar effect, mean effect, ARIMA model [(0,1,1)(0,1,0)], outliers: LS(Q4_2002), LS(Q2_2008), AO(Q4_2013), AO(Q4_2021), AO(Q3_2022); D.6M - log transformation; no calendar effect, ARIMA model [(0,1,1)(0,1,1)], outliers: AO(Q4_2007), LS(Q4_2004), LS(Q2_2019), AO(Q2_2020), AO(Q4_2021), AO(Q2_2022), AO(Q3_2022), AO(Q2_2023), AO(Q3_2023); D.4 - log transformation; no calendar effect, ARIMA model [(0,0,0)(0,1,1)], outliers: LS(Q4_2008), AO(Q4_2021), AO(Q2_2022), AO(Q3_2022), AO(Q4_2022), AO(Q2_2023), AO(Q3_2023), AO(Q4_2023); P.29+D.3+… - no-log transformation, no calendar effect, ARIMA model [(0,1,1), (0,0,0)]; Total revenue: D.2 - log transformation; mean effect, no calendar effect, ARIMA model [(0,1,1)(0,1,0)], outliers: AO(Q2_2004), TC(Q1_2009), AO(Q2_2020), AO(Q1_2022), AO(Q3_2022), AO(Q4_2022); D.4 – no-log transformation; no calendar effect, mean effect, ARIMA model [(0,0,0)(1,0,0)], outliers: TC(Q2_2012), AO(Q2_2020), AO(Q2_2021), AO(Q3_2021), LS(Q2_2022), AO(Q4_2022), AO(Q3_2023); D.5 - log transformation; no calendar effect, mean effect, ARIMA model [(1,0,0)(0,1,0)], outliers: AO(Q1_2020), AO(Q2_2020), LS(Q2_2009), AO(Q1_2023); D.61 - log transformation; no calendar effect, ARIMA model [(0,1,1)(0,1,1)], outliers: TC(Q1_2001), TC(Q4_2008), AO(Q4_2007), AO(Q3_2011); P.1+D.7+D9 - log transformation; no calendar effect, ARIMA model [(0,1,1)(0,1,1)], outliers: AO(Q4_2007), AO(Q4_2013), AO(Q4_2009), AO(Q4_2012). Portugal: X13-ARIMA on Demetra+. A manual pre-treatment is performed by identifying and deducting one-off measures. Additional pre-treatment is applied for outlier detection and correction. The seasonal adjustment is applied to total revenue, expenditure except compensation of employees and compensation of employees. Net lending (+)/net borrowing (-) is presented as a result of the difference between the series seasonal adjusted of total revenue and total expenditure. Total expenditure results of the sum of seasonal adjustment series of total expenditure (except compensation of employees) with compensation of employees. Total Expenditure (except compensation of employees): X-13-ARIMA; log-transformation; no trading days effects; no Easter effect; ARIMA Model [(0,1,1)(0,1,1)]; Outliers: AO (IV-2002), TC (III-2002) and SO IV [2012-2020] (user defined variable); Compensation of employees: TRAMO-SEATS; log-transformation; no trading days effects; no Easter effect; ARIMA Model [(0,1,1)(0,1,1)]; Outliers: TC (III-2005), TC (I-2013), LS (I-2011), LS (I-2012), AO (I-2001), AO (III-2014), SO II [2012-2013](user defined variable) and SO IV [2012 -2017](user defined variable);Total Revenue: X-13-ARIMA; log-transformation; no trading days effects; no Easter effect; ARIMA Model [(1,0,1)(1,1,0)]; Outliers: AO (II-2020), AO (II-2009) and SO III [1999-2008] (user defined variable). Romania: Tramo-Seats on Demetra+. P.51g series was not log-transformed and the model used was automatic Arima model. Total expenditure was log transformed and the model used was automatic Arima model. Total revenues was log transformed and the model used was automatic Arima model. B.9 is derived indirectly by the difference between seasonally adjusted series of total revenue and total expenditure. Seasonal Adjustments are made every time when Table 25 ESA are compiled. Slovenia: In September 2023 some changes in seasonally adjustment have been introduced. Time series are seasonally adjusted in two parts: with one model for the period from the beginning of the time series to the end of 2010, and with the second model for the period from the beginning of 2010 to the end of the time series. Results of seasonal adjustment are updated only for the second model. Average of both models is taken into account for 2010. JDemetra+ software (TRAMO/SEATS method) is used for seasonal adjustment. Time series from the beginning of the time series to the end of 2010: Total revenue: Log transformation, no trading days effects, no Easter effect; 2 pre-specified outliers: LS (Q1/2009), TC (Q3/2004); ARIMA (0,1,3)(0,1,1); Total expenditure: Log transformation, no trading days effects, no Easter effect; 1 pre-specified outlier: AO (Q1/2001); ARIMA (0,1,0)(0,1,1) + Mean correction; Gross fixed capital formation: Log transformation, trading days effects, no Easter effect; 1 pre-specified outlier: AO (Q1/2000); ARIMA (0,1,0)(0,1,1); Final consumption expenditure (P.3): Log transformation, no trading days effects, no Easter effect; 1 pre-specified outliers: TC (Q1/2007); ARIMA (0,1,0) (0,1,1). Time series from 2010: Total revenue: Log transformation, working days effect, holidays effect, leap year effect, no Easter effect; 2 pre-specified outliers: LS (Q1/2021), LS (Q1/2020); ARIMA (0,1,1)(0,1,1) + Mean correction; Total expenditure: Log transformation, no trading days effect, no Easter effect; 5 pre-specified outliers: AO (Q2/2020), AO (Q4/2014), AO (Q1/2013), AO (Q4/2013), SO (Q3/2020); ARIMA (0,1,0)(0,1,1); Gross fixed capital formation: Log transformation, working days effect, holidays effect, leap year effect, no Easter effect; 3 pre-specified outliers: LS (Q1/2016), AO (Q4/2019), LS (Q1/2018); ARIMA (0,1,0)(0,1,1); Final consumption expenditure (P.3): Log transformation, no trading days effects, no Easter effect; 2 pre-specified outliers: LS (Q1/2021), AO (Q2/2021); 1 detected outlier: LS (Q2/2023), ARIMA (0,1,1) (0,1,1). Slovakia: Tramo-Seats on JDemetra +. Total expenditure: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,1,1)(0,1,1)], 5 pre-specified outliers: LS(2000-IV), AO(2002-IV), AO(2015-IV), AO(2021-II), AO (2023-IV). Total revenue: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(0,1,1)(0,1,1)], 5 pre-specified outliers: LS(2001-III), AO(2015-IV), LS(2020-I), AO(2020-II), AO (2023-IV). Data are constrained to annual non-seasonally adjusted totals. Finland: The seasonally adjusted series are computed following an indirect approach. The components of the revenue and of the expenditure of the General Government are seasonally adjusted by Tramo-Seats / JDemetra+ 2.0.0, taking into account the presence of possible outliers and calendar effects. The data are benchmarked on annual totals of the non-adjusted series. The annual benchmarking is computed on each component by Denton procedure. Sweden: Tramo-Seats on JDemetra. Total expenditure: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(1,1,0)(0,1,1)], 7 pre-specified outliers: AO[1998-III], AO[2001-IV], LS[2002- II], AO[2002-III], AO[2010-IV], AO[2020-II], AO[IV-2022]. Total revenue: Log-transformation, no trading days effects, no Easter effect, ARIMA model [(1,0,0)(0,1,1)] with mean, one pre-specified outlier: TC[2020-II]. Iceland: Total revenue and total expenditure are seasonally adjusted using Tramo-Seats / JDemetra+ 2.2.3 taking into account the presence of possible outliers and calendar effects. For each subsector and S.13, total revenue and total expenditure are estimated directly, with B.9 for each subsector and S.13 derived indirectly from total revenue and total expenditure. Switzerland: The data reported is trend-cycle data. A Denton-Cholette method is used to temporally disaggregate annual data. The quarterly data is extrapolated using smoothed indicators. |
|
|||
|
|||
Manual on Compilation of Taxes and Social Payments on a Quarterly Basis (2002) Quality report on QNFAGG - Update of July 2008 Manual on Quarterly non-financial accounts for general government |
|
|||