PROPOSALS FOR IMPROVING QUALITY REPORTING

Transcription

1 EUROPEAN COMMISSION EUROSTAT Directorate F: Social Statistics and information Society UnitF-3: Living conditions and social protection statistics Doc SP/2010/05 /EN Working Group SOCIAL PROTECTION April 2009 BECH building Room Ampère Luxembourg PROPOSALS FOR IMPROVING QUALITY REPORTING 1

2 The purpose of the document To provide examples of good practice in reporting quality both for Core System and for the module on Pension Beneficiaries; To harmonize the quality reporting process across the Member States; The Member States are invited to Draw an opinion on the guidelines and recommendations made by EUROSTAT To follow as much as possible the approved guidelines and/or recommendations 2

3 Content of the document 1. The legal basis 2. Aim of the document and issues concerning quality reporting 3. Data sources reporting 4. QD specific tables best practice 5. PB specific tables - best practice 6. Reporting revisions 7. Recommendations 3

4 1. The legal basis During the years 2007 and 2008, three new regulations were introduced in the field of social protection statistics: 1. European Parliament and Council Regulation n.458/2007; 2. Commission Regulation n.1322/2007; 3. Commission Regulation n.10/2008. European Parliament and Council Regulation n.458/2007 provides: A methodological framework (based on common standards, definitions, classifications and accounting rules), intended to be used for compiling ESSPROS statistics on a comparable basis; Time limits for the transmission of statistics compiled according to ESSPROS; Commission Regulation n.1322/2007 complements EP and Council Regulation and gives: The appropriate formats for transmission and results to be transmitted; The criteria to follow in order to measure quality for the ESSPROS Core System; Time limits for the transmission of national Quality Reports; Commission Regulation n.10/2008, for the Core System of ESSPROS and for the module on Pension beneficiaries is focussing on: Definitions Detailed classifications Updating of the rules for dissemination 2. Important issues linked to quality reporting The aim of a quality report is to capture the multiple components of the quality of statistical data, respecting the criteria commonly used in assessing the quality of statistics. The European Statistical System (ESS) has defined the following quality criteria (to be applied to statistical data): Relevance, Accessibility and Clarity, Timeliness and Punctuality, Coherence, Comparability and Accuracy (see Consolidated Quality Report on ore System ) The data collection 2009 exercise of reporting quality In the last data collection all the countries belonging to EU, except for IE, sent the National Quality Report on Core System. CH, IS and NO supplied to EUROSTAT their Quality Report as well. Most of the countries belonging to EU sent their QR reports on the Core System before the deadline. As some Quality Reports arrived in a period very close to the transmission of the Core System, the information supplied by them could be used in the 2009 validation exercise for QD. 4

5 Some delays, in the analysis of the QR and in producing the consolidated Quality Report for the Core System as requested by Regulation 1322/2007, were encountered because: Some countries did not follow the latest template recommended by EUROSTAT for the structure of QR (LU, BE, EE, ES, FI, HU, mainly because they had sent in advance their QR); Some countries produced their QR in the original language (ES). Concerning the Quality Reports on Pension Beneficiaries, all countries belonging to EU, except for EL, IE and SE, sent the National Quality Report on PB. NO supplied to EUROSTAT the Quality Report as well. Most of the countries belonging to EU sent their QR report on the PB before the deadline. As some Quality Reports arrived in a period very close to the transmission of the questionnaires on PB, the information could be used for 2009 validation s exercise of data on this module. Some delays, to analyse the QR and to produce the consolidated Quality Report for the Core System as requested by Regulation 1322/2007, were encountered because: Some countries did not follow the latest template structure of QR as recommended by EUROSTAT (DK, EE, EL, ES, LU, NL and SK); Some information was provided in the original language (ES). Recommendation: Whenever possible, the national quality report should be transmitted together with the questionnaires. The aims of this document is: To provide some examples of good practice in reporting quality both for Core System and for the module on Pension Beneficiaries; To draw some short guidelines for better report the quality; To provide some recommendations useful for quality reporting. The use of the templates proposed by EUROSTAT is not compulsory and the quality reporting can be addressed by each country in a different way. However, EUROSTAT strongly recommends the use of the new templates proposed this year. A general recommendation is that whenever some information concerning quality of data is important to be reported, but the tables provided by the templates proposed by EUROSTAT are not enough complete, this kind of special information can be reported by: Reporting it inside the tables intended for comments; Reporting it as a special chapter at the end of the quality report; Short information can be reported by using footnotes. EUROSTAT encourages the use of any other ways of reporting information on quality if this is relevant for the data collection itself. 5

6 3. Data sources reporting The information concerning data sources is reported in table 1.1 of both the templates for the core system and for the module on pension beneficiaries. As the information to be reported is complex, this table is split in two parts; the first column is included in both the two parts of this table. Consequently, all the names of the data sources should be reported in the first column of the table 1.1 part 1 and copied in the first column of the table 1.1 part 2. Table 1.1 Coverage of data sources, part I The first part of this table should contain very condensed information about the names of data sources, types, frequency, time lag and schemes covered. The information supplied is supposed to be short for all the five columns of this table. During the process of validation, EUROSTAT found some information reported in this table, even if that information is intended to be reported somewhere else in the QR. Examples of redundant information in this table: Reporting additional information about data sources in the first column of the table. This is supposed to be done in the second part of the table; Reporting multiple data sources on the same row of the table. Each data source is supposed to be reported on a separate row; Reporting two different types for the same data source; Frequency or time lag reported in sentences explaining the way the information is supplied by the data source. The frequency and the time lag are supposed to be numbers; Reporting supplementary information about the schemes in the last column of the table. The last column should contain either the name of a scheme or a list of schemes. Example 1. Whenever possible, the best practice is to order the schemes on the column C5 and to report for each scheme the data source concerned (as an example, the first three rows of the table below). The columns C2, C3 and C4 should contain only the necessary information; all other information should be reported either in the second part of this table or at the end of the quality report, in the tables 4.1 or 4.2. If a scheme is covered by two or more data sources, it should be reported on different rows of the table (see rows four and five of the table below); When a data source covers several schemes, either total or partial, the last column should contain the list of the schemes covered (rows six and seven of the table below). We recommend providing this kind of lists as simple as possible, without supplementary information on them; When data is available online, the time lag is 0 months and it should be reported as such (row six of the table); 6

7 When either the data source or the schemes have associated some acronyms, the use of this jointly with the full name is advisable, especially if used also in the rest of the report (seventh row of the table); When using the first time an acronym, please use also the full name of the scheme or data source concerned. C1 C2 C3 C4 C5 Name of data source Type of data source used Frequency of the data source in months Time lag in months Schemes covered Row 1 Row 2 Row 3 Row 4 Row 5 Row 6 Row 7 Row 8 Rechnungsergebnisse der Rentenversicherungsträger Rechnungsergebnisse der Pflegekassen Rechnungsergebnisse der Krankenkassen Unemployment Insurance Fund Unemployment Insurance Register Unemployment Insurance Fund Annual Report Registers of the Social Insurance Institution The Financial Statements and statistics of the Social Insurance Institution National Authority for Disabled People (ANPH) Administrative data 12 months 12 months scheme 1 Administrative data 12 months 12 months scheme 2 Administrative data 12 months 12 months scheme 3 Administrative data Administrative data Year Year The time lag is 3 months The time lag is 4 months Scheme 4 Scheme 4 register online 1, 2, 20, 21, 32, 33, 34, 38, 40, 44, 47, 48 administrative data once a year 5 months 1, 2, 13, 14, 15, 16, 20, 21, 22, 32, 38, 40, 44, 47, 48 Administrative data 12 months 6 months Scheme 2 (PSHAND), Scheme 24 (TRANSP) Example 2. Whenever possible, instead of 'Health accounts' or 'Housing accounts' (or other types not provided in the template), is better to report either 'National accounts' or 'Administrative data' or one of the provided types; Or, at least to provide information about whether the data source is a section of National accounts or they are separate accounts. All this supplementary information should be provided in the second part of this table, namely in Table 1.1, part II, column C2. It is not advisable to report on the last column generic names for the schemes or categories of schemes (first two rows in the table below); The best is to report the schemes by their numbers; When special kinds of data sources are used, it is advisable to report like in the last two rows of the table below. 7

8 Row 1 Row 2 Row 3 Row 4 C1 C2 C3 C4 C5 Name of data source DREES (Direction of Statistics and studies at the ministry of Health and social affairs) Direction of statistics at the ministry of equipment and housing Austrian Institute for Economic Research Studies on the occupational pension scheme Austrian Institute for Economic Research) Additional information on occupational pensions. Type of data source used Frequency of the data source in months Time lag in months Schemes covered Health accounts Each year 1 year Social security schemes Housing account Each year 1 year Social security schemes Other (study). Sporadically. - Scheme 20. Other (expert opinion). 12 months. - Scheme 20. Example 3. Reporting the schemes together with their names in the original language, (first row of the table below) provides a plus of information but it is advisable that in the last column to report schemes only by their numbers and, eventually, by an acronym. No other information about the schemes in column C5 is required and/or advisable; As for the names of the data sources, they can be reported in both the original language and in English (the fourth row of the table below). C1 C2 C3 C4 C5 Name of data source Type of data source used Frequency of the data source in months Time lag in months Schemes covered Row 1 Row 2 Row 3 Information provided by Social Insurance Institution that is managing Social Insurance Fund The Norwegian Labour and Welfare Administration registers of private pension funds Administrative data 8 After adoption of the state budget law report September of the proceeding year. Administrative data - statistics 12 months Approximately 2-3 months register based data (data of different private pension funds** launching the scheme; structure of individual insurance programmes/plans is not single), the information if a beneficiary based on our requirement to get information for a reference date (31/12/2007) on specific number of beneficiaries/year (number on disability pension 9 months scheme no. 3 Scheme 1 - Fundusz Ubezpieczeń Społecznych Scheme 2 - Ubezpieczenie emerytalne Scheme 3 Ubezpieczenia rentowe Scheme 01 (The National Insurance Scheme): Includes disability pensions, old age pensions, survivors pensions (spouses and children) and pensions to residents living abroad

9 Row 4 Federal Ministry of Social Affairs and Consumer Protection (Bundesministerium für Soziales und Konsumentenschutz): Special tabulation of statutory pension insurance beneficiaries. receives a pension benefit from scheme no. 1 or 9 at the same time (eg. 1+3 and 3+9) is not available beneficiaries is not available) Administrative data. 12 months. Approx.10 months. Scheme 1. Table 1.1 Coverage of data sources, Part II This part of the table is intended to describe problems linked to every data source and the geographical coverage. As soon as a data source does not covers some benefits or other figures related to a scheme, or other problem are encountered, this should be reported and a detailed description is to be provided in Table 1. Different types of problems that can be reported in the second column ('Any additional information regarding this data source') of the Table 1.1 part II are like the following: Delays, either because the data source has a long time lag or the data source is not periodic; The schemes and the items that are uncovered by the data source and no other data source can supply the missing information; Cases of missing information for breakdowns either for a specific item (between several schemes) or for a specific scheme (between several items); Cases of data sources used only as complementary for defining structures in the data (like breakdowns); Inconsistencies between the data source and other data sources. The last column of the table below (column C3) should refer only the geographical coverage. This means that this table can contains two different kinds of coverage: In the column C2 can be reported the coverage (or lack of coverage) in terms of schemes or items; In the column C3 should be reported if the geographical coverage is for the whole country or only some specific regions are covered. Recommendation: Keep in mind that the column C3 in not intended for reporting if a scheme covers different occupational categories or other social categories. 9

10 Example 4. The table below gives an example of a description of the schemes and items within them that they are not covered. C1 C2 C3 Name of data source Federation of Austrian Social Insurance Institutions: Individual-related statistics (Personenbezogene Statistiken), in: Soziale Sicherheit. Any additional information regarding this data source (i.e. reports on problems -including delays- which lead to estimation of data) Data do not cover all beneficiaries of pensions paid by health institutions for some parts of civil servants (Krankenfürsorgeanstalten) (scheme 2: , , ). Number of double counting (civil servants survivors pension and civil servants pension) is not given (scheme 2: , , ). Geographical coverage (i.e. which parts of the country are covered by this data source) Scheme 1 and 2: Full coverage. 4. QD specific tables - best practice a) Table 1.2 of the template of quality report on core system This table is intended for the report on the schemes and/or the items that are covered by two or more different types of data sources. By now, some countries reported these cases even if the double coverage was supplied by data sources of the same type. When reporting this kind of information it is strongly advisable to provide the codes of the items together with their names; When reporting by benefits, it is also advisable to use the code and the name of the item concerned with that benefit; EUROSTAT encourages to report always the double coverage even if this is not compulsory; The last column of this table should contain always at least to different names of data sources; The last column should not contain supplementary information about the data sources or on the necessity of using them both nor to contain any other information. Example 5. The table below provides two examples of reporting cross coverage. Scheme concerned Benefit concerned Data sources cross-covering the benefits Scheme 21 (SPP) Old age pension ( ) National House of Pensions and Other Social Insurance Rights (CNPAS), Ministry of Culture, Household Budgets Survey (HBS) Scheme 1, Scheme 2, NMT Disability Pensions, NMT Old Age Pensions, Detailed Annual Report on Pension Expenditures; Register Pensions 10

11 Scheme NMT Anticipated Old Age Pensions NMT Survivors Pensions b) Table 1.3 of the template of quality report on core system. Geographical comparability This table is intended to provide information about the comparability of the final questionnaires across EU and should contain: Either scheme(s) uncovered (missed), in the first column (C1) of the table, or Uncovered item(s) within a scheme and in this case both the scheme and the item(s) should be reported in column C2, or Cases of benefits classified not in line with the methodology. These will be reported on the column C3. Example 6. If a scheme is partially covered and the percentage of coverage can be calculated or estimated, this information is to be provided in the QR (see as an example the second row of the table below) When there are no problems to be reported in this table, a mention should be made as for example those in the last two rows of the table. C1 C2 C3 Row 1 Row 2 Degree of coverage in terms of schemes) Social services provided by the non-profit organizations are not included as a scheme. Degree of coverage in terms of schemes 1, 2, 3, 4, 5, 7 is 100%. Degree of coverage in terms of schemes 6 is 36,1 %. The Local governments accounting does not enable to divide according to ESSPROS methodology. We accumulate the figure by the aid of Social Affair Ministry data collection. That is the Scheme 5 which data is social care expenditure. As we cannot use the local government accounting, the Ministry of Social Affairs is responsible for the correctness of the data. To compare the functions of the local government total expenditure, i.e. health, disability, oldage, family and children, social exclusion it is 36,1 % of the total expenditure. The Ministry of Social Affairs has collected the expenditure at 100% from the provider of the service. The periodic benefits, lump sums paid by local governments and expenditures are probably not included. The Ministry of Social Affairs does not collect data for these items. As a result of the comparison we can state that the data of the expenditure at the scope of 63,9 % are not collected on local government level, i.e. it accounts only for 3,0 % of the total expenditure of the Estonian social protection of all schemes. In co-operation with the Ministry of Social Degree of coverage in terms of receipts, expenditure and detailed benefits Expenditures for transport cards granted to students and pensioners are not included in Scheme 17. Degree of coverage in terms of receipts is 97,0 %. Degree of coverage in terms of expenditure is 97,0 %. Degree of coverage in terms of detailed benefits by the certain function of expenditure sickness/health care 0,02% disability 1,89% old age 2,40% survivor 0,38% family/children 7,98% unemployment 4,05% housing 0% social exclusion not else-where classified 20,53% Cases of nonapplication of the ESSPROS methodology in the form of a comprehensive list 11

12 Affairs and Ministry of Finance we try to find a possibility to solve this problem. some of expenditure are not divided by function Row 3 Nothing to report. Nothing to report. Nothing to report. Row 4 Full coverage Full coverage No cases c) Table 2 of the template of quality report on core system. Methodologies and assumptions used in the estimates of statistics The table 2 is intended to report on the methodologies used in the estimates of statistics. The following type of information is supposed to be reported in this table: Case 1: One or more schemes for which data sources do not provide detailed information and figures for all the items concerned are estimated. In this case, the column C1 should contain the number of the scheme(s) and the column C2 should contain the information about the method of estimation, as detailed as possible. The columns C3 and C4 should remain empty (see row 1 of the table below). If two schemes are estimated and the methodology differs, they should be reported in different rows of this table; Case 2: One or more benefits under a certain scheme are estimated because the data sources used contain no information on them. The scheme must be reported in column C1 of the table and the item(s), together with the method of estimation, should be reported in the C3 column. Columns C2 and C4 contain no information in this case (see row 2 of the table below). If several benefits are estimates under the same scheme, they should be reported either on the same row, if the methodology is the same, or in different rows, if the methodology differs; Case 3:If the data sources provides incomplete information and some estimates are to be performed by breaking down a total, the report should contain: o If a total is to be break between different items of the same scheme, then the scheme is to be reported on the column C1, and the column C4 contains in this case a list of the items concerned together with the method used. Columns C2 and C3 will contain no information in this case (see row 3 of the table below); o If the total is to be break for the same item between several schemes, all the information is to be reported on the column C4: the item, a list of schemes concerned and the method used. All other three columns will contain no information in this case (see row 4 of the table below). Recommendation: Any information contained by other cells than those supposed to be filled in for this table, is misleading. 12

13 Example 7. Correct way of reporting estimations for all the three cases described above. Row 1 Row 2 Row 3 Row 4 C1 C2 C3 C4 Scheme Estimates for schemes on which no data are available 08 The following indicators were used: Expenditures - Index of nominal and real wages ( lung/nach_branche.html), Work Volume Statistics WV ( quellen/blank/ blank/statistique_du_volume/anhang.html). Estimate the total of annually worked hours Estimate the total of work absences caused of illness in hours Estimate the theoretical maximum amount of payment of salary in the event of illness Subtract the estimated contribution of insurances Subtract the amount of paid sick leave in the event of maternity Receipts: Expenditures = Receipts Breakdown: Total expenditures contribution from insurances in the event of illness part in event of maternity = Total expenditures = Estimates for missing receipts, expenditure and detailed benefits, if: (a) data are completely lacking Scheme 12 Estimation of labour costs ( : Placement services and job search assistance) is derived from annual average gross wage (including premiums/contributions of an employer) and number of labour offices employees in the area of employment mediation and counselling (b) the breakdown for one benefit or group of benefit is missing Scheme , , , functional distribution estimated according to the known distribution of all health insurances. Item , Schemes 1, 2, 4, 8. distribution estimated according to the known distribution of all health insurances 13

14 d) Table 4.1 of the template of quality report on core system. Other information related to quality This is the place where one can report special cases of data sources, information about the way the questionnaire was compiled, short information about new schemes and/or benefits, and all other information that was not possible to be reported in tables 1 to 3. Some examples are given in the table below: Example 8. Different reports in table 4.1 In the following cases the benefits expenditure exists and values zero because of the rounding of the sum (everything less than 0.5 million of euro is rounded to zero): Schemes : disability pension Schemes : old age pension Scheme 15: survivors pension Scheme 25: old age pension Added new scheme 22 for miners. According to the base of legislation, this is the same as scheme 2 rules of granting pensions differ significantly from the others. This is DB scheme, while scheme, while scheme 2 is in transition from DB to NDC. A special mention have to be made for PT, who sent inside this table a complete temporal analysis (2006 and 2007), of the data contained in the questionnaire, both for the receipts and expenditure, including as well an analysis by function and detailed benefits. For those figures that varied between 2006 and 2007, the main reasons for that variation are provided. A little part of this analysis is displayed as an Annex, as Example 9. e) At the end of the quality report, any other information can be provided if it contributes to a better quality reporting. Example 10: MT included in the QR, important information on revisions: ''This year s transmission data included an extensive revision of expenditure on Survivor s Pensions and related cash benefits as these were reclassified as per ESSPROS regulations. Revisions are backdated by a five year period in every transmission till 2003 in this case. Social Protection expenditure was in aggregate increased by an average of 6.7 million each year. The main Scheme which accounted for most of the revision was Scheme 14- Sickness Days Employers Expenditure ( , upwards) due to revisions in the average weekly wage. Other revisions were for Scheme 13 (2006, downwards), Scheme 15 (2006, upwards), Scheme 16 ( , downwards) and Scheme 20 (2006, upwards). Data for which source are administrative records (i.e. most of the data) is not subject to change as yearly figures are approved by parliament.'' 14

15 5. PB specific tables - best practice a) Table 1.2 of the template of quality report on pension beneficiaries - Geographical comparability Table 1.2 is similar to the table 1.3 for the Core System. Therefore, only two examples of reports are presented in the table below: Example 11. C1 C2 C3 Degree of coverage in terms of schemes Scheme 20 Degree of coverage in terms of beneficiaries Beneficiaries of the new employer-based retirement income provision (Mitarbeitervorsorgekasse) are not yet included, as there are no pension beneficiaries yet. Cases of non-application of the ESSPROS methodology in the form of a comprehensive list RAFP (new scheme) Missing Until now, no data for this (very recent) scheme. But the pension ( ) in this scheme is paid in addition to other old-age pensions, so the number of beneficiaries would have been eliminated anyway as double-counting. - Example 12. If no problems should be reported, a sentence should be inserted in the first row of the table as one can see in the four examples below: C1 C2 C3 Degree of coverage in terms of schemes Degree of coverage in terms of beneficiaries Cases of non-application of the ESSPROS methodology in the form of a comprehensive list Data for all schemes are available The data on the total number of pension beneficiaries are available for all functions but the calculation is used for the data by gender. Data complete for all the pension schemes Data complete for all the pension beneficiaries - There are no cases of non-application of the ESSPROS methodology for the data for year Full Full Full Full coverage Full coverage No cases b) Table 2.1 of the template of quality report on pension beneficiaries. Estimates for schemes on which no data are available This table should contain all the cases of estimations on the number of beneficiaries. If a whole scheme is estimated, it is not necessary to provide a list of items. However, if all of the items concerned by that scheme are listed, the information is more complete, so we recommend listing always the benefits concerned. Some examples are listed in table below. 15

16 Example 13. Schemes Items (7) Type of estimate (8) Summary of estimation procedure Scheme , , , , , Based on known distribution. Scheme , , , Based on extrapolation Last known annual change The beneficiaries of an equalisation supplement were split up to the different benefits covered by using detailed information with slightly different figures. Scheme 15 Based on extrapolation Average annual change between the last two surveys. Last survey data are from 2003 and 2007 c) Table 2.2 of the template of quality report on pension beneficiaries. Information on the treatment of double counting 1. For a pension category inside a single scheme The information on the treatment of double counting type 1 should be reported by item and schemes. If for a specific item there is double counting treated inside more than one scheme, and the treatments are different, a new line should be created in the table as one can see below for the first two lines of the table below; If for a specific item the treatment was applied inside several schemes, the schemes can be reported on the same row of the table (see row three in the table below); If the number of concerned schemes is high, a list should be provided (last row in the table below). Example 14. Item Scheme Treatment (yes or not) Type of treatment (9) Scheme 1. Yes. Estimation based on previously known data. Summary of the treatment As multiple recipients are only counted once in the data source, some groups of double counting had to be added at benefit level Scheme 5. No. - There is no information available on double counting between disability pensions and survivors' pensions. It has been assumed to be not significant Schemes 01, 03 yes Using a personal identification number , 5, 6, 8, 10, 11, 12, 13, 14, 15, 16, 19, 20, 21, 22, 24 yes Using a personal identification number Eliminated double counting The elimination of double counting is carried out removing multiple fiscal codes inside each scheme belonging to this pension category 16

17 2. Between schemes (a pension category for all schemes) This table is very important for reporting double counting. This type of double counting raises some problems if, for a specific item, it concerns more than two schemes. If the treatment of double counting was performed by pairs of schemes, the report should describe this. It is important to report, as complete as possible, which are the pairs of schemes that have common beneficiaries and how this double counting was eliminated. If for a specific item, several pairs of schemes are concerned they should be separately reported. If there are two or more supplementary schemes and one single basic scheme, the report concerns all possible pairs of the type basic scheme supplementary scheme. If the basic scheme is Scheme A and the supplementary ones are the schemes B and C and if the treatment is the same, then the report on the column C2 is 'Scheme A versus Schemes B and C'. For some countries the basic scheme has double counting beneficiaries with many supplementary schemes. In this case, the report should contain the basic scheme and the list of all other schemes ( Scheme A versus schemes B, C, D,. ) If there are two or more basic schemes (i.e. A and B) and two or more supplementary schemes (i.e. C and D), and the treatment is the same for all of possible pairs then the report on the column C2 can be 'Schemes A and B versus Schemes C and D'. In this case, the reader will know that there is a double counting between the pairs: A versus C, A versus D, B versus C and B versus D. Recommendations: No report on horizontal double counting treatment for some item should be made without specifying the schemes concerned; No report on double counting should contain one scheme. There is no horizontal double counting treatment for only one scheme. Always it is about at least a pair of schemes; In case for a given item there is double counting between scheme A and scheme B, but the treatment was not possible, please fill in the row by indicating 'No' in the column C3 and 'Treatment no possible' in the column C4; Maintain all the rows of this table in the template in both cases of no double counting for an item and in case of double counting but no treatment possible. Example 15. In the table below, for the same item, there was a treatment between pairs of schemes '10 and 12' and '10 and 13'. For each pair, the methodology was reported. Keep in mind that if the method is the same, both pairs can be reported in the same row of this table but only as pairs. In the last row, we can see a more complex report where four different pairs are reported, namely pairs 'Scheme 01 versus Scheme 03', and Scheme 02 versus each one of the three schemes 01, 03 and

18 Row 1 Row 2 Row 3 C1 C2 C3 C4 C5 Item Schemes Treatment (yes or not) Type of treatment (10) Summary of treatment and 12 yes Using micro level data. Beneficiaries of State old-age pension (Scheme 12) received a State social insurance pension (Scheme 10) in most cases. To obtain the number of old-age pension beneficiaries without double counting, the number of beneficiaries who also received a State social insurance pension (Scheme 10) was subtracted from the total number of beneficiaries of a State old-age pension and 13 yes Using micro level data. Beneficiaries of an Officials and Military Personnel State old-age pension (Scheme 13) received State social insurance pension (scheme 10) in most cases (in 2007, 95 per cent). To obtain the number of old-age pension beneficiaries without double counting, the number of beneficiaries who also received a State social insurance pension was subtracted from the total number of beneficiaries of an Officials and Military Personnel State old-age pension Between schemes 01 and 03 Scheme 02 with Schemes 01, 03, 06; Yes Using a personal identification number Eliminated double counting 6. Reporting revisions Table 3, revision of statistics is similar for PB and CS and the reporting way is the same. Starting with the previous data collection (data collection 2009), some data were revised both for the QD and the module on PB. For the current data collection (collection 2010), there will be again revisions both for quantitative data and for the figures on the number of beneficiaries. The revisions should be reported categorized in five classes: 1. Changes in the data sources used; 2. Changes in the methods used for estimating data; 3. Revisions of data due to conceptual adjustments; 4. Revisions of data due to availability of final statistics; 5. Other revisions of data (e.g. due to quality review actions). 18

19 During the data collection 2009 some problems on reporting revisions were encountered, as listed below: Revisions reported as belonging to two different categories; Revisions belonging to one category, reported under another category; Revisions incomplete reported or not reported. All revisions should be reported by scheme and, if it is the case, by items, and EUROSTAT proposes this year a simpler table for reporting revisions. A revision will be reported by scheme, item(s), type of revision and, if necessary, a comment on why and how the revision was performed. The type of the revision will be reported by filling with an X the corresponding cell in the table; For the cases where revisions concern both 2006 and 2007 data, if the number of revisions is too big, the table can be duplicated, one for 2006 and the other one for If revisions were made for quantitative data for several years, a supplementary column can be inserted and the year of revision can be reported easier. In Annex II to this document, an example is given. The first table contains five revisions reported this year (data collection 2009) and in the second table the same revisions are listed using the new format. The third table shows how to proceed if revisions for different years must be reported in the same table. 19

20 7. Final general recommendations: 1. Whenever possible, the national quality report should be transmitted together with the questionnaires; 2. Do not let tables without filling in at least one row. If nothing should be reported, state a short sentence in each cell of the first row of the table; 3. Provide the full list of data sources both in Table 1.1 part I and in Table 1.1 part II; 4. The column C3 of the table 1.1 part II, is not intended for reporting if a scheme covers different occupational categories or other social categories. It should be reported only the geographical coverage; 5. If a scheme is missed this should be reported only in the table 1.3 (Core System); 6. If a scheme is missed but the figures are estimated, this should be reported only in the table 2 (Core System); 7. If an item is missed and no estimation is provided, this should be reported in table 1.3 but also in table 1.1 as a problem for the data source that provides the rest of the figures for the scheme to which the item belongs (Core System); 8. If an item is missed and an estimation is provided, this should be reported in table 2 but also in table 1.1 as a problem for the data source that provides the rest of the figures for the scheme to which the item belongs (Core System); 9. Never report the same scheme or the same item in both table 1.3 and table 2. They are disjoint as for the information they contain (Core System); 10. Pay attention to the footnotes at the bottom of the tables in the templates, they help with indication on what can be filled in those tables; 11. In the last column of the Table 1.2, there have to be listed at least two data sources and the names of these data sources have to be the same as the names in table 1.1. part I (Core System); 12. For each row of the table 2, only one of the columns C2, C3 and C4 has to be filled-in (Core System). 20

21 Annex I Example 9 Exanple of temporal analisis of the receipts and expenditure for the last two years of the questionnaire. Temporal analysis of receipts: Esspros core system continued to record a surplus. In 2007 the total receipts of social protection in Portugal increased +3.1% when compared to 2006, slightly higher than the increase of total expenses, which was +2.6%. A more detailed receipts analysis shows that the biggest variations are on the level of social contributions (accounts and ), and on other receipts (account and ): / The Employers' social contributions, accounts e , recorded a decrease of 11.0% and 35.1% in 2007, explained by negatives movements in schemes 20 and 22 and in scheme 43, respectively. In account , the scheme 16 recorded a decrease in receipts and in expenditures due to the reorganization of services that constitute this scheme; in scheme 22, has been a legislative restructuring in the receipts allocated to the scheme and will have own receipts of social contributions of employees (active and retired), causing a reduction in social contribution by the employers. In account , the decrease of receipts in scheme 43 followed the decline of social protection expenditure, was dully confirmed by the information providers / In the social / In Other Temporal analysis of expenditure: In 2007 the total expenditure of social protection in Portugal increased +2.6% compared to 2006, while the expenses on social benefits increased 3.3% (representing in % of total expenditure), and other expenses decreased 4.0% (representing 8% of the total value of expenditure in 2007). The decrease in Other expenditure is explained by negative evolution of other expenditure other (account ), with -12.1%, from 1530 m.e. in 2006 to 1344 m.e. in This behaviour is explained mainly by the negative movements of the scheme 42 (-97.5%) from 6 m.e. in 2006 to 0.2 m.e. in 2007; and in scheme 61 (-26.3%), from 948 m.e. in 2006 to 698 m.e. in 2007, dully confirmed by the information providers. The expenditure on social protection benefits showed an increase in most functions, except for Unemployment function (-5.0%) and Housing function (-18.7%). Analyzing the main changes by function, in 2007, we report the following: - Sickness ( ) - This function weighted 28% of the total expenditure on social protection in Portugal, being the second heaviest function. In 2007 there was a slight increase of 0.4% compared to Reimbursement of pharmaceutical products, non means tested This benefit has a significant negative variation (-86.9%) from 21 m.e. in 2006 to 3 m.e. in This situation is explained mainly by a) decrease in scheme 16 (-95.9%), due to the restructuring of the health system in an approximation of the rules followed in the scheme 22 ADSE; and by 2) the decrease in scheme 43 (- 80.4%) due to end of conditions in protocols between some companies and National Health Service Other benefits in Kind, non means tested - This benefit shows a increase (+46.0%), with 447 m.e. in 2006 and 652 m.e. in This situation is explained mainly by scheme 2 (from 286 m.e. in 2006 to 484 m.e. in 2007), duly confirmed by the final data provided by the information providers of this scheme, and returned to the EUROSTAT with the 2nd version of ESSPROS Core System 2007 in 27th July Disability ( ) - This function - Old age ( ) - This is the function - Survivors ( ) - This function represent 7% of - Family ( ) - This function represent 5% - Unemployment ( ) This function has - Housing ( ) This function - Social Exclusion ( ) This function 21

22 Annex II. Example of revisions reporting. Five revisions reported in the old format Scheme/benefit Changes in the data sources used (13) Changes in the methods used for estimating data (14) Revisions of data due to conceptual adjustments (15) Revisions of data due to availability of final statistics (16) Other revisions of data (e.g. due to quality review actions) 1/ Computation was corrected. 2/ New data source / , , Provisional estimation was replaced by biannual data missing in the last data collection data, Schema , , , , All scheme level / and The same revisions reported in the new format to put in line information supplied by data sources with the ESSPROS methodology Type of revision - For survivors pensions coming from old-age pensions, we used more detailed information in EIR. But it does not change the number of pensions in for this type of survivors pension. Scheme Items Changes in the data sources used (13) Changes in the methods used for estimating data (14) Revisions of data due to conceptual adjustments (15) Revisions of data due to availability of final statistics (16) Other revisions of data (e.g. due to quality review actions) Details on the revision Scheme X Computation was corrected. 22

23 Scheme X New data source. Scheme 02 2/ , , Scheme , , , , All scheme level and X Provisional estimation was replaced by biannual data missing in the last data collection. X to put in line information supplied by data sources with the ESSPROS methodology X For survivors pensions coming from old-age pensions, we used more detailed information in EIR. But it does not change the number of pensions in for this type of survivors pension. If different years for the revisions are needed, a supplementary column can be inserted. Type of revision Year Scheme Items Changes in the data sources used (13) Changes in the methods used for estimating data (14) Revisions of data due to conceptual adjustments (15) Revisions of data due to availability of final statistics (16) Other revisions of data (e.g. due to quality review actions) Details on the revision 2006 Scheme X Computation was corrected Scheme X New data source Scheme 02 2/ , , Scheme , , , , All scheme level and X Provisional estimation was replaced by biannual data missing in the last data collection. X to put in line information supplied by data sources with the ESSPROS methodology X For survivors pensions coming from old-age pensions, we used more detailed information in EIR. But it does not change the number of pensions in for this type of survivors pension. 23