ESSPROS Task Force on Methodology November Comparison of expenditure for social benefits between ESSPROS and national accounts

Transcription

1 EUROPEAN COMMISSION EUROSTAT Directorate F: Social statistics Unit F-5: Education, health and social protection Luxembourg, 03/11/2017 DOC SP-TF Task Force on Methodology November 2017 Comparison of expenditure for social benefits between and national accounts Meeting of the Task Force on Methodology Luxembourg, November 2017 BECH Building (Eurostat) B2/464

2 Comparison of expenditure for social benefits between and national accounts 1. Introduction This document describes the comparison between data for expenditure on social benefits () and the corresponding items from national accounts transactions. It takes into account presentations and discussions at the Social Protection Working Group meetings of 2016 and In particular, a detailed description of the flows to be compared is provided in order to facilitate a common approach between countries. The comparison is based on the guidelines contained in DOC SP and its Annex resulting from the analysis of links and s between and national accounts. DOC SP and its Annex were discussed at the Social Protection Working Group meetings of April Guidelines concern in particular the identification of transactions and the delimitation of the scope of social protection. The comparison is carried out separately for benefits in cash and benefits in kind. For both types, total benefits are compared as well as benefits broken down according to the functions of social protection. rerouted social contributions and administration costs are also considered in the comparison, which may apply differently to the various countries according to national treatments of these flows. A set of 4 standard comparison tables is defined, 2 for benefits in cash and 2 for benefits in kind. A checklist for possible systematic sources of s is also provided. Countries should indicate other recognized sources of s according to their national practices providing a description and quantification of the impact. Conclusions may include an analysis of not yet recognized sources of s and of specific s particular to the national system. Regarding timeliness, data are published 20 months after the reference period. Sector accounts are published 9 months after the reference period and Government Finance Statistics (GFS) data by COFOG 1 14 months after the reference period. Data by Classification of Products by Activity (CPA) in the Use table are published 3 years after the reference period. The comparison carried out in the context of this Task Force on Methodology refers to data for 2012 for which final data generally exist from all domains. If a more recent year were selected, other issues linked to release delays and revision policies would need consideration. Data used from national accounts For national accounts (ESA 2010 framework), three sets of data are considered: 1) Sector accounts. 2) Government finance statistics (GFS), in particular for the breakdown of social 1 Classification of functions of Government 2

3 benefits according to COFOG. 3) Final consumption expenditure of non-profit institutions serving households (NPISH), for the breakdown according to CPA available from the Use table. Sector accounts cover all sectors of the economy and provide the reference framework. GFS provide more detail by function of expenditure (COFOG) and by sub-sector of general government (in particular, social security funds). A cross-classification of the two dimensions (function, sub-sector of general government) is available for most countries. Final consumption expenditure of non-profit institutions is relevant for estimating a breakdown by function of social transfers in kind by NPISH. Consistency of data between sector accounts, GFS and the Use table should be checked, in particular - for data no more subject to revision - as regards possible systematic s in certain countries 2. In this respect, tables A3 and B3 also present an overview of the situation by country as regards consistency between the different national accounts datasets for Social benefits in cash Table 1 considers the data comparison between social benefits in cash and the national accounts item social benefits other than social transfers in kind (D62). Table 1 Comparison between benefits in cash and national accounts social benefits other than social transfers in kind National currency NA - 1 TABLE 1 Belgium with RRSC National accounts without RRSC S11 Non-financial corporations Social benefits other than social transfers in kind S12 Financial corporations S13 General government S14 Households S15 Non-profit institutions D9 Capital transfers Any sector All sectors Total benefits other than STK minus benefits identified as out of scope (*) Benefits to be compared with cash benefits Total Social benefits in cash Rerouted social contributions (**) 96 (*) for government, sum of D62 in COFOG functions outside scope and housing (COFOG 10.6) (**) if applicable, depending on national accounts treatment D62 % of NA 2.1 Datasets used The following published datasets have been used. Extractions are available in excel files. Annex I contains tables for cash benefits. All countries are presented in tables A1-A4, but 2 Vintage s for years still subject to revision may appear at specific intervals during a year, in particular between data for general government from sector accounts and GFS (to be further clarified with Eurostat national accounts units). 3

4 only for countries that are member of the task force Eurostat has prepared the comparison tables presented in the document for the case of Belgium 3. Cash benefits (table A1) all schemes, broken down by social protection function Transfers to other schemes rerouted to other schemes (table A2) all schemes, broken down by social protection function NATIONAL ACCOUNTS Sector accounts Non-financial transactions [nasa_10_nf_tr] 4 : NA_ITEM D62 Social benefits other than social transfers in kind, million units of national currency, paid, Sectors S1, S11,S12,S13,S14,S15 (table A3). Government finance statistics General government expenditure by function (COFOG) [gov_10a_exp]: NA_ITEM D62 Social benefits other than social transfers in kind, million units of national currency, general government (table A4). 2.2 Tables for data comparison As presented in the annex to DOC SP , the closest national accounts item to compare with Social benefits in cash is Social benefits other than social transfers in kind (D62). However, there is not a perfect conceptual match for the following reasons: i) D62 may include some benefits in kind; ii) the national accounts treatment of rerouted social contributions should be taken into account; iii) by convention, records all housing benefits as benefits in kind, so that housing benefits in cash are 0, while the national accounts do not have this convention 5 ; iv) some benefits in cash may be recorded as capital transfers in national accounts. Regarding conceptual s i) and iii), these would result in D62 recording benefits not considered in benefits in cash. Difference iv) on the contrary would result in benefits in cash recording benefits not considered in D62. The first row of table 1 is filled with data for D62 from sector accounts (table A3) broken down by sector. Table A3 also contains information by country on the consistency between sector accounts and GFS for D62. The breakdown from sector accounts gives an overview of 3 Two excel files will be distributed. Annex I contains tables for cash benefits. All countries are presented in tables A1-A4, but only for countries that are members of the task force Eurostat has prepared the comparison tables presented in the document for the case of Belgium. Countries are invited to send their comments. 4 The text in square brackets refers to the identification code of tables available on Eurostat database on line. 5 Housing benefits included in D62 are identifiable for the government sector from COFOG data, so that this source of can be neutralized as concerns government benefits. 4

5 the importance of the various sectors as providers of this type of benefits. In all countries the government sector was by far the most important provider, with the exception of Switzerland. The weight of households and NPISHs was below 1% for all countries, except for NPISHs in Sweden and in Cyprus. The importance of the sectors financial and non-financial corporations varied to a larger extent between countries. National specificities should be taken into account in order to orientate the data comparison. Benefits recorded in sector accounts cover all functions and not only those included in the scope of. The second row of the table therefore subtracts benefits that can be identified in national accounts as being out of the scope of. For the sector General government (S13), the amount of out of scope benefits is derived from table A4 (see below), which contains data by function according to COFOG 6. For the sectors Households 7 (S14) and NPISH (S15), benefits D62 paid are observed to be a small proportion of total D62 benefits and, for NPISH a small proportion of total benefits paid, where in the case of NPISH total benefits include social transfers in kind (D63). Initial assumptions used in the example (to be confirmed) are that all benefits D62 paid by these two sectors are benefits paid to their employees 8 and that they fall in the scope of functions. The lower part of table 1 is filled with data for benefits in cash (table A1) and rerouted social contributions (table A2). For data, a breakdown by sector is not available. 9 Total benefits in cash are compared with the final number for D62 from the upper part (75116 in the example). In the example, the is calculated including or excluding rerouted social contributions. The appropriate comparison has to be decided on the basis of the treatment of rerouted social contributions in national accounts. Subject to the validity of the adjustments for scope, rerouted social contribution and housing benefits, the remaining D62 minus benefits in cash calculated in table 1 is expected to be positive because no correction for scope has been applied to sectors S11 and S12, as well as because of the possibility of benefits in kind being included in D62. Table 2 considers the comparison of government D62 by COFOG function (source: GFS, see table A4) with benefits in cash broken down by function. Table 2 Comparison between benefits in cash and government national accounts social benefits other than social transfers in kind by function of social protection 6 If there is a between D62 from S13 in sector accounts and D62 S13 in GFS/COFOG data, the origin of the should be analysed in order to decide on the allocation in/out of scope. 7 Since benefits are analysed here as expenditure of sectors, the part related to households refers to households as employers (individual enterprises and unincorporated enterprises). 8 A comparison with data for employers' social contributions received by the sector Households and NPISHs seems to corroborate this hypothesis. 9 The availability of data broken down by sector could be investigated first with Eurostat national accountants (it may be that additional breakdowns are provided to Eurostat ion a voluntary basis) and then at national level with national accountants. 5

6 1 Table 2 Belgium Data by function (: all schemes. NA: only government) National accounts 2012 Functions Totals Social benefits other than social transfers in kind (*) Sickness, health and disability Old age Survivors Family and children Unemployment Social exclusion n.e.c. Sum of NA functions in scope/ Total Social security funds Other government Total government Social benefits in cash, all schemes Rerouted social contributions (**) Difference minus National accounts should be > Total (*) add relevant capital transfers by function if the information is available (**) if applicable, depending on national accounts treatment The upper part of table 2 is filled with national accounts data for the groups of COFOG functions for which a comparison with benefits by function is possible (see Annex to DOC SP and table A4). The housing function is considered out of scope for the comparison of benefits in cash. Data are available for most countries by government subsector, in particular for social security funds, as presented in the example. The lower part of table 2 is filled with data for benefits in cash and rerouted social contributions broken down by function (from tables A1 and A2). In the case of table 2, the benefits in cash minus government D62 (last row of the table) is expected to be greater than zero for each function. 3. Social benefits in kind Table 3 considers the data comparison between social benefits in kind and the national accounts item social transfers in kind (D63). Table 3 Comparison between social benefits in kind and national accounts social transfers in kind 1 national currency 2012 TABLE 3 Belgium National accounts D9 D63 Capital Social transfers in kind transfers S13 General government S15 NPISHs any sector sectors Total minus benefits identified as out of scope (*) Benefits to be compared with benefits in kind Total Social benefits in kind Administration costs (**) 2238 (*) for government, sum of D63 in COFOG functions outside scope (D62 in housing (COFOG 10.6) is subtracted to out of scope) - see table B4 For NPISH, calculated as P31 (final consumption expenditure) of NPISH from use table in some products - see table B5 (**) non-market administration costs only - in the example an estimate has been put indicatively All NA - % of NA

7 3.1 Datasets used The following published datasets have been used. Extractions are available in excel file. 10 Benefits in kind (table B1) all schemes, broken down by social protection function Administration costs (table B2) all schemes, total NATIONAL ACCOUNTS Sector accounts Non-financial transactions [nasa_10_nf_tr]: NA_ITEM D63 Social transfers in kind, million units of national currency, paid, Sectors S1, S11,S12,S13,S14,S15 (table B3). Government finance statistics Use table General government expenditure by function (COFOG) [gov_10a_exp]: NA_ITEM D63 Social transfers in kind, million units of national currency, general government (table B4) 11. Use table at purchasers' prices [naio_10_cp16]: INDUSE Final consumption expenditure by non-profit organisations serving households (NPISH) (table B5). 3.2 Tables for data comparison As presented in the annex to DOC SP , the closest national accounts item to compare with Social benefits in kind is Social transfers in kind (D63). However, there is not a perfect conceptual match for the following reasons: i) Social transfers in kind are recorded only for government and NPISH; Social benefits in kind may include benefits in kind provided by all sectors as employers; ii) Costs of administering non-market benefits are recorded in administration costs, but in national accounts these costs are recorded in D63 12 ; 10 Two excel files will be distributed. Annex II contains tables for benefits in kind. All countries are presented in tables B1-B5, but only for countries that are members of the task force Eurostat has prepared the comparison tables presented in the document for the case of Belgium. Countries are invited to send their comments. 11 Although the national accounts item D63 is not published in the COFOG database, D63 can be derived from Individual final consumption expenditure of government (P31 of government = D63 of government) or also, if P31 is not available, from data for Final consumption expenditure of government (P3) allocated to COFOG individual functions. 12 Exceptions may occur for costs explicitly charged by government schemes (social security funds) in some countries. ESA Table 29 experts group decided in April 2017 group meeting the following: "a non-zero figure would only be entered if the scheme required a fixed proportion of gross contributions or gross property income 7

8 iii) by convention, records all housing benefits as benefits in kind, while the national accounts do not have this convention. Regarding conceptual s i) and iii), these would result in benefits in kind recording benefits not considered in D Difference ii) on the contrary would result in D63 recording costs that are treated in as administration costs rather than as benefits in kind. The first row of table 3 is filled with data for D63 from sector accounts (table B3) broken down by sector. Table B3 also contains information on the consistency for the various countries between sector accounts, GFS and the Use table for D63. The breakdown from sector accounts gives an overview of the importance of the various sectors as providers of this type of benefits. For transfers in kind the government sector was by far the most important provider. The weight of NPISH ranged from a minimum of 2.4% for Slovakia to a maximum of 20.5% for Switzerland. Benefits recorded in sector accounts cover all functions and not only those included in the scope of. The second row of the table therefore subtracts benefits that can be identified in national accounts as being out of the scope of. For the sector General government (S13), the same approach as for benefits in cash can be used. The amount of out of scope benefits is derived from table B4 (see below), which contains data by function according to COFOG 14. For the sector NPISH (S15), the breakdown by 64 CPA categories of products of the NPISH Final consumption expenditure from the Use table can be a source generally available for estimating the part of benefits in kind provided by NPISH that falls outside the scope of. Final consumption expenditure of NPISH consists predominantly in expenditure on their own non-market output, calculated as sum of production costs. Therefore the breakdown of final consumption by product can give indications on the activity of NPISH. The last column of table B5 has been used as % of out of scope social transfers in kind of NPISH to compile table 3 in the example given. The lower part of table 3 is filled with data for benefits in kind (table B1) and administration costs (table B2) 15. For data, a breakdown by sector is not available. Total benefits in kind plus administration costs (36854 in the example) are compared with the final number for D63 from the upper part (38055 in the example). Subject to the validity of the adjustments for scope, administration costs and housing benefits, the remaining calculated in table 3 is expected to be zero or lower if some benefits in kind are included in D62 rather than in D63. to be set aside for covering the costs of operating the pension scheme. If the scheme costs are financed through other sources (most commonly general taxation), a zero should be entered in this cell of the table." 13 In order to eliminate the second reason for discrepancy indicated under i) the possibility to isolate in benefits in kind by employers using the classification of schemes by decision making (code 121: "schemes for employees") could be explored. 14 If there is a between D63 from S13 in sector accounts and D63 S13 in GFS/COFOG data, the origin of the should be analysed in order to decide on the allocation in/out of scope. 15 Only non-market administration costs should be considered. This has been estimated in the example as the between administration costs and item D61SC Social insurance scheme service charges from sector accounts. 8

9 Table 4 considers the comparison of government D63 by COFOG function (source: GFS, see table B4) with benefits in kind broken down by function. Table 4 Comparison between benefits in kind and government national accounts social transfers in kind by function of social protection 1 Table 4 Belgium National accounts Data by function (NA: only government, : all schemes ) Functions 2012 Totals Social transfers in kind (*) Sickness, health and disability Old age Survivors Family and children Unemployment Housing (including D62) Social exclusion n.e.c. Sum of NA functions in scope/ Total Social security funds Other government Total government Social benefits in kind, all schemes Administration costs (**) 2238 Difference minus National accounts should be > (*) add relevant capital transfers by function if the information is available (**) government administration costs only (estimated). Administration costs are included only for total and not for data by function Total The upper part of table 4 is filled with national accounts data for the groups of COFOG functions for which a comparison with benefits by function is possible (see Annex to DOC SP and table B4). Benefits D62 in the housing function are included under the column housing. Data are available for most countries by government subsector, in particular for social security funds, as presented in the example. The lower part of table 2 is filled with data for benefits in kind broken down by function (from tables B1). However, administration costs are not available broken down by function and only the estimated total amount is included in table 4 (number in orange font). The total included in the last cell to the right takes administration costs into account, but the s for the individual functions do not. 4. Total social benefits Total social benefits are clearly defined as the sum of the two respective categories in (in kind benefits and cash benefits) and in national accounts (social transfers in kind and other social benefits). The sum of the numbers from table 1 and 3 and tables 2 and 4 can be finally compared using the same structure. Tables for total benefits 9

10 National currency NA - 1 TABLE 1 Belgium with RRSC National accounts without RRSC S11 Non-financial corporations Social benefits other than social transfers in kind S12 Financial corporations S13 General government S14 Households S15 Non-profit institutions D9 Capital transfers Any sector All sectors Total benefits other than STK minus benefits identified as out of scope (*) Benefits to be compared with cash benefits Total Social benefits in cash Rerouted social contributions (**) 96 (*) for government, sum of D62 in COFOG functions outside scope and housing (COFOG 10.6) (**) if applicable, depending on national accounts treatment D62 % of NA 1 Table 4 Belgium Data by function (NA: only government, : all schemes ) 2012 Functions Totals National accounts Social transfers in kind (*) Sickness, health and disability Old age Survivors Family and children Unemployment Housing (including D62) Social exclusion n.e.c. Sum of NA functions in scope/ Total Social security funds Other government Total government Social benefits in kind, all schemes Rerouted + Administration costs (**) Difference minus National accounts should be > (*) add relevant capital transfers by function if the information is available (**) government administration costs only (estimated). Administration costs are included only for total and not for data by function Total 5. Topics for discussion at the Task Force a) Do the members of the Task Force have comments on the analysis and comparisons made in this document? Do you have any suggestions in order to improve the proposed comparisons (particularly for schemes run by NPISH)? Please note that: - the reference methodology for the comparison is included in document DOC SP-TF Annex (an updated version, with track changes, of the 2017 Working group document DOC SP Annex 1); - the data used in the document are those publicly available on Eurostat database on line. b) would you agree on the possibility to refine data comparisons/ analysis by using all the data available at Eurostat (both in and in National Accounts) but not disseminated in Eurobase? c) would you agree on the possibility to refine data comparisons/ analysis by using data available at national level? 10