Item 3.2 Improvement of expenditure data on education

Transcription

1 EUROPEAN COMMISSION EUROSTAT Directorate F: Social statistics Unit F-5: Education, health and social protection Doc 2017-ETS-02 Item 3.2 Improvement of expenditure data on education Meeting of the Education and Training Statistics Working Group Luxembourg, 1-2 June 2017 BECH Building Room QUETELET 1

2 Executive Summary In reply to the 2012 Communication of the Commission on "Rethinking Education: investing in skills for better socio-economic outcomes" and the recommendations of the Indicator Expert Group (IEG) related to indicators for monitoring education investment, Eurostat launched a project "Improvement of expenditure data on education". This project is fully in line with the needs of policymakers: in November 2016, the Council (ECOFIN) "recognises the importance of timelier, better and integrated data ( ) on major components of social expenditure" (to be understood as expenditure on education, health and social protection). A contract is supporting from January 2016 to March 2018 this long term project. The last ETS Working Group meeting in 2016 was the opportunity for Eurostat to present the objective of the contract, its timetable as well as the forthcoming pilot data collection on UOE finance data which is part of it. This document recalls the objectives, tasks and timeline of the contract; the main outcomes of the assessment of the UOE finance data, the results of the UOE finance pilot data collection exercise and the description of the reconciliation exercise that will take place in six volunteer countries. It also presents the main conclusions at this stage and the further steps of the project. The ETS Working Group is invited to take note of: The main preliminary outcomes of the assessment of the UOE finance data; The results of the UOE finance pilot data collection. Participants of the ETS Working Group are invited to express their views and discuss: The approach of the reconciliation exercise and their willingness to participate to this exercise by hosting a visit; The Eurostat proposal for further activities (in the medium and long term) to improve expenditure data on education in three key areas: o Improving the comparability of data and the UOE methodology on education expenditure (incl. the adaptation of the national quality report and a revision of the UOE manual); o Developing the consistency between the UOE methodology and the National Accounts methodology; o Developing a strategy to improve the timeliness of UOE finance data. The Eurostat proposal to organise a workshop in February 2018 to present the final results of the contract; to discuss the final objectives to be met in order to significantly improve expenditure data on education; to prioritize these objectives and to set up the timeframe for dedicated actions. 2

3 1 Context In reply to the 2012 Communication of the Commission on "Rethinking Education: investing in skills for better socio-economic outcomes" that proposed to "monitor actions taken by Member States to prioritise growthenhancing expenditure and developing an evidence-based framework to analyse the efficiency of public spending for quality education and training" and proposed to "initiate for the first time a debate at EU level with relevant stakeholders on the benefits of investment in different education and training sectors (vocational training, adult learning, higher education) and ways to increase the efficiency of spending", a new Indicator Expert Group (IEG) related to indicators for monitoring education investment was created in The Indicator Expert Group (IEG) was asked to provide recommendations from a user perspective for work on education investment statistics. The Indicator Expert Group (IEG) on education investment mainly recommended: To publish more data, to have them more timely (i.e. reducing the present nearly 3-year time lag) and with sufficient/logical breakdowns. In particular, regarding more timely data, it was recommended that a core set of variables should be identified to be provided with a shorter time lag. "Regarding Public expenditure i) to align the UOE methodology with the European Standard Accounts when relevant and useful (the use of accrual basis was underlined); ii) to better document education finance stocks and flows in UOE to avoid double counting issues; iii) to clearly define what should be included as public and private education expenditure; iv) to do some methodological work in order to distinguish better between capital and current expenditures." The Indicator Expert Group (IEG) on education investment also noted that "the quality and the implementation of the UOE finance methodology should be checked. In this respect, Member States have to document deviations from the UOE methodology (i.e., in addition to the data files, each Member State has to provide a quality report documenting the degree of compliance to the UOE methodology). This requirement is in the 'UOE' regulation (912, Sept. 2013). The quality reports would be a source of information for checking quality of UOE expenditure data with a view to revising the methodology. 1 As a follow up to the recommendations of the Indicator Expert Group (IEG) on education investment, Eurostat launched a new project "Improvement of expenditure data on education". In 2015, Eurostat informed members of the ETS Working Group about the start of a methodological work with the general objective of improving the quality of expenditure data. A contract is supporting from January 2016 to March 2018 this long term project (see below section 2). The last ETS Working Group meeting in 2016 was also the opportunity for Eurostat to present the objective, its timetable as well as the forthcoming pilot data collection on UOE finance data which is part of it. The "Improvement of expenditure data on education" is fully in line with the needs of policymakers: in November 2016, the Council (ECOFIN) "recognises the importance of timelier, better and integrated data ( ) on major components of social expenditure" (to be understood as expenditure on education, health and social protection). This document presents the objectives, tasks and timeline of the contract (section 2); the main outcomes of the assessment of the UOE finance data (section 3), the results of the UOE finance pilot data collection exercise (section 4) and the description of the reconciliation exercise that will take place in six volunteer countries (section 5). Section 6 presents the main conclusions and the further steps of the project. 1 Indicators Expert Group on education investment Final report June 2015 chapter 3 Data sources 3

4 2 "Improvement of expenditure data on education": objective, ongoing tasks and timeline 2.1 Objective and tasks The general objective of the ongoing project is to make available more timely expenditure data on education, with a better quality though methodological changes to the UOE data collection. As part of this project, Eurostat benefits from the support of a contractor. Under the on-going contract, five tasks were defined. Task 1 Assessing the quality of the existing 'UOE' expenditure data: the assessment of the quality of UOE expenditure data is based on the national quality reports and on the UOE questionnaires of three reference years. It follows the definition of quality as defined by article 12 of the Regulation on European Statistics and as detailed in the ESS Standard for Quality Reports Structure (ESQRS). Task 2 Organising a pilot data collection in view of producing fresher/more up to date and timely core indicators: the objective of this task is to select particularly relevant and robust core indicators and to collect the necessary data relating to the financial year 2015, by the end of 2016, from all volunteer Member States and EEA countries. The final purpose is to investigate the possibility to produce fresher, i.e. more up to date and timely core indicators. The selection of core indicators was based on the assessment (Task 1), on the main stakeholders needs and on the discussions during the ETS Working Group. Task 3 Inventory and documentation of the main methodological differences between the 'UOE' expenditure data and other finance data on education: The purpose of this task is to make the inventory of the main conceptual differences between the 'UOE' methodology for the collection of data on public and private expenditure, and the National Accounts methodology in recording government and household expenditure (ESA2010). Task 4 Quantifying the main methodological differences between the 'UOE' expenditure data and other finance data on education: the objective of this task is to quantify, as far as possible, the methodological differences between the 'UOE' expenditure data and the National Accounts (NA) data on education in view of a reconciliation of the two different sources. The framework of a reconciliation table between UOE and NA data will be tested during study visits in six volunteer Member States. Task 5 Comprehensive assessment of expenditure data on education, including recommendations to improve or adapt the existing underlying methodology: The objective of this task is to provide a complete report on the assessment of expenditure data on education, including recommendations to improve or adapt the existing underlying methodology on the basis of the outcomes of each task of the project. 2.2 Timeline The contract started in January 2016 and will run until March 2018 with the timeline presented below. Task Timeline Status Task 1 January 2016 November 2017 Task 2 April 2016 May 2017 Task 3 February 2016 November 2017 Task 4 May 2017 November 2017 In progress: a first report has been produced and is continuously adapted according to results of task 3 and 4, a synthesis of the results should still be produced. Completed In progress: the first part of the conceptual analysis has been performed. Analysis of transactions is in progress In progress: Preparation of the reconciliation framework. Six country visits (August to October 2017) Task 5 September 2017 March 2018 Planned to start in September

5 3 Main outcomes of the assessment of the UOE finance data This section presents some of the outcomes of the assessment of the UOE expenditure data. The quality assessment of the UOE expenditure data is based on the analysis of the UOE-2013, UOE-2014 and UOE-2015 questionnaires (covering reference years 2011, 2012 and 2013 respectively) and corresponding national quality reports. At this stage, several issues might already be highlighted. 3.1 Statistical processing Source of data One characteristic of the UOE finance data is that it uses mutiple sources which are not under the direct responsibility of the UOE national data provider. Only few countries declare using sources for which a single organisation is responsible while one third of the countries declares that the number of organisations responsible for providing part of the data equals or is above five. Most of the national data providers use survey data (sample survey or exhaustive surveys) (24 data providers out of 30 data providers) and administrative data (half of the data providers) to report the UOE finance data. Only one third of the national data providers uses registers. When they are used, specific surveys provide data about R&D (8 countries), about households expenditure (12 countries) and about private institutions (6 countries). The diversity of sources is caused by multiple factors: sources might change according to the ISCED level, the type of expenditure, the source of expenditure as well as the type of educational institutions. Such dependency on other data providers impacts the overall duration of the national workflow as well as the compliance with the deadline set by the EU Regulation. This also had an impact on the participation to the pilot data collection exercise Estimation The UOE manual encourages national data providers to estimate data that are not available. A quite large spectrum of variables are estimated in the finance UOE questionnaires. Over the period under scrutiny, estimation procedures were usually used to breakdown data by ISCED levels and/or to provide data for some components of households expenditure (six countries); of other private entities expenditure (one country); of transfers to households and students (three countries); of personnel expenditure (four countries); of regional and local government expenditure (three countries); about private institutions (five countries). Estimation procedures were also implemented to evaluate administrative costs (two countries), the rent of school buildings, the costs of schools books by ISCED levels, capital expenditure, R&D expenditure, ancillary services, expenditure on non-instructional institutions and expenditure on pre-school day care. 3.2 Availability of the data The complexity and the level of details of the UOE finance questionnaires (FIN1-source, FIN2-nature) render difficult to provide a global picture about the completeness and the availability of the UOE finance data. The availability of data can be considered at country level (analysing the number of countries that reports/does not report a value for a specific variable) but also at the variables level (identifying the variables and/or the ISCED levels for which missing data flag M and data included in another category flag X for the entire geographical scope represent a large share of the requested data) Educational expenditure by level of education, source and destination (FIN1-Source) Household expenditure and expenditure of other non-educational private entities show the highest percentage of missing data (M) or data included in another categories (X). They account for 30 % and 24 % of the requested data respectively. The availability of data is much higher for the government source than for private source. Within the governement sources, the local level is the one that shows the highest percentage of missing data or data included in another category (code M and X) with 15.4 %. This percentage is almost equivalent to the one of the central level (15 %). 5

6 Analysing the availability of data (against the total set of requested data) by single variable confirms the above statement. Missing data or data included in another category (M and X) account for more than one third of the requested data for each item of households payments outside educational institutions (variables H15, H16 and H17 2 ) as well as for the total household expenditure on education (H20). High percentages of unavailable data (M+X) also affect direct payment of households to governement-dependent private institutions and private institutions (variables H2 and H3) and fees paid by households for ancillary services (H5b). For "expenditure of other non-educational private entities", the percentage of unavailable data (X+M) is the highest (and ranges between 25 % and 30 %) for payment to independent private institutions (E3), for payment of private enterprises for specific educational activities (E5a) and for fees paid to institutions for ancillary services (E5b). For government expenditure, the variable "public grants attributable for tuition fees to educational institutions" (G10a) is the one that show the higher percentage of missing/included in another category data (26 %). For some variables of the FIN1-Source questionnaire, more than 80 % of the countries do not provide any value by ISCED level or reported that the phenomenon is negligible (i.e. the value for the variables is reported as null). This is for instance the case for "Transfers from international sources (for education) to local government (F8), Transfers from international sources (for education) to regional government (F7) or Regional government - Transfers and payments to other private non educational entities (R13) Education expenditure by level of education, type of institution and nature (FIN2-Nature) Expenditure in private institutions show the highest percentage of data coded M (missing) or X (included in another category). Altogether, these two codes account for slighlty more than 25 % of the requested data for government dependent private institutions as well as for independent private institutions. When considering the nature of expenditure, the percentage of missing data is lower than 8 % whatever the nature of the expenditure concerned but the percentage of data flagged as being included in another category (X) ranges from 15 % (expenditure for R&D activities) to 27 % (current expenditure non teaching staff). Whatever the type of institutions concerned, more than one fourth of data requested about expenditure on non-teaching staff is coded as being included in another category (from 25 % for independent private institutions and government dependent private institutions to slighlty more than 31 % in public educational institutions). "Other non salary compensation" is another nature of expenditure that presents high percentages of data included in another category: 24 % for governement dependent institutions and 26 % for public institutions ISCED levels The separation between the general and vocational programmes at upper secondary education (ISCED level 3) and post-secondary non tertiary education (ISCED level 4) raises some difficulties for the two UOE finance questionnaires. In both questionnaires, the percentage of missing information for upper secondary programmes is below 6 % but the percentage of data (out of the total requested data for a given ISCED level) included in another category reaches 22 % or more. A similar pattern is observed for post-secondary non tertiary vocational programmes. Another issue when considering the provision of data by ISCED level is to assess the extent to which countries are able to allocate some types of expenditure to ISCED levels. In FIN1-Source and FIN2-Nature, the category "not classified by level" was used by more than one fourth of the responding countries and mainly concerned total central government direct expenditure in public institutions (C1); central government expenditure designated for 2 Payment for goods required directly or indirectly by educational institutions (H15), payment for goods not required for participation (H16), payment for private tutoring (H17) 6

7 capital (C5a) and local government direct expenditure in public institutions (L1); expenditure for staff compensation (X6); expenditure for salaries (X7) and for retirement in public institutions (X8) as well as expenditure other than staff compensation and capital expenditure in public institutions (X13 and X15). 3.3 Comparability issues for expenditure data This section focuses on issues that are identified in the structure of the UOE quality report which is part of the UOE data collection for members of the ESS. The analysis of the comparability is sometimes hampered by the low level of details of the information available in the quality reports. Even if they relates sometimes to optional variables, several comparability issues can be identified. Expenditure at Pre-primary level Childcare component and evening child care: Eight Member States do not exlude the childcare component at pre-primary level. When detailed, the main reason is that of integrated programmes from which it is not possible to exclude the childcare component. Only three Member States declare including (at least partially) evening child care in expenditure data. Cost of a dual-system programme to the employer: Six Member States declare that their UOE return do not cover this type of expenditure or only partially cover it. Households'payment for private tutoring: Households payment for private tutering is not covered in six Member States or cannot be distinguished from the total expenditure outside educational institutions. Student living costs: Students living costs are unevenly covered across Member States. Four of them do not cover this item in their UOE return while two Member States report student living costs with some deviations from the UOE methodology (e.g. both subsidised and non subsidised student living cost are reported). Ancillary services coverage and expenditure: Among the countries that reported the coverage of ancillary services in their national quality reports, most of them cover meals, transportation and dormitories in their UOE return, nearly half of them report health services and dining halls while two of them cover special need services. 3.4 Coherence with national accounts In national quality reports, data providers are requested to specify the extent to which UOE finance data set is reconciliable with National Accounts in terms of calculation methods. The information available on these issues does not provide sufficient information about the coherence between the national UOE data and the implementation of the ESA10 in their national context. Most countries only mention that UOE expenditure data are compiled on a cash accounting basis while National Accounts data are compiled on an accrual accounting basis. Additional differences between the two methodologies are mentioned in some national quality reports but remain too generic which suggests that they are not drawn up from national practices but from a general analysis of the two methodologies. These additional elements covers the assignment of expenditure to institutional units (final expenditure vs production value); the treatment of flows e.g. transfers, intermediate consumption, the treatments of loans and child allowances. It is noticeable that few countries detail that some or all of their data follow some aspects of the National Accounts methodology. 7

8 4 UOE finance pilot data collection exercise It has been acknowledged from a long time by various stakeholders (e.g. DG Education and Culture, Eurostat, Member States themselves) that the length of time between data availability and the event or phenomenon they describe (i.e. timeliness) is too long for UOE data on expenditure to meet users needs. More recently the Indicator Expert Group (IEG) on education investment recommended "to publish more data, to have them more timely (i.e. reducing the present nearly 3-year time lag) and with sufficient/logical breakdowns. In particular, regarding more timely data, it was recommended that a core set of variables should be identified to be provided with a shorter time lag. Moreover, this lack of timeliness raises concerns since other sources of information such as COFOG 3 data which follow a different methodology and disseminate less detailed information could provide some data on education expenditure with a better timeliness. Such situation explains why a UOE finance pilot data collection was set up to investigate the possibility to produce fresher/more up to date and timely finance indicators. 4.1 UOE finance pilot data collection: characteristics and key milestones During the preparation of the pilot data collection exercise, the major stakeholders - i.e. EU Commission DGs (DG EAC, DG ECFIN, DG EMPL and DG RTD, EACEA-Eurydice) and OECD - were consulted. They expressed their needs in terms of timeliness and timing for the UOE finance data and in terms of indicators. The outcomes of the consultation process as well as the main characterstics of the pilot data collection exercise were presented and discussed during the 2016-ETS Working Group meeting (Item 3.2. of the agenda). It was agreed that participation to the pilot data collection would be on a voluntary basis. The meeting was also the opportunity for members of the ESS to raise some concerns about the scope of the pilot data collection exercise (i.e. the large number of data/indicators selected for the pilot data collection), about the challenging trade-off between accuracy and quality on UOE data and about the obstacles in their current administrative organisation, in particular in federal countries, to provide fresher UOE finance data. During the meeting, only three Member States (France, the Netherlands and Sweden) informed Eurostat that they would participate to this exercise while other Member States needed to consult internally or informed that it would be difficult for them to participate. The pilot data collection exercise was launched on the 11 July 2016 and aimed at collecting UOE finance data for reference year The questionnaire sent for the pilot exercise had the same structure as the regular UOE finance questionnaires (FIN1-Source, FIN2-Nature and FIN-Students) and highlighted the variables of special importance ("the core set") in order to allow participating countries to focus only on some specific parts of the questionnaire. Participating countries were asked to provide as much data as they could according to their possibilities (i.e. countries had the possibilities to provide part of the requested information only) and were requested to send back the completed questionnaire by the 30 December A reminder was sent on the 11 January 2017 and set up a new deadline to the 31 January Participation to the UOE finance pilot data collection exercise Seven countries (France, Hungary, the Netherlands, Poland, Sweden, the Former Yugoslav Republic of Macedonia and Serbia) 4 and the Belgium-Flemish Community participated to the UOE finance pilot data collection exercise and sent UOE finance data for reference year 2015 while 20 countries informed that they would not participate. 3 COFOG Classification of the functions of Government. 4 UK returned its regular UOE2016-Finance data for the purpose of the UOE pilot as the regular UOE2016 finance return covers the financial year ending in 2015 in the UK (i.e. 1st April 2014-end March 2015). For this reason, UK is not counted here as having provided data for the specific purpose of the UOE finance pilot data collection. 8

9 Seven countries did not send a specific feedback but some of them already declared during the 2016-ETS Working Group meeting that they would not participate. ESS Member States Pilot data collection finance data Participation BEnl, FR, HU, NL, PL, SE, UK*, MK, RS No participation BEfr, BG, CZ, DK, DE, EE, IE, EL, ES, HR, IT, CY, LV, LT, LU, MT, AT, PT, RO, SI, SK, FI, CH, IS, LI, NO, TR Figure 1: Participation to the pilot data collection (reference year 2015) The low participation in the pilot data collection on education expenditure data is explained by two complementary reasons: the lack of resources and the unavailability of data in due time at national level. ESS Member States Pilot data collection finance data Participation No participation BEnl, FR, HU, NL, PL, SE, UK*, MK, RS Lack of human resources LU, SI IS, LI, NO DE, IE, CY, LV, LT, MT, PT, FI Data not available BG, EL, ES, IT, AT, RO CH No specific reason BEfr, CZ, DK, EE, HR, SK, TR Figure 2: Participation and reasons for non-participation to the pilot data collection (reference year 2015) Thirteen countries pointed out that the lack of human resources prevents them to participate to the pilot data collection. This lack of resources was reinforced by the simultaneity of the pilot data collection with another data collection in Cyprus (AES), in Lithuania (CVTS), in Finland (UOE16) and in Norway (CEDEFOP). One specificity of the UOE finance data return is that it needs the consolidation of several data sources that have their own statistical processing timeline and which are usually out of the control of the UOE data providers. Therefore, in addition to the lack of human resources, the unavailabilty of finance data in time also explains why Germany, Ireland, Cyprus, Latvia, Lithuania, Malta, Portugal and Finland did not participate to the exercise. The time when finance data is available from other national sources (which should also be considered in cunjunction with the time needed for data treatment to complete the UOE finance questionnaire especially for federal or decentralised statistical systems) explains also why Bulgaria, Greece, Spain, Italy, Austria, Romania and Switzerland did not participate to the pilot exercise. The results of the pilot data collection exercise show three distinct groups of countries. A first group of countries (France, the Netherlands and Sweden and to a lesser extent Poland) provided as much data for the pilot exercise as for the regular UOE-2016 data collection. In these Member States, the percentage of missing data is usually null or below 7 % (France and Poland, both for FIN1-Source) whatever the status (mandatory or optional) of the variable concerned. In the second group of countries (Belgium - Flemish Community and Hungary), the availability of data in the pilot is far lower than the one in the regular UOE data collection. In this group, the percentage of missing data for mandatory variables for FIN1-Source is greater than the one for mandatory variables for FIN2-Nature. Finally, the two candidate countries (the Former Yugoslav Republic of Macedonia and Serbia) present a low availability rate of mandatory data (below 15 %) for both FIN1-Source and FIN2-Nature. In the case of Serbia, the percentage of missing data in the pilot exercise is more or less similar to the one for the regular UOE-2016 data collection round. 9

10 Most of the countries that participated to the pilot exercise used the statistical infrastructure usually dedicated to collect data for the regular UOE Finance data collection, be it an educational account (France and the Netherlands), a dedicated database on education expenditure 5 (Sweden) or such kind of database complemented by an ad hoc consolidation/compilation of data from various sources for the specific purpose of the UOE data collection (Hungary and Poland). Four of them would be able to provide the UOE-finance data more or less between 10 and 12 months earlier than the deadline defined in Commission Regulation 912/2013 on a regular basis. The main conclusion from the pilot exercise is that it is not possible for nearly all members of the ESS to collect fresher UOE finance data within the existing context. The exercise also reveals that very few Member States are able to provide the UOE finance data more or less one year earlier than the Commission Regulation 912/2013 deadline. It would be beneficial for all to learn from the experience of these few Member States: what is the statistical infrastructure they put in place? How long did it take to build such capacity? What was the impetus for such development? 5 Quantification and reconciliation exercise 5.1 Objectives An inventory of the main conceptual differences between the UOE framework and the National Account framework has been conducted as a task of the contract (Task 3). The report about this inventory is under completion and will be disseminated to members of the ETS-Working group for information and possible comments during autumn This report will cover various aspects such as the scope of education (incl. formal/non formal education, the implementation of the ISCED classification, the scope of R&D, ancillary services, etc.), the geographical scope, the sectors (i.e. the classification of institutional units) as well as the transactions (incl. the accounting principles, loans, transfers, etc.). The reconciliation exercise (Task 4) aims at quantifying the methodological differences between "UOE" expenditure data and National Accounts data on education and to summarise them in a reconciliation table. The reconciliation procedure will foster, as much as possible, the transparency between the two methodologies by defining whenever possible the sequences of adjustments from the one to the other. On the basis of the analysis of the main conceptual differences between the UOE methodology and the National Accounts methodology, a theoretical framework of the reconciliation tables is currently developed. This theoretical framework aims at defining, for each relevant concept, the step by step "walk" from one methodology to the other thanks to a description of the sequences of adjustment (i.e. data treatment). In addition, the exercise will list and document the elements used in the adjustment processes and corresponding data needs. In short, the main purposes of the reconciliation exercise are to document in details both production processes regarding education expenditure data to explain and evaluate the discrepancies between the two datasets. 5.2 Testing the reconciliation framework: study visits Once it is finalised, the theoretical framework should be tested in six volunteer countries: populating the reconciliation tables is the best way to ensure that such approach is feasible, operational and relevant. Besides refining the theoretical framework, this testing phase will also allow enrich the documentation of such process. 5 i.e. the consolidation of various sources processed on a regular basis at national level to provide data on expenditure on education that fulfil national or/and international data needs. 10

11 The study visits will take place between July 2017 and October 2017 and the participation to this exercise is on a voluntary basis. The on-site visit will be performed by a team of two experts (one UOE expert and one National Account expert) appointed by the contractor. Because the on-site visit is of a very short duration (1 day), the exercise implies a preparatory phase (including the preparation of the necessary documentation of the UOE and NA data sets) and a follow-up phase (including the validation of the results). This exercise would especially benefit from the participation of two categories of countries: Countries having already completed such type of work (because they have developed a satellite account on education/an educational account or performed a reconciliation exercise with National Accounts); Countries that are currently implementing such reconciliation process or any project aiming at better understanding the links between the UOE and the NA methodologies. Ideally, the set of participating countries should be geographically balanced and show sufficient diversity in terms of organisation of education system and funding. Finally, the list of topics/issue that will be discussed during the on-site visit will be adapted in cooperation with the host country and agreed prior the meeting. Whenever needed, the volunteer country will be asked to provide the documentation and the data for the full set of reconciliation tables that are part of the exercise. Indeed, since this exercise is limited to six countries, it will only deliver relevant outcomes if a large variety of national situations are covered for a given topic. 5.3 Topics/issues to be considered during the reconciliation exercise Several dimensions should be covered by the reconciliation exercise at national level. These dimensions relate to the data sources, the definition of the scope of education and institutional sectors, and the transactions. These dimensions are briefly presented below Characteristics and timeline of the various data sources At national level, the UOE data and the National Accounts data returns compile and consolidate data from various sources of different nature (e.g. administrative data, sample surveys, etc.). The reconciliation exercise could possibly be the opportunity to run a cross examination of financial data sources that are used by the UOE national data return and those used by national accountants. Besides being a preamble to the reconciliation exercise, this step would allow: To document the situation when different sources are used to compile data for similar institutional sectors or similar transactions/types of expenditure; To identify new potential source(s) of data for the UOE finance questionnaire or for national accountants and; To document the timeline of each source in relation to the compilation of data for the UOE questionnaire and for the National Account compilation process Scope of education The preliminary results of the theoretical analysis of the UOE framework and the NA framework show that they differ in their definition of the scope of education including the coverage of formal and non-formal education, the implementation of the ISCED classification, the scope of R&D, the treatment of teaching hospitals and the coverage of ancillary services/subsidiary services to education. The impact on the data of each issue will be assessed during the reconciliation exercise. Formal/non-formal education: would National Accounts be able to separate non-formal education from formal education? Whenever possible, how national practices (in both frameworks) are able to differentiate formal education and non- formal education? ISCED classification: the ESA2010 framework has not yet implemented ISCED2011 and the timetable for such implementation is not known. Has this issue been considered by national accountants at national 11

12 level? How potential divergences with UOE data mostly for early childhood education, post- secondary non tertiary education and tertiary education are envisaged at national level? Ancillary/subsidiary services to education: the UOE and the ESA2010 framework use different terminologies that need to be clarified in order to fully assess the scope of these services. Ancillary services as defined in the UOE correspond to part of the subsidiary services to education (COFOG classification) which themselves encompass also the administration and inspection of education. In COFOG, "subsidiary services to education" is a separate category that is not allocated to ISCED levels. Moreover, neither the UOE nor the ESA2010 provide an exhaustive or comprehensive list of services covered by this category. National practices in reporting such category are potentially different and should be considered in the reconciliation exercise. Scope of R&D: in the UOE framework, R&D is not domain-oriented and covers R&D performed in tertiary educational institutions while the ESA 2010 framework (COFOG) only covers R&D on education as well as grants, loans and subsidies to non-governmental bodies (e.g. research institutes) that may not be part of tertiary educational institutions. Are data providers using the same source of data to identify R&D falling in the scope of their framework? Is the granularity of data sufficient to cover the scope as defined in their respective framework? Teaching hospitals: only part of the activity of teaching hospitals (teaching costs, educational research and curriculum development) falls within the scope of education in the UOE framework. This is also what the ESA 2010 recommend 6 but allows to "assign all outlays by multifunction units to whichever purpose appears to account for the largest part of total outlays" when apportion of outlays by function is not possible. Are practices aligned at national level? Definition of institutional sectors/units The UOE framework does not define institutional units per se but merely defines the sources of funds (section of the UOE manual) and the services providers (section 2.4 of the UOE manual). Unlike the UOE, the ESA2010 manual explicitely defines the institutional units and their grouping because they are considered as one of the basic statistical units used in this framework. The reconciliation exercise will aim at identifying if discrepancies between the two frameworks exist in classifying financing units and producing units at national level. This would especially concern the allocation of other private entities, the treatment of the rest of the world in the UOE (financing units) as well as the allocation of producing units (e.g. public non-instructional educational institutions and independent private (instructional and non instructional) educational institutions) Reconciliation of transactions/type of expenditure: data processing/data treatment The step by step walk from one methodology to the other (i.e. mainly from the UOE perspective to the National Accounts perspective) needs to be detailed by type of expenditure (FIN1-Source) and by the nature of the expenditure (FIN2-Nature). Whenever relevant, the impact of the different accounting principles (cash vs. accrual) needs also to be considered for each type of transaction during the exercise. For instance, this implies to reconcile current expenditure other that compensation of personnel and expenditure for compensation of personnel (salaries, expenditure on retirement and expenditure on other non-salary compensation) with intermediate consumption and compensation of employees respectively; capital expenditure with gross capital formation; financial aid to households and students with social benefits other than social transfers in kind, etc. From a practical point of view, this would be achieved through the compilation of detailed data from the host country, the documentation of the data treatment from both methodologies at national level; the comparison of gathered data from the two methodologies and the documentation explaining the reasons for existing convergence and discrepancies. 6 If it is assumed that it follows article 19 of the UNSD Classification of Expenditure According to Purpose (article 19). 12

13 6 Conclusions and further steps Without prejudice to the final outcomes of the ongoing contract and in addition to the efforts to be made by countries to improve the coverage and the quality of their data, further activities to improve expenditure data on education probably need to be developed in three key areas: improving the comparability of data and the UOE methodology on education expenditure, developing the consistency between the UOE methodology and the national accounts methodology and defining a strategy to improve the timeliness of UOE finance data. Improving the comparability of data and the UOE methodology on education expenditure: o The assessment of the UOE finance data (Task 1) shows that some progresses still need to be achieved in terms of availability and comparability of the data especially for private expenditure and for the general/vocational breakdown at ISCED level 3 and 4; o The current UOE national quality report has limits for a comprehensive assessment of the comparability of UOE finance data. Eurostat will propose some amendments to improve the contents of the national quality reports. The up-dated version should allow a regular monitoring of the comparability of UOE finance data without increasing the workload of reporting countries. o Apart from reasons linked to the reality of each national statistical system, there is room for improving the current UOE methodology on education expenditure. Eurostat envisages different complementary actions in this respect: The section about expenditure of the UOE manual needs to be refreshed, clarified and improved. This would for instance mean to develop the set of existing definitions, to improve the consistency of the wording and to develop more detailed guidelines to support national data providers; Comprehensive work about specific concepts such as "transfers and subsidies", "student loans" would also allow a clarification of the indicators that are currently disseminated. Developing the consistency between the UOE methodology and the national accounts methodology: The improvement of the UOE manual will benefit from the cross analysis of concepts and the reconciliation exercise between the UOE and National Accounts and would possibly lead to a better alignment between the two methodologies. Defining a strategy to improve the timeliness of UOE finance data: Eurostat takes note that it is impossible to collect fresher UOE finance data in the short term and even in the medium term. For most countries, the results of the UOE finance data pilot exercise show that existing statistical infrastructures do not allow collecting UOE finance data earlier than planned in the current Commission Regulation. o For the short term, Eurostat suggests initiating discussions during the proposed workshop (please see below) on the best ways to better meet the current deadlines for UOE finance data as set up in the Commission Regulation; o For the medium and long term, Eurostat proposes to launch discussions with Member States to consider if it is worth considering alternative strategies (e.g. early estimates procedures for a limited set of core indicators) to better meet users' needs in terms of freshness of the UOE finance data. The ongoing contract is a component of the long term project aiming at improving the overall quality and the timeliness of expenditure data on education. Eurostat proposes to present the results of the contract and to discuss further steps of the project toward the improvement of expenditure data on education during a dedicated workshop that it will organise in February This workshop will be the opportunity for Eurostat and for Member States as well as for OECD and UIS to have an in-depth exchange of views about the developments needed to reach a substantial improvement of expenditure 13

14 data on education. For the three key areas mentioned above (improvement of the UOE methodology, better alignment with the National Accounts and improvement of timeliness), the workshop would aim at clarifying the final objectives to reach in order to better meet users' needs, prioritizing these objectives and setting up the timeframe for dedicated actions. If deemed relevant, the workshop would also provide recommendations to the 2018-ETS Working Group for setting up a dedicated Task Force in charge of these issues and its mandate. 14