National Accounts Working Party 3 5 October 2007

National Accounts Working Party 3 -5 October 2007 Paris OECD handbook on the measurement of volume output of health and education Paul Schreyer, OECD/STD Sandra Hopkins, OECD/ELS

Contents • • • Background General concepts Education Health Way forward

Background: OECD Project • Strong and continued demand for output measures of education and health by policy-makers • European Regulation • Project started in 2005, endorsement by CSTAT • Builds on previous work: Eurostat Handbook on Volume and Prices, Atkinson Report, country experiences • Cooperation with the UKCe. MGA and Eurostat • Financial support by INSEE (France), Government of Norway, United Kingdom • Workshops in London (2006) and Paris (2007) • Objectives: – OECD Handbook by end 2008 – Data development

Background: An old question – what is new? 1. Joint work with sector specialists • Elaborated jointly with OECD’s specialised networks – Network of education experts – Network of health experts • Both networks have strong interest in measuring appropriate volume output

Background: An old question – what is new? 2: Joint treatment of temporal and spatial dimensions • Education and health PPPs are of great importance to analysts • PPPs and national accounts have to be consistent • Handbook deals with both dimensions in parallel

Background: An old question – what is new? 3: Joint treatment of non-market and market production • Even for market producers of education and health services, price-volume splits are not obvious • In particular, quality adjustment is difficult in both cases • Handbook emphasises non-market production and volume indicators but not exclusively – the principles should be the same for market and non-market production

Concepts and terminology • Distinction must be made between inputs, outcomes • Best explained by way of a graph

If outcome indicators are used for quality adjustment, they: • Should control for any other factors that affect outcome for consumers (e. g. socio-economic background of pupils, environmental impact on health) • Should only capture marginal effect of process on outcome

Quality adjustment • First step towards capturing quality change is the correct stratification, i. e. , the comparison of products with the same or at least similar characteristics. • However, matching of services has its limits. • Also, stratification should be able to capture effects of substitution • However, avoid treating goods or services as substitutes that are in fact different products • Explicit quality adjustment may make it necessary to invoke outcomes

Cost and value weights: principles • In a market context, changes in the price or quantity of products are weighted by their expenditure share reflecting relative valuation by consumers/producers • In a non-market context, only cost observations are available and there is no guarantee that cost weights reflect relative valuation by consumers

Cost and value weights: principles • 2 possibilities to deal with this problem: – Assume that on average, cost shares reflect also relative valuation by consumers – Impute relative valuation by consumer but • • total value of non-market output ≠ costs; difficult measurement issues; asymmetry with regard to treatment of other products not within the scope of national accounts although value weights are useful for welfare analysis • Handbook recommends use of cost weights

Cost and value weights: practice • Note: – Compiling cost or value information in the required classification is not a trivial task – Example: no data may be available on the cost or value of medical care by disease because pricing mechanisms, or cost accounting are not defined over episodes of treatment

Education Scope of education services Handbook covers only formal education services Focus is on secondary education

Summary of proposed measures: 1) Stratification Level 0 Minimum stratification Preferred stratification Pre-primary education All classes Normal classes or pupils Level 1 Primary education or first stage of Special classes basic education handicapped pupils or Normal classes or pupils Level 2 Lower secondary or second stage Special classes of basic education handicapped pupils Level 3 Upper secondary education Level 4 Post-secondary education or General + pre-vocational Vocational non-tertiary General / available vocational if

Summary of proposed measures: 1) Stratification (contd) Level 5 B Level 5 A + 6 Minimum stratification More practical and occupationspecific programmes tertiary education More theoretically-based programmes tertiary education Preferred stratification All classes or by professional purpose By fields of education and/or prestige of education unit, or by equivalences of degrees Adult general education Adult vocational education Computer training Adult and other informal education Adult and other education, anticipating extension of “education” content in ISIC rev 4, class 8540. Driving lessons Music lessons Other cultural and artistic lessons Sport lessons Recreational lessons Education support activities According to what will be retained Other education activities in class 8550 of ISIC rev. 4

Summary of proposed measures for education services: 2) variables Stratum Pre-primary education Primary education: normal Primary education: special or handicapped pupils Lower secondary: normal Lower secondary: special or handicapped pupils Upper secondary education: general + pre-technical or prevocational Upper secondary education: vocational Post-secondary non-tertiary education More practical and occupationally specific programmes tertiary education More theoretically based programmes tertiary education Quantity Pupil-hours Pupils Credits (ECTS) as 1 st best Combination of time-lagged degrees as 2 nd best Enrolled tudents as 3 rd best Quality (educational) None Comment The sub-stratification Contribution to scores normal / special could be replaced by coefficients reflecting the extra costs for social services provided to handicapped pupils Entry education status has to be controlled for, this can be with the help of a model Relative future real Only incremental revenues earnings and employment must be considered - real rate if no scores available earnings and employment rate “without teaching” have to be subtracted from total earnings or employment Differentiation by field of Different concepts but close education figures in practice. Relative “value” of level of degrees could be estimated from labour market

Education services: conclusions and questions • Stratification can go a long way towards constructing volume indices – but are process measures an acceptable proxy for a full quality adjustment? • A mix of quality-adjustment approaches is suggested in the Handbook – e. g. , scores for secondary education, degrees or a human capital approach for tertiary education. Would a single approach be preferable?

Health services

1. Aggregation by disease or illness Aggregation of quantities of services: • Health volume output can be measured at 2 levels: disease or institution 1. Aggregation by disease or illness • Ideally, health volume output should be measured by complete treatments by disease as this is the product which an individual purchases from a health provider. • Complete treatment refers to the pathway that an individual takes through heterogeneous institutions – offices of doctors, hospitals, medical laboratories etc. – in order to receive full and final treatment for a disease or condition.

1. Aggregation by disease or illness Benefits: – “Our concern should be not where the money comes from and where it goes but what it buys. ” (Triplett 2001) – The summing of points of contact with the health system to estimate a complete treatment means that if clinical practice changes over time, and is associated with a change in the cost of providing the service, this will be reflected in the output measure e. g movement to day-only surgery and non-invasive types of surgery.

1. Aggregation by disease or illness Problems: – In SNA, total output of an activity is based on summing up outputs of various service providers. Principle is directly applicable only if the service provider is the same during the whole treatment. – Demanding data requirements e. g. linking patient treatment across providers, ability to determine the beginning & end point of treatment – Cost of illness studies require disease specific price indexes for conversion into volumes. Difficult in a nonmarket system

2. Aggregation by institutions Diagnosis Related Groups (DRGs) aggregate across a hospital treatment, usually acute episodes only – There is no international DRG system – Aggregation across other providers is problematic e. g. doctors, psychiatric hospitals etc.

2. Aggregation by institutions • Development and harmonisation of classification systems is required to ensure improvements in compatibility and comparability of health volume output both temporally and spatially. Developments proposed include a classification of health care products and international harmonisation of DRG systems for both inpatients and outpatients.

2. Aggregation by institutions • In the shorter run, it is possible to aggregate health volume output using currently existing DRG systems for hospital outputs, Resource Utilisation Groups for nursing home outputs and summing up activities in outpatient services.

Quality adjustment • Ideally, health volume output should be adjusted for the improvement in health outcomes which are due to the introduction into the health industry of new treatments as well as improvements in the existing practices. • Wealth of outcome measures and an ‘industry of quality measurement’ which compiles and records health outcomes, but at this stage the quality indicators which could be applied for adjustment to health volume output are rudimentary and under development • Developments should include consensus on what indicators should be used for quality adjustment and the role of some quality issues, e. g. waiting times, on health outcomes. Choice of quality indicators should emphasise internationally comparable and consistent measurement.

Way forward • Presentation of draft report to health experts next week • Possible input to Eurostat seminar November 2007 • Revision of report, and inclusion of education PPPs • During 2007/08: work of Taskforce on Health PPPs (supported by European Commission) • End 2008: complete draft report • 2009 and beyond: OECD will seek mandate to begin empirical implementation