A European Socio-economic Classification: How we got here and where we are going More David Rose & Eric Harrison Institute for Social and Economic Research University of Essex

Eurostat Statistical Harmonisation Programme Aims to create a common set of core units core variables and core classifications for use in European and national social statistics

ESeC Expert Group appointed by Eurostat in January 2000 Proposals for an ESeC made in 2001 Feasibility Report (see This report forms the basis for the project

Form of classification. The outline classification takes the form of a two-level nested hierarchy, similar to classifications such as the French PCS. In its disaggregated form (level 2, what we term ‘Socio- economic Groups’ - SEGs) it covers the whole population at the individual level. There are SEGs to cover various ‘other active’ and inactive groups. Individuals who are allocated to these groups on the basis of current status may then be allocated to ESeC classes in a variety of ways, depending partly on analytic purpose and partly on the group concerned.

Possible ESeC ‘Classes’ (Level 1) 1.Large employers, higher managerial and professional occupations 2.Lower managerial and professional occupations 3.Intermediate occupations 4.Small employers and own account workers 5.Employers and self-employed in agriculture 6.Lower supervisory and lower technician occupations 7.Lower services etc occupations 8.Lower technical occupations 9.Routine occupations 10.Never worked and long term unemployed

Conceptual basis for the NS-SEC (Goldthorpe) Employment relations and conditions are central to delineating the structure of socio- economic positions in modern societies

The Derivation of the NS-SEC Basic SEC Positions EMPLOYERSSELF-EMPLOYED WORKERS EMPLOYEESEXCLUDED

Dimensions of work as sources of contractual hazard Difficulty of monitoring Specificity of human assets low high

Typical elements of the Labour Contract Short-term exchange of money for effort Payment by the time or piece No occupational pension or health scheme Contract easily terminated Low level of job security

Typical elements of the Service Relationship Long-term exchange of service for compensation Greater job security and employability Salary Incremental or similar payment systems Occupational pension and health schemes Greater control over the job and thus trust between employer and employee

Dimensions of work as sources of contractual hazard, forms of contract and class locations Difficulty of monitoring Specificity of human assets low high Labour contract Service relationship mixed

The conceptual derivation of ESeC Basic SEC Positions EMPLOYERSSELF-EMPLOYEDWORKERSEMPLOYEESEXCLUDED LABOUR CONTRACT Form of employment regulation SERVICE RELATIONSHIP MIXED Higher prof Lower prof/ Tech OtherAgric etc Higher prof Lower prof/ Tech OtherAgric etc Higher prof Lower prof/ Tech Other Lower SupProf /Tech ManProf Man Higher Never worked LargeSmall Super- Serv- Lower routine visory/ ices technical Technician Clerical Sales Services Clerical Sales Services Professional managerial, etc Unemp- loyed

Underlying ESeC ‘Socio-economic Groups’ (Level 2) 11.Employers (other than in agriculture) with 10+ employees 12.Farmers with full-time employees (or ‘large business’ farmers) 13.Higher managerial occupations 14.Higher professional occupations (employees) 15.Self-employed professional occupations Class 1 Large employers, higher managerial and professional occupations

Other active groups 01.Other unemployed 02.Unpaid family workers 03.National service Inactive groups 04.Retired 05.Students (full-time) 06.Children 07.Permanently sick and disabled 08.Looking after home Not classifiable 00.Not classifiable (occupations not given or inadequately described etc.)

Classification rules for the individual level of ESeC The ‘other unemployed’ in SEG 01, unpaid family workers in SEG 02, national service personnel in SEG 03 and the inactive SEGs do not immediately collapse to any class. Rather, individuals in these groups are (re-) allocated to the group of their ‘career typical’ (usually last ‘main’) job or to their household class.

Household level rules The household level of this classification would work in a similar way, except that the ESeC class position (level 1) would be allocated through a household class measure. In this case, those in SEGs and 00 would be allocated to their household class. Equally, those allocated to SEGs would take on the ESeC values of their household.

Flexibility One of the advantages of a nested two-level schema such as this is that it will permit analysts to look ‘inside’ classes. This will assist them in understanding how life-chances may vary between groups with the same employment relations. For example, do higher professionals in SEGs 14 and 15 have better health outcomes when compared with higher managers in SEG 13?

Number of SEGs (1) As far as the number of SEG categories to be recognised within each class is concerned, this is partly an issue of face validity, i.e. of grouping together in sub-categories similar types of occupations that share similar employment relations. It should be noted that the SEGs within this outline classification are only postulated ones, designed to help illustrate how a possible two-level classification might work.

Number of SEGs (2) However, it is the classes themselves that will need to be validated. Which SEGs we then wish to recognise within each class will be largely a matter of contingency, depending upon, for example, what might be useful for the internal analysis of classes, face validity issues, etc. The question of which SEGs ‘exist’, therefore, relates to which useful class sub-divisions we might wish to make among those combinations of occupation and employment status that share similar employment relations.

Criterion validation of ESeC (1) We have a measure similar to ESeC that has already been created and validated using the method of collecting employment relations data at the level of occupations. This is the UK NS-SEC. We will also build on previous European research aimed at developing a comparative measure of social structure similar to ESeC, for example the Comparative Analysis of Social Mobility in Industrial Societies (CASMIN).

Criterion validation of ESeC (2) Given the broad similarities of market economies and occupational and industrial structures across the EU, we can expect that employment relations will also be similar. Thus, it is reasonable to begin by creating an ESeC derivation matrix with cell values based on UK employment relations data. These data were collected in the 1996/97 winter quarter of the LFS.

What we have to do We have constructed a derivation matrix on the basis of the best available evidence we have on the employment conditions typical for the occupational unit groups of ISCO88(COM). This evidence was drawn from work for the project that produced the UK National Statistics Social- economic Classification (NS-SEC), as well as from earlier work on social mobility by Erikson, Goldthorpe and their colleagues and more recent work by academics on employment relations in Europe. In addition, the matrix is now being examined by NSIs, partners in the project and other nominated academic experts.

Constructing ESeC In order for an ESeC to be fully operationalized in line with our theoretical model, at a minimum we require measures of occupation, status in employment and, in some cases, enterprise size. We also believe that labour market position should be part of what an ESeC measures. In addition, some measure of farm size may be necessary, too, in order to distinguish capitalist farmers from other (e.g. subsistence) farmers. How, precisely, are these common elements to be measured? Do all the datasets we intend to use have these measures in the form required?

Occupation For the most part occupation is measured either by (4-digit) ISCO88(COM) or by a national occupational classification similar to it. France is exceptional in this regard, but has developed a Table des Correspondances between the Catégories Socioprofessionnelles (CSP) and ISCO88(COM). ISCO88(COM) is a core variable for the Eurostat harmonisation programme and so is the obvious measure of occupation to use for ESeC.

Status in employment All SECs distinguish between employers, the self- employed (own account workers) and employees. In the EU context, we may need to add the category of family worker. The EU harmonised variable is ICSE-93.

ICSE-93 1.Employees 2.Employers 3.Own account workers 4.Members of producers’ co-operatives 5.Contributing family workers 6.Workers not classifiable by status

Labour market position It is necessary to distinguish more than activity status. Our theoretical model requires us to discriminate between employers by size, the self- employed, and between managers (by size of enterprise or preferably managerial level), supervisors and other employees. Managerial status will be dependent on allocation to Major Group 1 of ISCO88(COM). Thus, labour market position involves a combination of ICSE-93, enterprise size and supervisory status.

Number of employees The size cut-off for enterprise size in the non- agricultural sector varies across the national SECs: 1-9, 10+; 1-24, 25+; 1-49, 50+ or combinations of these. However, since ISCO88(COM) is the harmonised occupational classification, then the initial simple rule for ESeC will need to be that employed by ISCO for managers and employers – 1-9 and 10+.

Example illustration of parts of the ESeC derivation matrix Note: in this table, for simplicity’s sake, we assume a seven-category empstat (i.e. that ‘farm’ can be established via ISCO). Empstat ISCO OUG S/emp 10+ S/emp <10 S/emp None Manager 10+ Manager <10 Super- Visor Employee 12xx111XX113XXX 13yyX441442X221XX 3xxx XX222 3yyy XX222 5xxx XX226333

Reduced ESeC Some data sets may not contain all the elements required to create ESeC in the prescribed manner. However, it would also be possible to produce a ‘reduced’ form of ESeC for use where data on establishment size are not given. Naturally, the costs and benefits of this would have to be assessed for each member state. The reduced form could be derived in essentially the same way as the full form of ESeC, except that (ignoring the agricultural sector again) the employment status variable would only have five categories: 1.Self-employed with employees; 2.Self-employed with no employees; 3.Manager 4.Supervisor 5.Employee The ESeC category for self-employed with employees and for managers would be based on the modal employment status category for each occupation.

Simplified ESeC The simplified form of ESeC would be for data sets in which only information on occupation (i.e. on 4 digit ISCO OUG) is available. The primary rule would be that occupations (OUGs) are allocated to the ESeC category for ‘other employees’, except where these are in a minority within that occupation or an occupation has no ‘other employee’ status (e.g. managers). In these cases the ESeC category of the modal occupation by employment status combination would be used. Hence, for example, if within a particular OUG supervisory status predominates, then the ESeC value for supervisors in that OUG will apply.

Next Steps Create derivation matrices: done Matrices + report to partners, NSIs, Eurostat and experts for responses - done Statistical Compendium – being undertaken Validation studies – May to November 2005 Validation conference – January 2006 ESeC User Guide – mid 2006 NSIs’ Workshop – mid 2006