Document Prepared by Cecile Brousse, France

02 CLASSIFICATIONS

Doc. ISCO/08/08/FR 26November 2008

______

Data collection on occupations

Document prepared by Cecile Brousse, France

______

Workshop on the data collection of occupational data

28 November 2008

Agenda item 3.3

STATISTICAL OFFICE OF THE EUROPEAN COMMUNITIES

02 Statistical governance, quality and evaluation

Tel. (352) 4301-33622, Fax (352) 4301-33899

Data collection on occupations

Eurostat, November the 28th 2008

Presentation by France

PLAN

 ISCO-2008

Before implementing ISCO-08 in the EU, some imprecisions in the new classification should be removed (page 2 to 5)

How to identify supervisors in household surveys? Presentation of a short survey on supervisory responsabilities (page 6 to 14)

 Prototype ESeC

The ESeC classification applied to occupational mobility in France: Another perspective (page 15 to 23)

How easy is it to classify oneself in the prototype ESeC? The first results of a module on self-classification (page 24 to 29)

A pilot survey on the understanding of the draft ESeC by the general public (page 30 to 34)

Before implementing ISCO-08 in the EU, some imprecisions in the new classification should be removed

Cécile Brousse (Insee, Employment unit)

Jérémie Torterat (DARES)

Some areas lacking clarity in ISCO-08

The tripartite meeting that adopted the new version of ISCO concluded with some areas of imprecision in the classification not being completely resolved. A certain degree of porosity between some major groups can be pointed out: 1/2, 1/7, 3/7 (no conceptual definitions of managers, supervisors, and the boundary between the two), 61-62/63 (no criteria given for distinguishing commercial agriculture from subsistence agriculture).

Company manager, tradesman, supervisor: definitions need to be more precise

ISCO-08, like the preceding version, does not offer a close analysis of the specifics of supervisory positions. Even though the definition of a manager might have been improved (someone who “supervises/manages a team”), no definition is given for a 'team'; this restriction is however already useful in English contexts since it at least excludes project managers. However, ISCO-08 does introduce the concept of supervision in certain occupations where the task of 'supervisor' would involve specific tasks. However, these specific tasks are not specified, nor is it clear whether they would be the same for different jobs where ISCO-08 imposes a distinction. This leads to a lack of clarity in the coding (see appendix).

The boundary between management and supervision is also left unclear. ISCO-58 defined supervision as the first hierarchical level. This criterion has the advantage of being relatively operational, but may not correspond to any social reality. It would be necessary to specify in a consistent way the specific tasks and responsibilities of management and of supervision in order to decide between borderline cases, such as co-ordination/coordinator[AJP1] in the first case and monitoring in the second...

These theoretical considerations may be directly illustrated by showing some of the difficulties of making a classification in ISCO.

-Category 1 or 2: where would one classify a ‘professional’ who coordinates a small team of two other people?

-Category 1 or 7: how do we classify a mechanic who, in addition to his work as a mechanic, manages his own garage which employs three salaried mechanics? It is all the more problematic that a shopkeeper (in a fairly similar situation) would be classified in major group 5.

-Category 1, 3 or 7: how do we classify a mechanic who runs his own small garage, where he also works as a mechanic, supported by a supervisor to whom he delegates some of his authority over eight salaried mechanics?

Occupations that may be carried out at different skill levels

Numerous social science research studies on social stratification across different countries disregard the influence of different skill levels between countries, with the methodological assumption of a convergence of occupations under the influence of a convergence of national economies. This hypothesis is more difficult to sustain in the context of public statistics: the labour market does not in practice have the flexibility at the European-level for such convergence to occur in the medium term - for example, some occupations are strictly controlled at State level, especially in the health field - and the situations in each country will have to be looked at carefully to harmonise coding.

The way an occupational field is divided into the various occupations included within it cannot a priori be assumed to be the same in different countries. For example, do we consider two nurses who are not authorised to perform the same tasks to be carrying out ‘the same job’? Identifying occupations internationally is far from obvious and cannot rely on the names that categories are given locally, within the same classification system. It is necessary to produce standard definitions of occupations.

These difficulties with identification naturally lead on to the problem of differences in training and skills in different countries for similar or comparable occupations. This question has particular importance in the context of the coding for the draft European Socio-economic Classification (ESeC), as differences in skill levels can affect an individual’s employment relationships. In addition, this aspect is also important to the extent that natural use of ISCO in the context of the EU could ultimately constitute an internationalisation of the labour market.

Non-commercial agricultural jobs

The fact that there is no rule to distinguish commercial from subsistence farming is also very problematic. This distinction does not exist in other international conventions, and there was a long discussion on the subject at the tripartite meeting. Several participants (Tunisia, France, Eurostat, the FAO, etc.) were against maintaining the separation. On the question of criteria, China cited the use of motorised machines as a criterion for making a local distinction in its classification. There was no consensus on whether to keep subsistence farming (the secretariat of the ILO argued that to remove this category would be to go beyond it mandate simply to update), and a decision on criteria was avoided.

It is clear that some farmers in the European Union could be considered close to a situation of agricultural subsistence (in France, in particular in overseas communities and départements); it is necessary that a European-wide standard is proposed (possibly the removal of the category, as was the case in ISCO-88(COM)).

Towards a European interpretation of ISCO-08?

The resolution adopted by the ILO in December 2007 sets out the following definitions:

-ISCO classifies jobs. A job is defined for the purposes of ISCO-08 as a set of tasks and duties performed, or meant to be performed, by one person, including for an employer or in self employment.

-An occupation is defined as a set of jobs whose main tasks and duties are characterised by a high degree of similarity. A person may be associated with an occupation through the main job currently held, a second job or a job previously held.

Differences in skills between European countries

Even assuming that the definition of ‘job cores’ is relatively easy, we still have the problem of possible differences in training and skills between comparable occupations (for comparable ‘job cores'). Hence our attempt at coding the most common occupations (see Appendix 5).

The commitment to remove parallel groups in ISCO-08 (comparable occupations carried out at different skill levels, identified and classified in different major groups) both simplifies the coding and also complicates the interpretation of data.

Coding is simplified by the relative automation of the coding. Parallel groups caused natural problems for harmonisation as the differentiation of skill levels was standard and therefore not necessarily directly compatible with the situation in each country (in relation to level, but also in relation to the number of different levels nationally).

Conversely, the use of data becomes problematic as the same unit group in ISCO will equate very different occupational situations. Since harmonised coding has an operational aim, it is necessary to consider problems linked to training and skills.

***

We will outline three independent but complementary paths for moving towards harmonised data coding. The result of the aim to standardise coding - obtaining optimal classification in a European socio-economic classification system - should guide attempts at harmonisation at every stage; it is also the opportunity to remove some ambiguities from ISCO, as a standard classification and to root it in a context. We will devote an additional, more specific piece of work to this.

Occupations that could belong to either the public or private sector

The de facto removal of ISCO-88(COM) categories corresponding to administrative civil service occupations means that individuals with training in multiple roles (and therefore with a specific and broad national ‘job core’, meaning that mobility between job domains is different to the mobility of individuals with careers in business) have to be classified in the functional categories in ISCO-08.

From an employment relationship perspective, differentiating occupations according to whether they are public or private sector may be debateable as far as occupations classified in the higher major groups in ISCO-08 are concerned. But from an occupational mobility perspective it is generally justified in relation to employment policies, which have to consider employment status, or in relation to a job’s skill requirements (competitive entry in the civil service).

Harmonising and enhancing the alphabetic index of occupations

According to the recommendations of the ILO, coding occupations in ISCO should be done using reported job titles. To this end the ILO provides indexes in English, French and Spanish.

The use of an index of job titles is unquestionably a major help in coding jobs in countries where one of the official ILO languages is spoken.

However, such a standard index does not solve all the coding problems in English, French and Spanish-speaking countries. Identifying and naming occupations is rooted in socio-linguistic contexts (national interpretations of the type of associated tasks, i.e. the ‘job core’ as defined above), which has to be taken into account to guarantee a classification standard. Comparable occupations may be given different names even in the same language (for example an ‘institutrice en maternelle’ (nursery school teacher) in France is a ‘jardinière d’enfants’ in Switzerland); or, conversely, the same name can refer to significantly different jobs (for example, ‘nurses’ are authorised to perform different tasks in different countries).

But quite apart from these problems of interpretation, non- English, French and Spanish-speaking countries don’t even have this partial coding standard. We cannot rule out the possibility that, in these countries, the socio-linguistic interpretation of ISCO, as translated for national purposes, (and therefore national codings in the classification) are different; this means that data collected in countries without an ILO index are even less likely to be comparable across Europe.

One way of improving the comparability of data coded in ISCO at the European level might be to establish a harmonised index of occupations with greater detail than the ILO indexes by taking into account nationally used indexes and national and European ‘job cores’.

The key point about a European index that takes into account national job titles, if necessary combined with skill level, is that it should be possible to match any job title to the best-fit European 'job core' (and thereby to a unit group in ISCO-08), regardless of the way the title is coded in the national classification system.

A quasi extensive list of occupations could be put forward in the context of an internet survey allowing the use of search trees; assigning an individual to a detailed category can then be considered fairly simple. However, in some sources (e.g. censuses) the standard procedure is voluntary reporting of occupations, and the quality of coding can be affected if individuals use inadequately precise job titles.

It is therefore standard practice to use supplementary variables, collected at the same time as the job title, to make sure the coding is robust.

Making coding more robust with the use of supplementary variables

The use of supplementary variables makes it possible to deal with uninformative job titles (from a statistical point of view) reported by individuals, but also to monitor direct coding, where possible.

If an individual reports that he is a ‘manager’ without giving further information, it is not possible to classify him in ISCO at a detailed level, but only at a higher level. It is necessary to know what exactly his role is and, in some cases, the line of business of the company that he works for.

To distinguish between occupations that can be performed at different identified levels (level of skill, an explicit element in ISCO) in the classification (for example, for someone reporting that he is a 'legal professional', this could be at professional or associate professional level), it is also necessary to have variables indicating professional level, or an approximation using qualification level.

For workers distinguished by their skill speciality (second explicit element in ISCO), it is also necessary to know the company or organisation’s line of business to make an uninformative job title more specific.

The combination of job title and supplementary variables makes it possible to define in detail the ‘job core’ to be assigned to individuals. Harmonising coding in ISCO may also therefore happen via a harmonisation of codings of core variables in Europe-wide surveys.

Choosing which variables to include and to examine in harmonising the occupational coding must happen not only according to the classification itself, but also with a view to offering consistency with the socio-economic categories derived from it. It is thus reasonable to consider using variables relating to company's line of business, number of employees, occupational position and role.

How to identify supervisors in household surveys? Presentation of a short survey on supervisory responsabilities

Daniel Verger, Cécile Brousse, François Gleizes (Insee)

Loup Wolff (Dares)

The revised ISCO includes 6 headings for ‘supervisors’. Most fields of activity are covered, as the following are now distinguished:

• mining supervisors (3121)
• manufacturing supervisors (3122)
• construction supervisors (3123)
• office supervisors (3341)
• cleaning and housekeeping supervisors in offices, hotels and other establishments (5151)
• shop supervisors (5222)

In French the term ‘supervisor’ (‘superviseur’) is rarely used, but the introduction of headings relating to the first level of supervision should facilitate correspondence with the French classification system as the PCS classification already distinguishes at level 2 the group ‘Foremen and supervisors’ (‘contremaîtres et agents de maîtrise’), 80% of whom have supervisory responsibilities (see section 1.1.2). This category contains 17 headings:

-Agriculture and forestry foremen and supervisors (non-managerial)480a

-Commercial shipping and fishing boatswains480b

-Works supervisors (non-managerial)481a

-Site supervisors (non-managerial)481b

-Electrical and electronic equipment manufacturing supervisors482a

-Machine manufacture and metal work supervisors483a

-Manufacturing supervisors: agri-foods, chemistry, plastics, pharmacy.484a

-Manufacturing supervisors: metal, heavy materials and other processing industries484b

-Energy, water and heating production and distribution supervisors and technicians485a

-Manufacturing in other industries (printing, flexible materials, furniture and wood) supervisors485b

-Electricity, electromechanical and electronic maintenance and installation supervisors486a

-Machine maintenance and installation supervisors486d

-General maintenance, installation and new works supervisors (except mechanical, electromechanical, electronic)486e

-Warehouse and storage supervisors487a

-Sorting, packing, dispatching and other handling supervisors487b

-Restaurant supervisors: kitchen/production488a

-Restaurant supervisors: running the restaurant488b

The absence of supervisor categories in ISCO-88 meant we had to classify foremen and supervisors in groups 4, 5, 7, 8 and 9 (see appendix 2), at the same level as the people they supervise (as recommended by the ILO). This solution was not very satisfactory.

The difficulty of defining supervisors

When we come to actually using these new headings for supervisors, there are three apparent difficulties. The first is not unique to France: several groups of supervisors do not seem to have a named place in ISCO-08, even though they work in important jobs such as warehousing, sorting, maintenance or catering. In our attempt to code some of the most common occupations in ISCO-08 we frequently came up against this problem (see appendix 3). Without adequate guidance, it does not seem possible to classify cooks, caterers, policemen with responsibilities for subordinates, and head stock-keepers. The ILO guidelines should offer some solutions.

Secondly, and this also perhaps applies in other countries, the industrial sector is organised around two levels of supervision: sub-foremen and foremen-supervisors; the PCS classification groups sub-foremen together with skilled workers in the same occupational groups (level 4) without being able to identify them as such, whilst foremen/supervisors are an integral part of the intermediate professionals group, forming a clearly defined sub-group (SC 48). If supervisor categories should include both sub-foremen and foremen, we will have some difficulties in establishing a direct correspondence between ISCO-08 and the French classification.

The third difficulty, and by far the biggest, is the lack of explicit criteria for distinguishing supervisors from managers and tradesmen. This point, which concerns all countries, will be discussed in section 2.3.

Supervisors’ reported occupations: insufficient information for them to be recognised

The 2006 Employment Survey (see appendix 3) makes it possible to identify those who have “at least one person under their authority or under their responsibility”. The proportion of supervisors varies from 67% amongst legislators and senior officials to 3% amongst unskilled services and sales employees. Two-thirds of supervisors do not spontaneously report their hierarchical position. The importance placed on hierarchical responsibilities is very variable across socio-occupational groups, with well-known differences between mainly male and mainly female occupations. Thus in sales and personal services, only 30% of unskilled employees who carry out supervision mention it explicitly. The terms most often reported are 'head', ‘team leader' or ‘organiser/coordinator’. Amongst sales and demonstration supervisors the rate is even lower: only 20% indicate a hierarchical position. Job titles are very varied but often include the term ‘head’. At the other end, half of operators (of plant or machines) and half of assembly workers with a supervisory role specify it in their job title: the most commonly used terms are ‘chief’, ‘head’, ‘co-ordinator', ‘supervisor’ and ‘controller’. Office workers are somewhere in between with 40% of ‘supervisors’ being identifiable from job titles such as ‘head’ or ‘chief’. For associate professionals, the differences between the tertiary and industrial sectors are less clear: 27% of technicians and 32% of health associate professionals voluntarily mention their hierarchical position.