EUROSTAT
Directorate F: Social statistics
Unit F-4: Quality of Life
Doc. 2013/2/5/EN
document for point 5of the agenda
TOPICS ON EMPLOYMENT
4th meeting of the expert group on quality of life indicators
Luxembourg, 8/9 October 2013
Bech – Room Ampere
At the previous meeting, the EXPERT GROUP on QUALITY OF LIFE INDICATORS decided to adopt a new structure for the dimension Productive or Main Activity, more in accordance with dimensions currently developed by UNECE/Eurostat/ILO Task Force and to further develop topics on Employment, both related to the Quality and Quantity sub dimensions.
The purpose of this document is to present the analysis done by Eurostat in developing new (in the sense of not yet published in the Eurobase) indicators on Employment, based on data coming generally from the Labour Force Survey. Relevant data will be presented, together with considerations about the meaning of the respective indicators and their quality.
The indicators and subsequent topics covered in this paper are:
- Sub dimension 2.1.3 Underemployment, quality: Over qualification rate
- Sub dimension 2.2.2 Health and Safety at Work: Hazardous Working Conditions
- Sub dimension 2.2.3 Work Life Balance: Long working hours, Atypical Working Hours, Parental Leave, Child care
- Sub dimension 2.2.4 Temporary Work: Long job tenure
1. Over qualification rate
The importance of this indicator is twofold, at individual level and at societal level. For the individual, working in a job that requires a lower qualification than the maximum achieved by him or her can have an important impact on self-esteem, job satisfaction and overall quality of life assessment. For the society it represents a suboptimal usage of its stock of human capital, which can hamper social and economic development both in the short and the long term.
Although its importance cannot be denied, the measurement of this indicator is not very straightforward and several approaches have been developed in this sense. The most well-known is constructing a matrix with educational level as one dimension and the ISCO code of the occupational classification on the other. The persons working in a occupational category below (or rather above, giving the fact that the ISCO classification ranks the most educational intensive occupational categories first) the level expected to be equivalent to their highest level of education achieved are considered to be overqualified. We propose to the Expert Group to use a subcategory of this group, namely the highly qualified (people having tertiary education) who are in this situation. Although the data is not harmonized enough between the countries (some occupations are not easy to be placed into a category or the other, and it is also difficult sometimes to assess the level of education needed to perform a certain occupation, and this can vary across countries or time), the indicator is currently being used in Eurostat publications and by policy makers.
The following is a graph presenting the data for each European country, for 2012. The indicator measures employed persons, with tertiary education (ISCED 5-6), working in jobs in an ISCO category higher than 3 (starting with category 4, clerical work support workers). Those working for the Armed Forces (ISCO code 0) are not included in the calculation of the indicator at all, due to the difficulties of assigning this occupational category to a certain educational level. The figure seems to be rather stable at the EU level (EU28 average), being at around 21% from 2007 until 2012.
Percentage of Overqualified employed persons, 15-64, 2012
Another way to look at this indicator is the subjective assessment of the person. A dichotomic variable on the subjective appreciation of being overqualified (Qualifications and skills of the person would allow more demanding tasks than current job) will be included in the 2014 LFS Ad-Hoc Module on The labour market situation of migrants and their immediate descendants. The data will be available in 2015.
The Expert Group is invited to give its opinion on the possibility of including this indicator in the Quality of Life disseminated scoreboard, using for the beginning the operationalization presented in the table above, at least until another one becomes available.
2.Hazardous working conditions
At the moment, there are 5 indicators disseminated for this subtopic: 1 coming from administrative sources (Accidents at work: standardised incidence rate) and 4 coming from the LFS Ad-Hoc Module on Health and Safety at Work (Accidents at work, Work-related health problems, and Exposure to factors that can adversely affect mental and physical well-being) which has been in collected in 2007 and is currently being repeated in 2013.
For parsimony reasons, a reduction of their number is desirable. There is at least a potential overlap between the indicator coming from administrative sources and the one coming from the survey and having the aim of measuring a similar issue (the rate of accidents at work). But the data is not easily comparable, due to the fact that the administrative data refers only to incidents that were fatal or severe (at least 3 days of leave), and that the incidence rate is standardized according to the structure of the economy (NACE codes) in order to account for the fact that some sectors are by default riskier than others. The standardized rate of fatal injuries is considered to be the most comparable indicator across member states[1], although it is affected by problems of underreporting in the New Member States. For the purpose of Quality of Life it may be preferable to use non standardized rates (not disseminated in the original table), because we are interested in the probability an employee has to suffer a fatal accident at work, regardless of the structure of the economy in the respective country.
However, the indicators coming from the survey refer to hazardous working conditions that were not fatal or necessarily severe, and therefore offer complementary information. Building a synthetic out of the 4 variables can be a solution for reducing the number of indicators. But the assessment of the module shows that issues of comparability between the countries are quite present[2]. For example, the use of proxy interviews can have an important impact, and the practice varied between countries in this regard. Also, the wording of the question on work related health problems was significantly different in 8 of the countries (not including mental health). Furthermore, it seems that these indicators cannot be clearly associated with either a formative or a reflective model. There is a degree of severity between them which is clearly visible in the average percentage of Europeans affected by them (3.2% reported accidents, 8.6% health-related problems, 27.9% exposure to factors that can affect mental health and 40.7% to those that can affect physical health).The one on Health-related problems could offer good complementary information to the indicator coming from administrative sources (as it is not extremely severe, nor extremely vague), although it is difficult to see patterns in 2007 data. The 2013 Ad-Hoc Module will be available next year in May, and the results will probably be more harmonized. It would be advisable to wait for this data in order to see if a synthetic is the way to go forward, or maybe choosing one of the indicators (for example Health-related problems) with a better discriminant power and more relevant for the framework.
The data in the table below shows that the two indicators measure different aspects of the phenomenon, they are not correlated. Romania, Bulgaria, the Baltic countries, Italy and Spain show a pattern of significantly higher than the EU average risk of having a fatal accident, but quite low rates of employed persons declaring they had a work-related health problem. This may be related to cultural reasons, but also to the way the question was formulated (including or excluding stress-related problems). At the other end, Nordic countries show remarkably low rates of accidents, but significantly higher than average percentages of employed persons declaring they had a health problem caused by working conditions.
Eurostat’s proposal is to use for the moment in the framework only the standardized rate of fatal accidents, and when data from the new Ad-Hoc Module will become available (May 2013) to decide if to choose one of the 4 indicators measuring Health and Safety at work, or to build a synthetic out of them.
The Expert Group is asked to give its opinion on the feasibility of this proposal.
Work related health problems / Fatal accidents(standardized rate)2007 / 2008 / 2009 / 2010
EU27 / 8.1% / 3.21 / 2.54 / 2.61
Belgium / 11.5% / 4.97 / 2.64 / 2.57
Bulgaria / 4.8% / 5.73 / 3.38 / 3.55
Czech Republic / 8.5% / 3.93 / 2.36 / 3.3
Denmark / 12.9% / 2.08 / 1.88 / 2.54
Germany / 5.2% / 2.44 / 1.39 / 1.59
Estonia / 9.0% / 2.95 / 3.09 / 3.82
Ireland / 3.0% / 2.54 / 2.54 / 3.79
Greece / 6.4%
Spain / 5.8% / 3.96 / 3.18 / 2.88
France / 1.83 / 2.9 / 2.69
Croatia / 8.1% / : / : / 2.8
Italy / 6.8% / 4.22 / 4.02 / 3.89
Cyprus / 8.4% / 6.56 / 2.86 / 6.72
Latvia / 4.0% / 5.71 / 3.98 / 3.45
Lithuania / 4.0% / 6.19 / 5.39 / 4.85
Luxembourg / 3.8% / 3.8 / 1.77 / 5.28
Hungary / 5.4% / 4.61 / 3.41 / 3.23
Malta / 4.0% / : / 4.49 / :
Netherlands / 10.9% / 2.71 / 2.1 / 1.58
Austria / 15.0% / 4.65 / 6.47 / 6.38
Poland / 21.9% / 3.91 / 5.66 / 3.96
Portugal / 4.5% / : / 5.6 / 5.26
Romania / 5.2% / 9.77 / 5.93 / 6.44
Slovenia / 10.2% / 3.62 / 3.2 / 3.5
Slovakia / 6.0% / 4.63 / 2.28 / 2.67
Finland / 24.5% / 1.47 / 1.71 / 1.88
Sweden / 14.3% / 2.14 / 1.92 / 2.11
Great Britain / 4.6% / 0.9 / 1.54 / 1.63
3. Work Life Balance
3.1 Long Working Hours
At the moment, the indicator included in the framework is the average number of hours worked per country. But this can be influenced by the proportion of employees working part-time, and it can conceal differences between groups of employees who are overworked and some which are underworked. Both of these two situations can be detrimental for individuals’ well-being, and therefore should be accurately measured within the framework. We propose using the percentage of the population working longer than the ILO threshold of 48 hours as a more accurate indicator than the average number of hours worked. “Underworked” groups are measured through low work intensity and involuntary part time indicators.
The following is a graph presenting the percentages of employed persons working more than 48 hours a week, by country, for 2012. The two bars represent the total employed persons and a subset of them, the ones with full-time contracts. For both categories, there seems to be a slightly decreasing trend, if we look at the evolution of the EU28 average from 2007 until 2012. For the full time employed persons it decreased from 14.8% to 13.2%, and for the total population from 12.2% to 10.6%. This change is less visible in the average number of hours worked (37.3 in 2008, 36.7 in 2012), although the trend seems to be similar.
The Expert Group is invited to give its opinion on the adequacy of replacing the average number of hours worked with the percentage of employed persons working more than 48 hours a week, and for the latter if it is better to use the total employed population or only the ones with full-time contracts.
3.2 Atypical working hours
In the framework at the moment there are 5 indicators measuring the percentage of persons working during atypical hours, namely on Saturdays, Sundays, evenings, nights and shifts separately. The Expert Group endorsed at the previous meeting the proposal of building a synthetic out of these variables, based on a simple logical addition.
In the following graph we present the percentage of employed persons that have any of these atypical working hours (Nights or Evenings or Sundays or Saturdays or shifts), on an usual basis. If a person is included in more than one of these categories (for example works both on Saturdays and Sundays on a regular basis), he or she will not be counted more than once.
At the EU28 level, the indicator seems to be quite stable, being at around 39% since 2006 (with the exception of 2008 when it was 34.6%). As expected, the percentage is slightly higher (2.3 percentual points) for males than for females, and it becomes lower with an increase in educational levels (being at 44% for people with ISCED 0-2 and at 30% for people with ISCED 5-6).
The Expert Group is invited to give its opinion on the adequacy of replacing the 5 indicators from the scoreboard with this new one which synthetizes the information obtained from them.
3.3 Parental leave
Although the importance of parental leave in reconciling work and care responsibilities has been proven to be quite high, comparable European data on the topic is difficult to obtain at this stage. The main reason is related to the differences between national policies, which are reflected in the surveys. In countries where parental leave is really long (up to 3 years) the person is considered inactive, while in others countries she or he may be still counted as employed (because there is a strong attachment of the person to the job, sometimes due to an insurance system through which the person receives an allowance only if employed by that employer, and not as social benefits).At the moment, information about being on parental leave is only collected for the latter category.
A better harmonization of this type of data is proposed to be included in the LFS Revision, but this relates especially to a better definition of which groups of people on parental leave are counted as being employed. A proposal has been made to include a question on whether the person is on parental leave on not to be addressed to the whole population. Still, calculating rates can prove very problematic, what should be the denominator when part of this population will be considered employed and part inactive? Using absolute numbers doesn’t tell us much, given the differences in population in European countries. A solution could be to use the whole population of fertile age and to calculate the incidence of parental leave in this group.
Eurostat’s proposal to the Expert Group is to keep the indicator in the framework, but not disseminate data on it, because there is none sufficiently comparable. It should be labelled under development in the LFS. Eurostat will follow up on the developments that will take place and will inform the Expert Group about them. The Expert Group is asked to express its opinion about the issue.
3.4Child care
The use of child care services is another type of measure that can be used for the reconciliation of work and care responsibilities, and thus can have an important impact on work life balance. Also, research has shown that there is a positive impact on educational achievement and reducing initial inequalities for children who attend pre-primary school. Within the EU-SILC information is collected (and disseminated) on the use of formal childcare arrangements, by age group (under 3, 3-minimum compulsory school age, above this age-12). These indicators, especially the second one, are important both for the Productive or main activity and the Education dimension of Quality of Life. The incidence of using child care services for children under 3 can be more relevant for employment related policies, while for the second age category its usage is more related to educational policies.
Therefore, we suggest including these two indicators in the Frameworkfor Quality of Life, under the heading of the two relevant dimensions. Information exists also on the average number of hours, but given that percentages of non-attendance vary a lot between the countries (in 2011, it varied between 2% and 59% for children aged 3-minimum compulsory school age), this second one seems to be a more relevant indicator.
The following table presents the data for these two age categories, for 2011. Very big differences between the countries are easy to notice.
Less than 3 / 3-compulsory school ageRO / 98% / 59%
PL / 97% / 56%
SK / 96% / 25%
CZ / 95% / 26%
BG / 93% / 41%
LT / 92% / 35%
HU / 92% / 25%
MT / 89% / 27%
AT / 86% / 15%
HR / 85% / 49%
LV / 84% / 27%
EE / 81% / 8%
EL / 81% / 24%
IE / 79% / 18%
CY / 77% / 27%
DE / 76% / 10%
IT / 74% / 5%
FI / 74% / 24%
EU28 / 71% / 16%
PT / 65% / 19%
UK / 65% / 7%
SI / 63% / 8%
BE / 62% / 2%
ES / 61% / 14%
LU / 57% / 27%
FR / 56% / 5%
SE / 49% / 5%
NL / 48% / 11%
DK / 26% / 2%
The Expert Group is asked to give its opinion on the usefulness of including indicators on using formal childcare arrangements in the Framework, and on the optimal placement within the dimensions.
4. Long job tenure
At the previous Expert Group meeting, concerns have been expressed that the percentage of temporary contracts taken alone is not a sufficient indicator to measure job security. There are countries in which permanent jobs are very well protected, and countries in which this seems to be less the case. There are also professions in which temporary contracts are the norm, but the contract is renewed almost automatically, and so the holder is not in an insecure position. Very strong constraints for firing permanent employees may sometimes lead to a dual labour market, because companies may tend to keep a stock of easily dismissible temporary works in order to face changes in demand, while the regular employees are extremely well protected.Therefore, frameworks that try to capture employment quality typically include also indicators related to job tenure, even though this can be affected by factors related to both the employers and the employees. In the OECD Employment Outlook (2013) it is shown that there is a strong correlation between the percentage of workers with long job tenure and the stringency of national dismissal rules (measured by the OECD employment protection index), and thus this is considered to be a good proxy measure for job security.
The following table shows the two indicators comparatively, based on data from the 2012 LFS:
Country / Long job tenure / Temporary contractsAT / 44.5% / 9.3%
BE / 47.8% / 8.1%
BG / 35.4% / 4.4%
CY / 37.5% / 15.1%
CZ / 43.9% / 8.3%
DE / 49.1% / 13.9%
DK / 34.8% / 8.6%
EE / 34.6% / 3.5%
EL / 42.6% / 10.0%
ES / 43.3% / 23.7%
FI / 50.4% / 15.5%
FR / 48.8% / 15.1%
HR / 56.1% / 12.8%
HU / 38.0% / 9.4%
IE / 40.5% / 10.1%
IT / 51.6% / 13.8%
LT / 29.6% / 2.6%
LU / 44.0% / 7.6%
LV / 32.1% / 4.7%
MT / 48.7% / 6.8%
NL / 46.1% / 19.3%
PL / 39.6% / 26.8%
PT / 49.6% / 20.7%
RO / 35.5% / 1.7%
SE / 41.8% / 15.9%
SI / 49.2% / 17.0%
SK / 41.9% / 6.7%
UK / 37.7% / 6.2%
EU28 / 44.7% / 13.7%
It can be noted from the data that two different types of patterns emerge that could be interesting from the job security (and implicitly, Quality of Life) perspective. The countries highlighted in yellow are those in which it seems that the dual labour market may be in place, as both the percentage of temporary contracts and that of long job tenures are higher than the European average. Especially in Portugal and Slovenia this may be the case. The countries highlighted in orange are those where the percentage of temporary contracts doesn’t look like a very good measure of job insecurity, because although it is low, the incidence of long job tenures appears to be low as well. Romania, Bulgaria and the Baltic States are in this category, amongst others. Of course, other specific factors (cultural or structural) may be at play, but the added indicator does bring valuable new information.