Skip to main content

An inquiry into the health and safety at work; a European Union perspective

Final Report Summary - HEALTHATWORK (An inquiry into the health and safety at work; a European Union perspective)

Executive Summary:
Executive summary
The HEALTHatWORK project aims to review current knowledge and issues related to the economic impact of health at work, to assemble, organise, analyse and synthesise data from national projects and surveys in the participant countries, and to recommend future actions for research and policy development aiming at improving health and safety at work in a changing labour market environment in the European Union.
The aims of the project are achieved through, co-ordinated reviews; the development of common datasets regarding indicators of health and safety at work and consistent statistical modelling across EU countries of the various aspects of occupational safety and health. The use of the GIS analysis capability facilitates the identification of patterns of the above indicators across countries and time. The statistical analysis of appropriate databases explores the interactions between indicators of health and safety at work and labour market conditions. A pilot study designed to evaluate the cost and benefit of investing in health and safety at work from the employee and employer perspective is utilised. In co-ordination meetings and workshops the status of health and safety at work is reviewed, its repercussions for the quality of work are examined and a comparative EU-wide assessment of the structure and dynamics of the health and safety at work is carried out. Policy recommendations aiming at improving the health and safety in the context of changing labour market environment and its repercussions in the competitiveness of European labour markets are also proposed.
Main research outline
Importantly, the project provides a detailed review of the current state of knowledge on occupational health and safety (OSH), the economic and social value of its improvement and the OHS contribution to the improvement of the job quality, job satisfaction, and reduction of lost time at work (due to absenteeism and due to occupational illness/accident) in the participant European countries. It identifies gaps in knowledge and the best methods available to collect necessary data. It reviews and document current data collection on the incidence of accidents, illnesses and absenteeism, the management practices for health and safety at work, the experience in meeting economic, social and institutional goals for increasing job quality and job satisfaction in the European Union.
The project statistical analysis and modelling is based on a harmonisation in data management and any data collection for all the participant EU countries, analysis and reporting. It uses Geographical Information Systems (GIS) to generate, maintain and map existing health and safety at work indicators and identifies problematic areas, industries or occupations. In addition an up-to-date review of large scale questionnaire-derived datasets (such as LFS) is offered to uncover if there suitable questions able to produce different indicators of health and safety at work. Two new indicators are proposed; absenteeism for 27 EU countries for the years 2004-2006 and prensenteeism piloted for the UK and Germany. Furthermore how the health and safety at work indicators vary with country, industry and occupation is investigated. The common trends in the above indicators by country, industry and occupation are identified and the effect of indicators of health and safety at work in terms of socio-economic and occupational and demographic factors on the quality of work across the participating EU countries is studied. Importantly, macro-level multivariate analysis to assess the effect of different institutional arrangements or labour market conditions on the health and safety indicators is studied. Finally, the employees' and employers' stated preferences with respect to health and safety provision at workplace with respect to their perceptions of health and safety at work and their willingness to pay for OHS. The scenarios are established and the content of the questionnaires is designed suitable for collecting the relevant data. The survey uses data on employees stated preferences with respect to health and safety at workplace. By using this dataset from a purpose build questionnaire the preferences of the employees and employees regarding OSH are analysed. The synthesis of the results provides information on the willingness of the employers and the employees to pay for facets of OSH in industries with high rates of occupational hazards. Finally, comprehensive Strengths, Weaknesses, Opportunities and Threats (SWOT) analysis is provided that offer a comprehensive synthesis of the available qualitative and quantitative evidence on the state of OHS within each Member State participating in the project.

Project Context and Objectives:
A summary description of project context and objectives
There are considerably large economic costs associated with the lack of provision of health and safety in the workplace (about 3% of lost life years due to the factor 'work', work-related diseases and accidents account for financial losses as high as 4% of world-wide GDP). For each worker in the EU-15, an average of 1.3 working days is lost each year due to an accident at work and 2.1 days are lost because of other work-related health problems. Evidence from psychology suggests that individuals consistently underestimate the probability of an accident/illness at work. The social costs of occupational injuries and illnesses (e.g. the cost to families/communities/national social insurance systems that have to bear the burden of disabled and idle workers, and the early retirees etc.) are largely overlooked in the relevant discourse. Indicators that capture the efficiency of OSH management by firms are inadequately followed at present. Information regarding aspects of OSH management is not readily available in European or international labour market datasets.
This project aims to review current knowledge and issues related to the economic impact of health at work to assemble, organise, analyse and synthesise data from national projects and surveys, and to recommend future actions for research and policy development aiming at improving health and safety at work in a changing labour market environment in the European Union in an era of ageing populations, feminised labour markets and increased incidence of Small and Medium Enterprises. This will be achieved through
(a) Co-ordinated reviews;
(b) Development of common datasets regarding indicators of health and safety at work in the participant countries (the incidence of accidents and illnesses of work as well as the incidence of absenteeism and early retirement due to illness/ accident), and associated GIS analysis capability, leading to:
(c) A series of co-ordination meetings and workshops at which the status of health and safety at work will be reviewed, its repercussions for the quality of work will be examined and a comparative EU-wide assessment of the structure and dynamics of the health and safety at work will be carried out
(d) A pilot study to evaluate the cost and benefit of investing in health and safety at work
(e) Policy recommendations aiming at improving the health and safety in the context of changing labour market environment and its repercussions in the competitiveness of European labour markets.


1. To review the current state of knowledge on health and safety at work in the participant European countries, and the methods available to collect necessary data, on a indicator by indicator basis, and to identify gaps in knowledge;
2. To review and document current data collection regarding the incidence of accidents, illnesses and absenteeism at work and early retirement due to occupational illness/accident and management practices for health and safety in the European Union;
3. To review the current available datasets (such as the national Labour Force Surveys (LFS)) regarding their ability to offer new indicators on the incidence of accidents, illnesses and absenteeism at work and early retirement due to occupational illness/accident in the participant countries and if so to explore ways and construct such indicators;
4. To review the current knowledge on the economic and social value of improvement of occupational health and safety and its contribution to the improvement of the quality of the job, job satisfaction, and reduction of lost time at work (including lost time from work due to absenteeism and early retirement due to occupational illness/accident);
5. To review current knowledge of aspects of occupational health and safety at work related to gender and age issues in light of the feminisation of the labour markets and ageing populations and the increasing incidence of small -medium enterprises (SMEs);
6. To review management options for current state of affairs regarding health and safety at work, to meet economic, social and institutional goals for increasing job quality and job satisfaction, considering issues such as feminisation of the labour market, ageing populations and increasing incidence of self employment, and SME;
7. Review evidence of how improvement of health at work appear to affect the quality of work hence the competitiveness of the Europe;
8. To hold workshops on assessment and management, with invited experts, using existing data to make assessments of health and safety at work, including explicit evaluation of associated state of affairs; to provide baseline management proposals and recommend future research directions;
9. To explore the possibility of developing new health and safety at work indicators (incidence of accidents, illnesses and absenteeism at work and early retirement due to occupational illness/accident) by using large scale available surveys (such as Labour Force Surveys (LFS).
10. Invite research collaborations aiming in securing datasets (such as the LFS) for EU countries not covered by the current partners;
11. To use GIS to map the incidence of accidents, illnesses and absenteeism and other health and safety at work indicators and data in the participant European countries;
12. To analyse the databases and possibly GIS to explore the possibility of developing models of interactions between investments on health and safety at work, job satisfaction (as a measure of job quality) the incidence of accidents, illnesses and absenteeism and early retirement due to occupational illness/accident.
13. To design appropriate a data collection protocol for the incidence of accidents, illnesses and absenteeism at work ant early retirement and assess their appropriateness;
14. Piloting a novel approach to eliciting values of both employers and employees regarding the health and safety at work in chosen medium enterprises in a number of participant countries (at least two or three). This pilot study will provide information on the cost and benefits of investing in improving the health and safety at work.
15. To disseminate the information synthesised and results generated by the Support action as peer-reviewed papers and conference presentations, in addition to reports to the European Commission.
16. To develop new research proposals to address perceived gaps in the information required for improving occupational health and safety indicators in the light of changing labour market environment in an era of ageing populations, feminisation of the labour markets and increased incidence of SMEs
17. Propose appropriate protocols for collecting additional data (including proposing questions for the large scale surveys such as the LFS) that can contribute to the improvement of the accuracy of health and safety at work indicators

Project Results:
PLEASE ALSO SEE THE ATTACHED PDF FILES - A description of the main ST results_foregrounds_2012.pdf AND References_S_T_results foregoundsx.pdf

This Final Report provides a review of the research carried out by the HEALTHatWORK team on issues related to theory and empirical evidence of the market for occupational safety and health (OSH). The increasing competition prompted by globalisation, the predominance of service-oriented industries, the rising job insecurity associated with labour market flexibility and demographic developments in the composition of the workforce (e.g. ageing, feminization), pose important challenges for the health and safety of workers in modern economies. In addition, many governments have recently paid greater attention to the need to tackle the non-trivial costs to both individual and societal welfare that the lack of OSH entails, as part of their overall strive to overhaul insolvent social security regimes.
(i) Research on reviewing the state of knowledge
K Pouliakas and I Theodossiou (2011) survey the state of affairs in the OSH literature and engage in an interdisciplinary survey of the current state of knowledge related to the theory, determinants and consequences of occupational safety and health (OSH). It first describes the fundamental theoretical construct of compensating wage differentials, which is used by economists to understand the optimal provision of OSH in a perfectly competitive labour market. The plethora of incentives faced by workers and firms in job and insurance markets that determine the ultimate level of OSH are discussed in detail. The extensive empirical evidence from the hedonic wage and stated choice approaches used to assess the value of OSH is reviewed. The causes of inefficiency and inequity in the market for OSH, such as externalities, moral hazard in compensation insurance, systematic biases in individual risk perception/well-being and labour market segregation are subsequently examined. The implications of government intervention and regulation for tackling the aforementioned inefficiencies in OSH are then considered. Finally, the survey identifies areas of future research interests and suggests indicators and priorities for policy initiatives that can improve the health and safety of workers in modern job markets.
Barnay, Sauze and Sultan-Taïeb, (2010) investigated the links between health and work - a major policy issue, in terms of retirement, public health, and employment. The review focuses on two different types of studies. Economic analyses have focused on measuring (1) the impact of health on employment and (2) the effect of working conditions on health. Poor health clearly appears to impede professional activity and may cause earlier retirement. Studies have to deal with two difficulties, the measure of individuals' health status and the implementation of longitudinal data collection. The associations between working conditions and health have been established by epidemiological studies for a large number of diseases. Economic studies also provide econometric estimations of the relationship between working conditions, health and absenteeism (Rubin model). In addition to public health measures designed to improve the overall health of the population, these studies raise questions about the specific role of work environment in preserving employees' health. Moreover, in a context of under-reporting and under-recognition of occupational illnesses and injuries, estimations of the costs imputable to occupational exposures require specific economic evaluations. Overcoming these obstacles remains a challenge for economic analyses.
Kopnina and Haafkens (2009) discussed the differences in implementation of policies in relation to chronically ill employees in the context of organizational culture. This is a literature study using systematic hand-search strategy involving medical, statistical, management and social science databases (Web of Science, MedLine, Pub Med, Psych Info, etc.). For the purpose of this study include physical and not psychological or mental disorders. It appears that an appropriate organizational culture is required to be able to take measures to retain chronically ill employees. Various stakeholders view organizational culture as a magic bullet to help introduce company policy to retain chronically ill workers. Within functionalist approach to organizational culture, the three perspectives (integration, fragmentation and differentiation) can be distinguished. According to these perspectives, organizations are classified in accordance to the decision-making and hierarchical structures. The three organizational culture perspectives can be used for understanding what may constitute 'best practice' or 'best strategy' in to address the question of what the organizations can do to facilitate sustained employability for chronically ill workers. The main objective was to determine what type of organizational culture is more effective for policies and practices in case of optimal functioning of chronically ill employees.
Serrier, Sultan-Taïeb, Sauze, Béjean (2009) consider that studies have estimated the scope of under-reporting and under-recognition of occupational diseases and injuries in France. Non availability of data is a major obstacle to the evaluation of the costs imputable to occupational exposures. Moreover, it is often difficult to measure the effect of occupational and safety health interventions. Available studies have difficulties in measuring the fractions of diseases that could be avoided by prevention interventions. The occupational and safety health field has very different characteristics from the clinical health field, as regards randomized controlled trial, blind clinical testing. Job insecurity, workers turnover and temporary work contracts impede the follow-up of workers' health and working conditions. Therefore, a broader use of cost-benefit methods in the occupational and safety health field would require a wider access to data and a more systematic measure of interventions costs and effects.
Drakopoulos, Economou, Grimani, (2012) present a detailed review of current state of legislation framework and empirical research regarding OSH issues in Greece and point out the knowledge gaps & methodological shortcomings. The Greek legislative framework regarding OSH issues has been greatly updated and advanced in the past two decades following the relevant EU legislation. Several shortcomings in the Greek labor market hamper the efforts to achieve efficient OSH measures. Such shortcomings are the lack of trained personnel (physicians etc.) at workplaces; the lack of occupational health inspectors to monitor the enforcement of the law; and the lack of education and information among employees and employers regarding occupational health. Greece exhibits one of the lowest spending on social public policies compared to the EU average and the national social policy is often criticized for the lack of preventive measures and treatment of occupational accidents and diseases once they occur (Christopoulou and Makropoulos, 2007). Only a few studies utilised large samples of workers for a larger time span, to examine the frequency and the determinants of working accidents while the remaining studies provide findings based on small-scale personal interview surveys (Alamanos et al., 1986; Alexe et al., 2003). According to the IKA reports (1999-2006), men sustain the greatest number of fatal and no fatal accidents at work. Most accidents occur in the construction sector and involve upper and lower extremities. During the 1997-2006 decade, there was an annual drop in the number of accidents, at a 4.8% mean rate of reduction. In addition, most prevalent causes of fatal accidents are collision with immobile objects & falling from a height. Most fatal accidents occur in the construction sector and involve cranio-cerebral injury. There is a substantial lack of data on occupational diseases. For the period 2003-2007, when the recording of occupational diseases begun to be implemented in Greece, 103 cases were found. The basic diagnoses are allergic contact dermatitis, toxic effect of metals and asthma. Chemical and industrial agents are described for most cases as casual agents of exposure to disease. According to a cross-sectional study in 2007 of ESYE at recording mental health problems due to work, males are found to be slightly more sensitive to the negative working conditions than females. The main indicators related to OSH that can be drawn from the Social Insurance Institute, Body of Work Inspection, National Statistical Service of Greece are occupational and fatal occupational accidents, occupational diseases, mental health and absenteeism.
(ii) Research on Sickness Absence
Can one explain the individual behaviour regarding absenteeism from work by utilising the concept of a 'market for absenteeism'? Barmby (2012) argues that it is certainly what workers' will demand for various reasons, sickness being one such reason. It is also something which though various, either firm specific arrangements or governmental regulations, firms and governments will facilitate (or supply) to workers. This research is motivated by a recent book on the issue (Treble and Barmby (2011)). The problem is identified by E. J. Working (1927) who gives a framework within which to think about observed market data. How are the various cases depicted below informative?

- Case A Little can be said, we observe one equilibrium point, which isn't going to tell us much about either the demand or supply schedule
- Case B and C are the most promising, here we have either shifts in the supply curve (C) which allows us to see the demand curve or shifts in the demand curve (B) which allows us to see the supply curve
- Case D Here we have shifts in both supply and demand curves at the same time so the data we observe doesn't tell us much about either.
Is it more useful to know something about the supply curve or the demand curve? The supply curve is essentially decided by either firms or governments. Demand reflects worker's preferences, and will give some basis for predicting the effect of policy. In Barmby, Orme and Treble (1991 the effects of a particular experience-rated sick pay scheme was studied. These schemes may be problematic since absence behaviour itself determines the "price" one end up paying, (or at which absence is supplied to him/her). So high absence in one year will make the cost of absence higher in subsequent years in the diagram below you move from SA to SC for instance. One idea which might be worth exploring is the extent which health data might help. The idea is if data on, say, flu, incidence, in previous years (remember the way experience rated schemes work), could be found, then one may have an instrument for "price" of absence which is uncorrelated with present absence behaviour. However detailed localised data on flu incidence is difficult to be found. So the research uses Pneunomia which is measured more consistently across geographical areas in the UK and which will have a correlation with flu. Pneunomia can develop (mainly in old people) from flu. Using data from two factories (in England and in Scotland), variation in death from pneunomia per 100,000 of population is an instrument for "price" of absence. Higher pneumonia deaths will be correlated with higher incidence of flu, and higher incidence of flu will cause higher cost of absence in the following year. The results are suggestive that this may be a useful empirical strategy. One problem is that the health data cannot be measured at a sufficiently defined geographical locality. The factories are in 2 large cities in Scotland and England but the health data refer to either the whole of England and Wales or Scotland. The Table below shows the regression for full time male employees from three factories (N=715) in 1988 in UK. The dependent variable is the mean absence over the year. The empirical results support the accompanying figure. Estimating not taking account of the fact the data is generated in a market context can underestimate the effect of cost.

Cost not instrumented Cost instrumented with Health data
Coeff t statistic Coeff T statistic
Constant 0.0309 7.83 0.0309 7.27
Grade B 0.0104 1.46 0.0137 1.96
Grade C 0.0186 0.88 0.0527 2.52
Cost -0.0009 0.95 -0.0027 2.29
Table 1: Source: Treble and Barmby (2011)
So high absence in one year will make the cost of absence higher in subsequent years in the diagram below you move from SA to SC for instance. One idea which might be worth exploring is the extent which health data might help. The idea is if data on, say, flu, incidence, in previous years (remember the way experience rated schemes work), could be found, then one may have an instrument for "price" of absence which is uncorrelated with present absence behaviour. However detailed localised data on flu incidence is difficult to be found. So the research uses Pneunomia which is measured more consistently across geographical areas in the UK and which will have a correlation with flu. Pneunomia can develop (mainly in old people ) from flu. Using data from two factories (in England and in Scotland), variation in death from pneunomia per 100,000 of population is an instrument for "price" of absence. Higher pneumonia deaths will be correlated with higher incidence of flu, and higher incidence of flu will cause higher cost of absence in the following year. The results are suggestive that this may be a useful empirical strategy. One problem is that the health data cannot be measured at a sufficiently defined geographical locality. The factories are in 2 large cities in Scotland and England but the health data refer to either the whole of England and Wales or Scotland. The Table below shows the regression for full time male employees from three factories (N=715) in 1988 in UK. The dependent variable is the mean absence over the year. The empirical results support the accompanying figure. Estimating not taking account of the fact the data is generated in a market context can underestimate the effect of cost.
Ercolani (2012) considers the proportion of UK work hours lost due to sickness absence calculated for all employees across time and across various economic and demographic characteristics, using the UK Labour Force Surveys (LFS). The total sample is over four million cases spanning 1984 to 2008. Quantifying sickness absence is of interest to practitioners and researchers because it represents lost work hours for employees and employers. It highlights issues relating to the health of the workforce. The techniques presented show how these rates of sickness absence can be calculated for most countries that administer a LFS. The paper also shows now confidence bands around these average rates can be computed. It also shows the special statistical techniques that are needed to calculate the confidence bands for these absence rates because we are calculating the mean value of ratios. The absence rate is defined as the ratio of hours reported absent due to sickness (excluding overtime) to work hours (excluding overtime). More precisely, for all workers i in category j the estimated ratio of sickness absence to work hours for the population is defined by equation (1), the proportion of work hours lost due to sickness absence in the employee workforce:
where j can index any category, or categories, of interest; such as year and/or gender. is the number of hours absent due to illness for worker i during the reference week. is the number of work hours for worker i during the same reference week. i indexes the individual respondent. Although the absent hours and work hours ( ) are based on the survey week, the proportion of absent hours for any given time interval, such as month, quarter or year, can still be generated and no bias is induced by the aggregation. Equation (1) can obviously be reformulated as a ratio of means if we divide both the numerator and the denominator by the number of observations n. Based on either Cochran (1977, section 2.11) or Kendall et al. (1994, section 10.6) equation (1) can be used as a basis for calculating the standard error for the ratio of two means:
s.e. (2)
where by the central limit theorem the variances and covariance of the means are given by , and . calculations of the rate of sickness absence across each year from 1984 to 2008 suggest that approximately 3% of usual working hours are lost to sickness absence each year. Though this annual rate is relatively constant, the same calculations by quarter and month show that there is a very high degree of seasonal variation with peaks at 4.5% and troughs at 2.8%. The seasonal variation in the absence rate is illustrated in Figure 1 below. Calculations for the annual rates of sickness absence by gender show that females have a systematically higher absence rate than males and annual rates of sickness absence by public/private sector splits show that public sector employees have systematically higher absence rates. These patterns may be driven by additional factors such as differing levels of responsibility in household production and the prevailing personal labour market circumstances. Calculations of the rate of sickness absence across economic and demographic characteristics highlight additional features of interest. Rates of sickness absence seem to increase with age but are remarkably low for employees who work beyond retirement age. This is illustrated in Figure 2 .

Figure 1 UK sickness absence rate R by year and quarter Figure 2 UK sickness absence rate R by gender and age
The rates of sickness absence with respect to children being present in the household seem to be very different between females and males and suggest that young children have a substantial detrimental effect on females' sickness absence rates. Rates of sickness absence seem to rise and fall with the number of usual working hours with the peak being at around 30 usual working hours per week. Rates of sickness absence also vary dramatically with the industry in which employees are working as illustrated in Figure 3.

Figure3: UK sickness absence rate R by industry
Industries dominated by the public sector tend to have higher absence rates and industries in the services sectors tend to have lower absence rates. For manual industries the picture is much more mixed. In some industries the number of observations per industry is relatively small (SIC A and B) and it is therefore difficult to give much weight to these results. In other manual industries (SIC C, D, E and F) the absence rates tend to be in the middle of the range. Though these are industries where injury and illness are more likely, those who are long-term ill or injured are unlikely to continue working in these industries and they are therefore not in the sample because no longer working there.
Barmby and Ercolani (2012) show that there exists a North-South divide in the UK regarding the rates of sickness absence. This is the first study to examine these differences using a systematic statistical analysis. Differences in employees' sickness absence rates are analysed as a function of a rich set of economic, social and demographic characteristics. A Blinder (1973) and Oaxaca (1973) decomposition is used. Differences in characteristics are captured by the mean values of the characteristics; differences in behaviour are captured by differences in parameters of estimated models for the two groups. The North-South divide in region-of-work is used to define the two groups of employees and their rates of sickness absence are analysed. The Table 2, based on UK LFS data 2006-2008, shows that the difference in absence rates between the north and south of the UK is 0.374 percent. This 0.374 percent (a 13.37 percent difference) has two components. The first component is the explained one and is useful for policy-makers because it identifies how differences in employees' characteristics affect the outcomes. The second component is the unexplained one, sometimes referred to as discrimination. It is useful for policy-makers because it identifies how differences in outcomes are due to differences in the estimated parameters which reflect differences in behaviour, either by the individuals themselves or by others' attitudes toward those individuals.
Mean Percent
Absence Observations
North: 2.975 75,269
South: 2.602 55,468
Raw difference (R) 0.374
Note: North: Tyne and Wear, Rest of North East, Greater Manchester, Merseyside, Rest of North West, South Yorkshire,West Yorkshire, Rest of Yorkshire and Humberside, East Midlands, West Midlands and Metropolitan Area, Rest of West Midlands, Wales, Strathclyde, Rest of Scotland, Northern Ireland. South: East of England, Inner London, Outer London, South East, South West.
Table2: Difference in Mean Percentage Sickness Absence.
Two-fold and three-fold Blinder-Oaxaca decompositions are carried out and the summary findings are presented in the Table 3. The raw difference R=0.374*** is statistically significant. The explained/endowment component E=0.315*** is statistically significant and accounts for the majority of the raw difference. The other components (U,C,I) are not significant statistically and are not significant in magnitude. Detailed analysis of various socio-economic sub-components of these decompositions characteristics reveal that the significant explained/endowment components are driven by being female, holding a second job, working in the public sector and, predominantly by the hourly wage rate. The findings suggest that differences in absence rates across regions are not caused by differences in estimated parameters. The differences in sickness absence rates are driven by differences in characteristics.
Decomposition Twofold Threefold
North: Mean PercentAbs 2.975*** 2.975***
(49.3) (49.3)
South: Mean PercentAbs 2.602*** 2.602***
(40.3) (40.3)
R: raw difference 0.374*** 0.374***
(4.23) (4.23)
E: explained 0.315***
U: unexplained 0.0581
E: endowments 0.315***
C: coefficients 0.0284
I: interaction 0.0297
Observations 123,105 123,105
Table 3: Blinder-Oaxaca Decompositions of Percentage Sickness Absence
The characteristics in absence equations are either characteristics of employees (demographics) or of the jobs they occupy (wage incentives). It appears that demographics are not the main source of the absence differences. Half of the raw difference in sickness absence is due to regional differences in the wage.
Beblo (2012) investigates sickness absences of men and women from a longitudinal perspective. The paper derives and test hypotheses on determinants of the private and the job absences of male and female employees using the German Socio-Economic Panel (GSOEP), 1985 and 2001. The results of ordered probit estimations confirm that women's and men's sickness absences are related to both working and household conditions. Higher female absences due to family reasons do not seem to represent actual behaviour of German employees. Differences between genders are identified with respect to the relative importance of working conditions for absences and the relative importance of household structure versus amount of time spent in household production. Lack of autonomy has a larger impact on male absenteeism, whereas the work-related relationship with colleagues and supervisors is more important for female employees (Figure 4, where the marginal effects of autonomy on the probability to be absent from work for 1-7 days, 8-14 days and so on is always larger for men and the marginal effects of a supportive environment are larger for women). This finding is robust even conditional on occupation, although the autonomy-absence pattern is evidently related to the occupational segregation by gender. The results support the difference-in-vulnerability argument that women and men react differently to strains and benefits even when they hold the same job (measured by position and occupation) due to their different resources and coping strategies.

Figure 4: Marginal effects of working conditions Figure 5: Marginal effects of the household context
Note: GSOEP, 1985, 1987, 1995 and 2001; own calculations. Pillars represent the marginal effects of the pooled regression and bullets indicate the marginal effects of the separate regressions for 1985 (N = 2,140/1,282 men/women), 1995 (2,773/2,133) and 2001 (3,890/3,159). Control variables include age, education, job position, working hours, tenure (and year dummies).
As for the role of the household, the amount of actual household work and child care is positively related only in the case of male absences, while female employees in general stay home more often when living with a child under the age of 16 (see Figure 5). Women living with children are less absent when they have a partner in the household. They also indicate the potential empirical relevance of the 'double burden' for both women and men. This finding is particularly interesting in light of the potential selectivity of the (female) sample, as those not included in the analysis are likely to experience an even larger 'double burden' if working.
Schreiber, (2011) propose a time-series based approach to test the existence of presenteeism as a phenomenon induced by high unemployment. Presenteeism means that workers show up at their workplace even though they are sick. The study uses official diagnosis-specific sick leave data for Germany. In order to measure any discrepancies between claimed sick leave and the true sickness incidence over time, the volume of internet searches for certain symptoms is used as a proxy for the actual sickness incidence in Germany. To this end the publicly available Google search data from "Google Insights" is used. The paper tests for the influence of labor market conditions on the propensity not to claim sick leave. The tested hypothesis was whether the level of the (non-stationary) unemployment rate has a longer-run influence on the discrepancy between claimed sick leave and actual sickness incidence, using standard tools from time-series econometrics such as cointegration analysis. Many of the relevant variables appear to be non-stationary in the short sample used. The cointegration tests find some evidence for longer-run relationships between the unemployment rate and the sickleave/sickness discrepancy for the sicknesses related to (a) the metabolism, and (b) the digestive apparatus. The sign of the relationship would contradict the unemployment-induced presenteeism hypothesis. For the remaining three sickness groups defined there is no evidence in favour of a relationship between unemployment and the sick leave/sickness discrepancies. The findings are not compatible with the presenteeism hypothesis. This could partly be explained by the very short samples available from Google Insights.

Figure 6: Missed workdays in comparison
Another reason could be the historical sick leave trends in Germany that have been different from those of some other countries (see Figure 6 for a comparison with France and UK). The finding of an existing relationship carries over to the aggregate series which is calculated as the sum of the variables for the five sickness groups that we consider. The unexpected sign is also sustained. In all of these potential long-run relationships, it appears that the unemployment rate is long-run exogenous, i.e. the adjustment to the longer-run relationship is ensured only by the sick leave/sickness discrepancy variables. This causality direction seems plausible.
Drakopoulos and Grimani, (2011) argue that most of the literature on absenteeism suggests that absence from work is a complex issue influenced by multiple causes, both of personal and of organizational nature. There is ample literature focusing on the relationship between job satisfaction and absenteeism (Steers & Rhodes, 1978; Scott & Taylor, 1985; Brooke & Price 1989; Hoque & Islam, 2003; Bockerman & Ilmakunnas, 2008). However, there is a lack of attention to the injury related absenteeism and its relationship with job satisfaction. The data used in this paper was drawn from the EU Commission funded project, SOCIOLD ( By using the Tobit model, the results indicated a statistically significant inverse relationship between the number of days employees stay absent due to occupational injury, and their job satisfaction levels indicating that a specific type of absenteeism is associated with job satisfaction. Although prior research suggested that all of the predictors should relate to absenteeism, only three exhibited significant relationship in this paper. No significant relationship between age and injury absenteeism has been found. There is a significant relation between injury absenteeism and gender (males). Although there is no effect of type of employment and education on absenteeism, permanent workers exhibit less absenteeism rates, while individuals with middle education are more prone to absenteeism. There is no effect of career neither of type of industry on absenteeism. The results of the present paper indicate that a specific type of absenteeism might offer more empirical information to the complex absenteeism job satisfaction relationship.
Blazquez, (2012), highlights the trends of absenteeism among spanish employees using the Spanish LFS micro-data. Absenteeism is an important part of the individual decision on actual working hours. It might be an efficient individual response in the presence of institutional constraints - such as minimum working hours - that affect individuals' choice between work and leisure (Dunn and Youngblood, 1986). However, significant efficiency costs may arise when absence costs are not internalized by workers. This moral hazard problem leads to lower values of output and employment in equilibrium, owing to the imperfect substitutability of absent workers. If insurance costs are mainly borne by the government, as is the case in most European countries, significant fiscal costs will also arise. Employers are aware of the direct and indirect costs of sickness absence. The direct costs include statutory sick pay, cost of replacement staff and loss of output. The indirect costs, difficult to quantify, are also related to sickness absence. These include low morale among staff who have to carry out the extra work for those who are absent, the cost of managing absence and the impact on training and development, all of which impact on the overall levels of output of the organization. Using data from the Spanish LFS (1996-2004) provides evidence for the proportion of contracted work hours lost due to sickness absence, its evolution across time and its main determinants. The estimation results reveal that females are significantly more likely to be absent from work due to sickness than males. Although there are no formalized theories that explain gender differences in sickness absence rates, empirical findings suggest that such differences exist. One way to understand it is to relate the disparity between men and women in sickness absence to the way they cope with adverse working conditions: men and women are on average different in how vulnerable they are to adverse working conditions (Koheler et al., 2006; Väänänen et al., 2003; Mastekaasa, 2000; Messing, 1998). Gender differences can also be explained by a combination of biological, psychological and socio-cultural factors. For instance, if women must conciliate family and work activities they are more likely to have a higher level of 'life stress' and thus to be more vulnerable to work-related stress. The presence of children in the household does not seem to increase the likelihood of absenteeism neither for males nor females. However, the effect of marital status significantly differs between males and females. Taking singles as the reference category, married females present a higher probability of being absent from work, while the opposite is observed among the male subsample. Sickness absence is found to be positively related with age. This relationship between age and sickness absence ? especially long-term absence ? can be partly explained by the positive correlation between age and prevalence of chronic diseases. Furthermore the results reveal that higher educational attainments are associated with lower rates of absenteeism due to sickness. The estimation results reveal that increased job insecurity, captured by temporary contracts, has a negative effect on absenteeism. Public sector employees are more likely to be on sick leave than those in the private sector. There may be several reasons behind these differences in absenteeism. One is that private employers have stronger incentives to prevent absence, since it is costly to the employer, whereas public employers have weaker direct incentives to minimize costs to their organization. Another explanation is based on the possibility of self-selection: workers with preferences for frequent absence self-select the public sector because of its higher degree of employment security. The results also reveal remarkably differences in sickness absence rates between occupational groups. In line with the existing evidence, blue collar workers exhibit a higher sickness absence rate than white collar workers (Taylor, 1979; Steers and Rhodes, 1984). This result can be partly attributable to the greater likelihood of injury at work among manual jobs.
(iii) Research on exposure to job strain and psychological or physical health
Sultan-Taïeb, Sauze, Vieillard, Sultan, Niedhammer, (2011) explore the validity of a measure of job strain, defined by the combination of high psychological demands and low decision latitude, in order to compare job strain exposure. They examine the role of welfare state regime in the differences in this exposure between countries using a random sample of 9,953 male and 11,462 female employees in 27 European countries from the European Working Conditions Survey (EWCS) in 2005. Two measures of job strain (12- and 19-item) were constructed. Multilevel logistic regression analyses were performed controlling for covariates: gender, age, occupation, economic activity, public/private sector, part/full time work, and number of workers in household. Both measures of job strain had satisfactory psychometric properties: internal consistency, correlations, and factorial validity. The lowest prevalence of exposure was observed in Sweden and the highest in Greece. Significant differences in job strain exposure were observed between countries even after controlling for covariates. Working in a Southern, Eastern European or Bismarkian welfare state regime was associated with a higher risk of exposure for men. Welfare state regime contributed to explain the differences in job strain exposure between countries especially for men. This comparative analysis of job strain exposures may assist decision-makers in orienting prevention policies in order to improve working conditions at European level.
Sauze, Sultan-Taïeb, Viellard, (2010) use two waves of the (EWCS) (2000 and 2005) to study the evolution of exposure to psychosocial risks in six countries (Germany, Spain, France, Italy, United Kingdom, Sweden). Differences in exposure between countries tended to increase during the period. Swedish workers had the lowest (or one of the lowest) exposure to psychosocial risks, except for psychological demands (Figure 7). They have experienced the largest decrease in exposure during the period. These differences can be explained by the form of work organization (Lorenz and Valeyre, 2005). The legal context can be also a factor of explanation; the work environment act compels a systematic work environment management including psychosocial risks.

Changes in psychological demands in skills utilisation in decision authority
Figure 7 Changes in psychological demands
Jones, Latreille and Sloane (2011) use matched employee-employer data from the British Workplace Employment Relations Survey (WERS) 2004 to examine the determinants of employee job anxiety and work-related psychological illness. Organisations have been keen to highlight the business case for improving employee psychological health, and highlight sickness absence, employee turnover and presenteeism (being at work, but working at less than full capacity) as mechanisms through which psychological health may influence firm performance (see, for example, Sainsbury Centre for Mental Health, 2007). The methodology underlying this evidence relies on estimating work time 'lost' as a result of psychological ill-health using responses from employees and converting this into an aggregate economic or monetary cost by multiplying it by an estimate of the value of work, often measured using hourly wage rates (Stewart et al., 2003 and Goetzel et al., 2004). These studies, do not distinguish between the costs of psychological ill-health and work-related psychological ill-health, the latter of which may be within an employer's more direct control. This paper extends the analysis of Leontaridi and Ward (2002). It contributes to the literature by first using a large scale nationally representative data of workplaces in both the public and private sector and it provides a comprehensive analysis of the determinants and effects of employee psychological health. Second, by focusing on work-related psychological health, it examines an aspect of psychological health over which employers have more control and to which modifications of work practices may have more influence. Third, the data contain an extensive set of controls for the influence of job characteristics and the workplace on psychological health, including measures of co-worker psychological health. Fourth, the matched nature of the data facilitates the examination of the relationship between psychological health reported by employees and workforce psychological health and workplace performance reported by the manager. Measures of absence, quits and labour productivity are utilised in an attempt to identify the channels through which work-related psychological health may influence workplace performance. The paper provides evidence on personal and employment related characteristics that are correlated with employee job anxiety. Overall, the results are largely consistent with the existing literature which finds that psychological ill-health is more strongly related to factors relating to the job than worker characteristics (see Groot and Maassen van den Brink, 1999 and Michie and Williams, 2003). There is strong support that job anxiety is positively associated with job demands, consistent with Wood (2008). Average levels of job anxiety have a positive influence on managers reporting workforce stress, suggesting that employee reports contain valuable information. A one unit change in the job anxiety index is associated with a 12 percentage point increase in the probability of manager-reported workplace stress. There is also evidence of a positive relationship between job anxiety and absence with a one unit increase in average workplace level of employee job anxiety being associated with a 2.90 percentage point increase in the absence rate. No relationship between job anxiety and the workplace quit rate is found. There is some evidence from the subjective and objective measures of labour productivity to suggest a negative association between work-related psychological ill-health and workplace labour productivity. The paper considers the possible endogeneity of job anxiety in the analysis.
Haafkens, Kopnina, Meerman, van Dijk, (2011) suggest that chronic diseases are a leading contributor to work disability and job loss in Europe. The aim of this qualitative study was to explore and compare the perspectives of Dutch LMs and HRM on what is needed to facilitate job retention for chronically ill employees. To achieve this aim we held two focus-group sessions, using the concept-mapping methodology. During the first part of the session participants were asked to generate statements to complete the following focus prompt: In order to ensure that chronically ill employees can continue to work, it is necessary that ...? During the next part of the session participants were asked to rate the importance of the statements on a five point scale and to sort them according to theme. The results from the two concept mapping meetings were analyzed separately for HRM and LMs. The means of the importance ratings the participants assigned to each statement were calculated at a group level. This resulted in a rated list of statements for HRM and LMs. Using a series of statistical analyses (multi-dimensional scaling techniques and then hierarchical cluster analysis, the participants' statements and sorting results were aggregated at a group level. The brainstorm session yielded 35 statements expressing the thoughts of the participants on what is necessary in order to ensure that chronically ill employees can continue to work. Analysis of the sorting activity revealed that each group sorted the statements into a set of six distinct thematic clusters, referring to conditions they perceived as necessary to ensure that chronically ill employees can continue to work. Figures 8, 9 show the cluster maps.

Figure8: Perspectives of line managers on what Figure 9: Perspectives of human resource managers
is needed to ensure continued employment on what is needed to ensure continued employment
for chronically ill employees: cluster map. for chronically ill employees: cluster map
Line Managers regarded "good cooperation between manager and employee" as the most important condition to ensure continued employability for chronically ill employees (Cluster 1). The statements grouped under this theme indicate that this involves mutual trust, contact, shared responsibilities between manager and employee, and attentiveness from the manager, but also the ability of the employer to make demands on the employee. Cluster 2 indicates that LMs also find it important that "a manager must have basic knowledge of how chronic illness can affect work". A relatively high priority was also assigned to the "role of employees themselves" (Cluster 3). Almost just as important is the theme that "work should be accommodated to the condition and needs of the employee", within the capabilities of the organization (Cluster 4). Cluster 5 indicates that "good information and knowledge transfer between managers, occupational physicians and HRM "is also perceived as a prerequisite for facilitating the employability of chronically ill workers. Cluster 6 concerns the need to develop a "company policy" with respect to chronically ill employees. Although this theme has the lowest average score (3.1) a relatively high score was assigned to the statement "an organization should reflect on what it means to be a good employer for the chronically ill employee." Other statements grouped under this theme refer mostly to organizational policies and practices that need to be in place according to the LMs. HRM assigned the highest importance to "company policy" as a condition to facilitate sustained employability for chronically ill employees (Cluster 1). The statements they grouped under this theme refer to what the employer, the manager , the organization and the employees must do in order to develop such a policy: the employer must realize that the employee cannot continue to work in an unhealthy situation; the work must be suited to the condition of the employee; managers must have the right to make demands on employees and evaluate the consequences of an employee's illness for his or her colleagues; the organization needs to reflect on what good employership involves; and the employees must understand their capabilities and limitations. A second, almost equally important cluster is that there must be "a culture of trust, openness and communication" within the organization. This theme contains statements regarding the relationship between managers and employees. The third cluster indicates that HRM also feel that chronically ill employees and managers must "share responsibilities" in order to ensure continued employability for the employee. From the part of the LMs this means removing the employee's fear of repercussions or shame about his or her condition, while the employee must realize that privacy is not always possible. The fourth cluster contains statements indicating that both "managers and personnel officers" should have sufficient "knowledge about chronic illness and its impact on work" to be able to act proactively. The fifth cluster indicates that "work adaptations" are also seen a condition to help ensure job retention for chronically ill employees, and the statements in this cluster refer to who should be responsible for providing these accommodations. Cluster 6, "support services within the company", was given the lowest average priority. It contains some low-rated statements suggesting possibilities for centralized services for chronically ill employees within the organization. Both similarities and differences were found between the views of LMs and HRM on what may facilitate job retention for chronically ill employees. Mutual trust between the manager and the employee was rated as the most important statement by both groups. Four thematic clusters were mentioned by both groups; two uniquely by LMs; and two uniquely by HRM. LMs saw "good employee/manager cooperation" as the most important starting point for enabling job retention. For HRM the most important starting point was "corporate policy and culture".
Serrier, (2011) evaluates the social cost of respiratory cancer attributable to occupational risk factors in France and reviews the available epidemiological data in the literature. By using the Medline database, a review of the literature restricted to meta-analysis highlights the relative risk data available. The method of Attributable Risks (AR) is mobilized to estimate the numbers of lung, sinonasal and mesothelioma cancer cases caused by asbestos, exhaust fumes from diesel engines, painters, crystalline silica, wood dust and leather dust. The author then assesses the costs of these cancer cases for the French society using the Cost of Illness (COI) method. To take into account all indirect costs a decision tree is developed to estimate the probability of being involved in each cost category. Cancer-specific assessment models for each category of costs are set up to allow the estimation, according to incidence-based and prevalence-based approaches, direct costs (hospital and ambulatory care), indirect costs of morbidity (absenteeism and presenteeism) and mortality in the market and nonmarket spheres. The study estimates that exposure to asbestos in the workplace is responsible for 5 618 lung cancer cases and 3 676 deaths in 2010. The number of lung cancer cases attributable to exposure to exhaust fumes from diesel engines is estimated between 2 367 and 3 283 for 2010 and the number of deaths between 1548 and 2146. Occupational exposure to crystalline silica would be responsible for 1209 to 2 241 lung cancer cases and between 790 and 1 475 deaths. For 2010, the social cost of lung, sinonasal and mesothelioma cancer cases caused by asbestos, exhaust fumes from diesel engines, painters, crystalline silica, dust wood and leather dust in France is estimated between 986 and 1248 million euros according to prevalence-based approach and between 1223 and 1586 million euros according to incidence-based approach among which 760 to 806 million euros only for asbestos.
(iv) Research on working conditions and health at work
Böckerman, Petri, Johansson and Kauhanen, (2011) argue that there is scare literature on the effects of innovative work practices on employees. Contradicting results dominate the empirical literature focusing on the impact of innovative work practices on employee health. One reason for this unsatisfactory situation might be that the existing studies typically investigate cumulative disorders and other specific injuries or illnesses. Accordingly they overlook possible concomitant effects on other illnesses and the general well-being of employees. This paper provides a more comprehensive picture of the effects of innovative work practices on employee health by studying their impact on sickness absence and accidents at work. The analysis is based on the Finnish Quality of Work Life Survey for 2008. The Finnish case has a broader interest for several reasons. While innovative work practices have rapidly gained popularity in Finland, the country is also characterised by a conspicuously high share of sickness absenteeism. The high unionization rate in combination with close co-operation between employees and employers should provide an exceptionally fertile ground for the benefits of innovative work systems to emerge and, hence, for minimizing the potentially negative effects of such practices on employee outcomes. The impact of workplace innovations on employee health in terms of sickness absence and accidents at work is assessed by estimating the joint effects of innovative work practices (self-managed teams, information sharing, employer-provided training and incentive pay) since workplace innovations are known to be complementary in their effects rather than substitutes. However, the effects can be expected to differ significantly between employee groups as well as absence measures (any absence, long-term absence, accidents at work). Thus a distinction is made in the analysis in these respects. The results provide no clear-cut support for workplace innovations causing an increase in long-term sickness absence or in the prevalence of accidents at work. Also the effects of these systems on short-term sickness absence are found to be close to negligible. The overall conclusion of the paper, therefore, is that high-performance workplace systems have little impact on the overall health of employees in Finland.
Keith A. Bender K and I. Theodossiou (2012) point out that an interesting and little explored potential unintended consequence of performance pay is its effect on health. Except for a limited number of studies this observation has not been empirically tested, and the small literature that has examined this topic has focused very narrowly on injuries at work. Previous research shows that performance related pay generates increased effort and productivity at work. However, it may also generate a series of unintended consequences. Adam Smith was the first to observe in The Wealth of Nations, "Workmen when they are liberally paid by the piece, are very apt to overwork themselves and to ruin their health and constitution in a few years". The goal of this paper is to use a nationally representative dataset to broaden the definition of health to examine other health outcomes, particularly ones affected by stress, to see if Smith's observation holds, particularly with respect to the implication that increases in the length of time paid using piece rates will negatively impact 'health and constitution in a few years'. Using multiple waves of the British Household Panel Survey dataset, duration models find that increasing the time in which workers' pay comprises at least in part of performance pay generates higher odds of falling into bad health either measured by subjective health or along one of three physical health dimensions (Table 4).
Cox Prentice-Gloeckler
Sample w/o Frailty w/ Frailty
Overall 1.004*** 1.005*** 1.013***
Overall with Past health controlled 1.003*** 1.004*** 1.009**
Female 1.004** 1.004** 1.011**
Male 1.004*** 1.005*** 1.016**
Income<=median 1.005*** 1.005** 1.021***
Income>median 1.003* 1.004*** 1.011**
Manual occupation 1.007*** 1.008*** 1.028***
Nonmanual occupation 1.003** 1.004*** 1.010**
Notes: z-statistics under odds ratios. All regressions also include a constant and indicators for gender, noncompletion of secondary education (excluded), completion of secondary education, completion of postsecondary education, age less than 26 (excluded), age between 26-35, age between 36-45, age between46-55, age above 55, married, currently smoking, labor income quartile, broad occupation and region where the P-G regressions also include log of time. Numbers under odds ratios are asymptotic z-statistics. *, **, and *** indicate statistical significance at the 10, 5 and 1 percent level. The sample is those people who were working in wage and salary jobs for the entire time we observe them and who start out with excellent or very good subjective health in wave eight. For the overall sample, there are 2,443 observations where 51% are censored.
Table 4. Selected Results for Hazard Ratios for Overall Subjective Health: Odds Ratio on Percentage of Time Spent in Performance Pay
These results are robust to variations in variables or estimation procedure, and serve as a lower bound of the negative effect if the endogeneity controls included are not capturing all endogeneity given the likely positive bias due to healthier workers selecting performance pay jobs. Performance pay can generate a variety of efficient labor market outcomes. However, the findings here are firmly in the camp of a potential unintended consequence of performance pay. Like other research that finds that performance pay can lead to workers, for example, focusing on quantity rather than quantity or overusing physical capital, long term exposure to performance pay is related to worse health, suggesting that firms may face increased health insurance or workers compensation costs.
Cottini, E, (2011) investigates the relationship between health, working conditions and pay in Europe (OECD, 1996, 2008; Lucifora and Salverda, 2009). The decline of manufacturing jobs, the growth of service sector jobs and rapid technological progress put raising pressure for labour market flexibility, which led to a rapid diffusion of temporary contracts, low-paid jobs and part time work. It has been argued by the European Commission that low-paid workers suffer from a double penalty as their jobs are also of intrinsically bad quality. Standard economic theory commands that workers search for jobs that will provide higher utility, and quit any time this opportunity arises. If a worker has preferences for some non-wage attributes, this model predicts that in equilibrium the labour supply facing each establishment should be rising in both wages and non-wage components. Firms, by offering jobs that yield higher utility to the worker can attract and retain workers more easily. Non-wage components, such as job quality and workplace safety, are instruments to achieve a profit maximizing labour supply. While bad working conditions and low pay may harm workers' health, poor health can make it more difficult to search for jobs and more physically or mentally costly to work. Equally, illnesses may also increase absenteeism and reduce job performance, which can affect earnings, increase the probability of dismissal and reduce the chances of promotion. Employers may also discriminate against workers who have a physical or mental disability even when their performance is satisfactory. Knowing how health depends on work environment and employment arrangements is clearly of policy relevance as it provides key equity considerations to complement the efficiency argument advocated by employers (Bardasi and Francesconi, 2004). The extent of these problems seems to vary across countries according to the legal and social protection for worker's health and safety (Cox et al., 2004). In this paper the 2005 wave of the EWCS is used to analyse the link between working conditions and low pay on health at work and provide cross-country evidence for 15 European countries. Results show that, controlling for personal characteristics, firm attributes, industry and occupational structure characteristics (adverse) working conditions are associated with poor health both physical and mental at work. Low pay plays a role, mainly for men and when interacted with working conditions, suggesting that stigma and deprivation effects may be correlated with health at work. There is evidence that the association of health with poor working conditions is attenuated by the low-pay status. Women seem to derive their Health problems at work from bad working conditions, while for men being low paid is strongly associated with bad health (Table 5).
Table 5: Health problems at work
Even accounting for the potential endogeneity arising from workers sorting by firms and job types with different working conditions, the results support the hypothesis of a causal effect of (adverse) working conditions on the probability of experiencing health problems. Working conditions are an important determinant of health status at the workplace, and that health policies directed to workers should pay attention at improving working attributes and pay.
Blasquez, Cottini and Herrarte (2011) use the Spanish Living Conditions Survey (2005-2008) to investigate whether the socioeconomic gradient in health remains when alternative measures of socioeconomic status are taken into account. People in lower socioeconomic status groups have much worse health outcomes than those in higher socioeconomic groups (Fox, 1994; Nazaroo, 1998; Marmot and Wilkinson, 1999; and Smith, 1999). Which dimensions of socioeconomic status actually matter to determine agents' health status? The level of income has been a commonly used indicator. After controlling for significant determinants of health, such as sex, age, educational level or occupation, a striking relationship between self-reported health and income persists. Other forms of deprivation, apart from that of income, might also exert a significant influence on health status. Material deprivation is closely related to social exclusion, a concept that in the last years has received increasing attention among social scientists discussing the attributes, differences and novelties of it with respect to more traditional concepts such as income poverty, multidimensional poverty and inequality. This shift from the concept of poverty to material deprivation and social exclusion reflects the need for a multidimensional approach to study social disadvantage. In this respect, the multidimensional aspect refers to the failure to attain adequate levels of various functionings that are deemed valuable in the society (Sen, 1985). Another fundamental element identifying the concepts of material deprivation and social exclusion is relativity. Relativity comes from the idea that a person's feeling of deprivation in a society arises out of comparing his situation with those who are better off (Runciman, 1966). This paper makes an important contribution to the literature on the socioeconomic gradient in health. It examines the effect of material deprivation in terms of financial situation, basic necessities, durables and housing conditions on individual self-assessed health (SAH). The authors start with the "counting" approach of Atkinson (2003), and construct two deprivation indicators by summing up the number of dimensions in which the person is deprived, assigning respectively equal and different weights to the various dimensions considered. Then they follow Bossert et al., (2007), and define a third deprivation index as the product of two terms. The first term corresponds to the Yitzhaki (1979) index, that is, the average of the functioning-failure differences between a person and those who are better off. The second term, the share of agents with fewer functioning-failures, captures the capacity of an individual to identify with other members of the society. The results suggest a negative relationship between the deprivation indicators, in any of the life domains considered, and self-assessed health. This holds even when we consider a relative measure of deprivation (Bossert et al., approach), thus reflecting a comparison effect, by which unfavourable comparisons with the social peers depress individual levels of SAH. Information on health is derived from the question "How is your health in general". It is a five-point response scale ranging from very bad to very good. The approach followed in this paper is to exploit the panel structure of the data and estimate a random effects model using Mundlak (1978) to allow for correlations between the individual random effects and the observable variables. Finally, as in Contoyannis et al. (2004), the authors carry out separate estimations for males and females. They find that the estimated coefficients of material deprivation are larger in absolute terms than that of income, which is found to be positive and between 0.02 and 0.025 among males, while considerably lower and non significant among females. These results confirm that the socioeconomic gradient in health is more latent when, apart from income, deprivation in a variety of life dimensions is considered to measure individual's socioeconomic status. Except the domain of Financial Difficulties, the larger differences with respect to income appear when the deprivation scores are based on Bossert et al., approach. This suggests the importance of unfavourable comparisons with the social peers in depressing individual levels of SAH. For instance, among males the coefficient representing the health effects of lower societal position in the domain of Financial Difficulties is of -0.122 significantly above the corresponding value in absolute terms to income. The same occurs in the rest of deprivation domains, with estimated coefficients of -0.109 -0.116 and -0.045 for the domains of Basic Necessities, Housing Conditions and Durables, respectively. A similar pattern is observed for the female subsample. However, as it was mentioned before, negative comparison information (higher deprivation scores) in the domains of Basic Necessities (-0.152) and Housing Conditions (-0.154) exert major negative effects of health.
Mertens and Beblo (2011) ask whether the recent financial market crisis has caused a deterioration of satisfaction not only for the unemployed but also for those out of the labour force and especially those in employment. The focus of our analyses is on the pattern of life, job and health satisfaction over time and the influence of unemployment rates, inflation rates and GDP growth. We compare the UK and Germany, two countries with different employment protection regulations and different consequences of the crisis for the labour market. The authors found significant differences between the formerly separated parts of Germany even twenty years after re-unification. While people living in the western part of Germany report somewhat lower satisfaction in 2009 compared to the previous years, those living in the eastern part report higher levels than in 2006, the year before the financial crisis started (except for health satisfaction as illustrated in the Figure 10).

Figure 10: West Germany East Germany
Source: Authors' calculations with SOEP 1996-2009
This could be due to the largely different employment structures in both parts of Germany. As exports and not so much services were hit by the crisis, western German federal states were facing stronger negative demand pressures. Our findings for the UK are similar to the East German evidence.
Looking at the impact of macroeconomic indicators such as GDP growth, inflation and unemployment, we find diverse effects for Germany and the UK. Positive reactions to an increase in GDP are observed in West Germany only. As illustrated in the Table 5 below, the strongest and most robust result across all the subsamples is detected for the relationship between self-reported satisfaction (regarding all aspects, life, job and health) and the regional unemployment rate: The higher regional unemployment the more satisfied people seem to be irrespective their current labour market status. Not only unemployed workers feel better when unemployment rises a potential result of changing social norms in times of rising unemployment but also employed workers who seem to be more satisfied with their lives and jobs when unemployment is rising around them. The overall level of unemployment has a significantly negative effect, though for Germans only. Interestingly, similar results are obtained when looking at job and health satisfaction. In the UK, the significance varies with the strongest effects on job satisfaction.
Life satisfaction Job satisfaction Health satisfaction
Germany UK Germany UK Germany UK
West East West East West East
Crisis years +/- + + - - + - - +
Temporary job in crisis years + + + + + + + + +

GDP growth rate +
Inflation rate -
National unemployment rate - - + - - -
Regional unemployment rate + + + + +
# obs 69,486 22,824 62,229 68,669 22,534 62,268 68,669 22,534 62,268
Table 8 Summary of the economic determinants of satisfaction measures
Note: Summary of the qualitative results of the paper. + indicates statistically significant positive relationship, - indicates statistically significant negative relationship.
Blazquez, Cottini and Herrarte, (2012) use the Spanish Living Conditions Survey (2005-2008) to investigate whether there is a socioeconomic gradient in self assessed health (SAH) when alternative measures of socioeconomic status, apart from income, are considered. Material deprivation in terms of financial difficulties, basic necessities and housing conditions exerts a direct effect on individual health. Deprivation influences individual health through two channels. First, lack of monetary resources and/or inaccessibility to specific items has a direct negative impact on SAH. The second relates to comparison effects with the societal peers. Following Bossert et al. (2007), relative deprivation as the product of two terms: The first, is the average of the functioning-failure differences between a person and those who are better off (Yitzhaki (1979) index). The second is the share of agents with fewer functioning-failures, captures the capacity of an individual to identify with other members of the society. A multidimensional index of material deprivation is used. Since the relative deprivation effect is driven by unobserved factors correlated both with low income and poor health, the panel structure of the data is exploited to include individual effects by proposing a RE model extended to include a Mundlak term that allows for correlations between the errors and the observable variables. Since health dynamics may be influenced by gender (Contoyannis et al. (2004)), the authors carry out separate estimations for males and females and provide a series of sensitivity tests that support our results. In line with previous works (Graham 2007; House et al., 2005), the results provide evidence that socioeconomic gradient in health is not unidimensional. They find interesting differences regarding the influence of income and material deprivation on SAH. Higher distances between own income and others' income significantly depresses individual health, which supports the relative income hypothesis. Once comparisons in terms of income are considered as determinants of health, own income does not play any role. Additionally, material deprivation leads to a substantial decrease in SAH. However, unlike income, the SAH effects of being deprived in these domains mainly operate through a direct channel, with inaccessibility to specific items exerting per se a negative impact on health.
Cottini, and Lucifora, (2010) investigate the links between contractual arrangements, working conditions and mental health using time-series cross-section data for 15 European countries. In recent decades industrialised countries have experienced substantial changes in the functioning of labour markets and increased pressure for higher labour flexibility pursued by reforming labour market regulation and working arrangements (reducing employment protection legislation, introducing non standard work arrangements and, at the firm level, increasing demand over workers performance, demanding job tasks and lower worker's control). In most European countries, working conditions have progressively deteriorated (OECD, 2008). These changes are expected to impact on worker's health conditions and their overall well-being. The Mental Health Foundation, 2000 has recognised the prevalence and impact of mental health disorders in the working population. Employee performance, rates of illness, absenteeism, accidents and staff turnover are all strongly associated to employee's mental health status. Mental illnesses contribute to the burden of disease and disability benefits (Marusic, 2004; Heijink et al.2006; Mark et al. 2007). This paper contributes to the existing literature by documenting the links between working conditions and mental health for a large set of countries using comparable data (Figure below). All indicators of working conditions show a positive relationship with mental health problems such that countries with the worst combinations of psychosocial conditions and physical hazards also score badly in terms of worker's mental health.

Figure 11: Correlation between mental conditions and health
EWCS 1995, 2000 and 2005 is used to investigate recent patterns in mental health at the workplace across 15 European countries, and to assess how working conditions - such as shifts, repetitive work, job autonomy, job intensity and job complexity and physical hazards - are related to worker' s mental health conditions. The paper also investigates the potential sources of these differences accounting for personal characteristics, firm attributes, industries, occupational structure and the institutional context (as shown in the Table 11). As workers may sort across jobs according to their preferences and risk aversion, the potential endogeneity of working conditions and mental health within the workplace is assessed. Overall the results support the perception, currently debated word-wide, that adverse working conditions can affect worker' s mental health conditions.
Cottini E. and P. Ghinetti (2011)aim to investigate whether employee health is affected by the environment in which the individual works in terms of both physical and psychosocial working conditions and by his or her lifestyle. The period of rapid transformation and changes in the organisation of the production system has modified the work environment, with an increase in the share of atypical jobs and a reduction of hierarchical levels, as well as a growth of service oriented work. The traditional sources of adverse physical working conditions are declining, whereas the share of workers subject to psychological job stressors is increasing (Cappelli et al., 1997). A greater importance of "immaterial" job attributes such as stress and work overload relative to strenuous physical working conditions may have non neutral effects on health at work, with a worsening in its mental versus its physical component (OECD, 2008; Cottini and Lucifora, 2010; Cottini 2011). From a policy perspective, the EU commission has recognised the importance of several job quality dimensions and decent working conditions for the implementation of the European Employment Strategy (EU 2001) Whilst the relationship between lifestyle indicators and self-assessed general health where the former plays an input role in the production of the latter - has been recently investigated (e.g. Contoyannis and Jones, 2004), the role that working conditions could play in the same context has not received the same attention yet. Adverse environmental job aspects and, more in general, organisational factors are important determinants of perceived health. Once the mental and physical health are considered separately, substantially different gradient relative to variables capturing differences in lifestyles and working conditions are observed. Overall, the causal effect of lifestyles and working conditions it is in general negative, but not as strong as one may expect, and it is concentrated on the subjective evaluation of health (Self-assessed health). When the more objective measures of mental work-related and physical health are considered many effects vanishes (with the exception of drinking and job insecurity) and, if any, they are concentrated on mental health.
Cottini, Kato and Westergard-Nielsen, (2011) using a Danish linked employer employee data find that: (i) exposing the worker to physical hazards leads to a 3 percentage point increase in the probability of voluntary turnover from the average rate of 18%; (ii) working in night shift results in an 11-percentage point hike; and (iii) having an unsupportive boss leads to a 6-percentage point jump. High involvement work practices are found to play a significant role in mitigating the adverse effects of workplace hazards. Finally, the worker under adverse workplace conditions is found to improve the 5-year odds of rectifying such workplace adversities by quitting the firm. The study provides new evidence on the exit behavior of workers who are exposed to workplace hazards and the potentially important role that high involvement work system (a cluster of complementary human resource management practices designed to promote employee involvement) can play in mitigating such exit behaviour. Thus the paper integrates two very different streams of research: that of workplace hazards and that of high involvement work systems, and provides researchers, practitioners and policy makers with fresh insight on interplay between workplace hazards and high involvement work systems. The authors link the 2000 data to the 2005 data and study whether the worker exposed to adverse workplace conditions can improve her odds of rectifying such workplace adversities in 5 years significantly by separating from the firm voluntarily. By separating voluntarily from her firm, the worker exposed to physical hazards at the beginning of year 2000 can improve her 5-year odds of eliminating such physical hazard exposure by 17 percentage points. Since the average odds of escaping from physical hazards in 5 years is 40%, the estimated marginal effect of voluntary turnover amounts to almost a 50-percent increase. For all workers with bad boss, voluntary turnover will result in a 10-percentage-point increase in their odds of rectifying this adverse workplace condition in 5 years (which amounts to a 14-percent rise from the average 5-year odds of 70%). The worker without strong voice in the workplace can improve her odds of gaining such voice in 5 years by over 20 percentage points (which amounts to a 30-percent hike in the 5-year odds from the average odds of 67%) if she quits the firm. The data support that changing the firm voluntarily does boost the odds of improving workplace conditions in 5 years, even if such turnover does not accompany any change in occupation, industry, firm size, education, location, and wage (see Table 6)

Table 6 Five years odds of workplace improvements with and without turnover
Cottini (2011) investigates how different dimensions of working conditions affect the health of female and male workers in 15 European countries. The impact of health problems at the workplace has serious consequences for the productivity of the firm. Poor health might results in higher absenteeism rates, lower productivity and performance. An adverse work environment may damage workers' health and make more difficult to search for jobs and more physically or mentally costly to work. This paper Analysing gender differences, in empirical studies dealing with working conditions and health, has been shown to be very important (Bardasi and Francesconi (2004), Robone and Jones (2008)) as both health conditions and working arrangements can vary considerably according to the gender of the worker. It has been suggested that the increasing participation of women into the labour force over the last decades has positively affected their health, and may in fact enhance their health (Hall, 1992). This positive association could be explained by a selection effect, as women might exit the labour force because of health or family reasons. Even if women have the same job title as men, their specific job tasks are often different. Women are generally engaged in work activities different from those of men, thus they may be subject to different exposures (Stellman, 1994). Health responses to exposure may be different for men and women due to biological differences such as the effects of body fat or endocrinological factors. Differences across gender in terms of quality of working conditions that affect health at work (Vermeulen and Mustard, 2000). This study includes the distinct effects of both physical and psychosocial conditions relating to the work environment (Cox, Griffiths & Rial-González 2000) on workers' health. Key mental conditions comprise psychosocial strain, work arrangements, and work organizational factors, whereas physical hazards typically focus on exposures such as noise and workload (Cox, Griffiths & Rial-González 2000, Stock et al. 2005). The contribution of both these dimensions on the health at work are studied. Given the cross country evidence we are able to make generalisations of results in a wider context compared to previous literature. The results suggest that controlling for a wide range of personal and job attributes, (adverse) working conditions are associated with more work related physical and mental health problems. It emerges a different pattern by gender (as shown in Table 7).
Physical health
Physical hazards 0.012** 0.032***
Psychosocial job conditions 0.033*** 0.041***
Mental health
Physical hazards 0.013*** 0.023***
Psychosocial job conditions 0.07*** 0.097***
R-squared 0.082 0.072 0.089 0.084
Nobs 8035 8035 9126 9126
Note: ***1%,**5%,*10%. Sector, occupation, firm size, country fixed effects and cross national weights are included.
Table 7: Working conditions and health by gender
With respect to mental health at work, male workers suffer more from high work demands/ low job autonomy compared to female workers. A less clear pattern across gender is found with respect to physical health problems at work. Results persist also when controlling for the endogeneity of working conditions, in fact the instrumental variable estimation offers evidence in support of a causal effect of (demanding) working conditions on individual' health at work.
Cottini and Buhai, (2011) provide a fresh analysis of the theory of compensating wage differentials (CWD). The growing interest for job satisfaction data among labour economists has, for instance, generated a debate about the main factors explaining the worker's rating of her work environment (Clark, 2005). Many studies point out a trend of declining job satisfaction despite rising real wages. It is uncontroversial that non pecuniary work conditions are considered crucial job aspects by individuals searching for jobs. In seven OECD countries in the 1990s (Clark (2010)) employees overall do not rate income or hours as most important job features, the two aspects with highest ratings being job security and job interest, followed by work independence. These preferences appear consistent across the two genders. Job quality become an important economic policy issue, for instance through the definition of 'decent work' (ILO, 1999) or through the inclusion of 'employment quality' indicators in the European Employment Strategy. These definitions concern a wide series of job dimensions, like minimum wage level, job security, representation rights, job safety, training opportunities, all of which can be affected by recommended labour and social policies. The standard hedonic wage theory has been challenged by follow-up models of equilibrium search, (Burdett and Mortensen (1998)) extending their wage dispersion feature to dispersion of utilities that depend on both pecuniary and non-pecuniary attributes; existing search frictions in matching workers to firms might then lead to equilibria with different configurations of wages and amenities, where the slope of the worker's indifference curve need not equal the slope of the wage-amenities relationship, (Hwang et al (1998), or Lang and Majumdar (2004)). Most research based on estimation of hedonic wage equations has not been able to account for unobservables in the worker and the firms utility This paper proposes a bridge between these competing explanations, using to that end uniquely suited Danish data, controlling for both worker and firm time-invariant unobservables, and for the sorting of workers cross jobs according to their attitudes towards risk, in the conventional hedonic wage framework. The study compares the range of CWD estimates with duration model estimates obtained using the worker employment histories, as justified by the utility dispersion equilibrium-search framework.
Lucifora C.and F. Origo, (2010) draw the attention that in recent years many countries of the European Union (EU) have implemented comprehensive smoking bans to reduce exposure to tobacco smoke in public places and all indoor workplaces. This paper contributes to the literature in the following ways. First, the focus is on the effects of a specific type of public smoking control policy the so called "comprehensive" smoke-free law -- on workers' health within workplaces. These types of smoking bans, covering all public indoor places and all workplaces (either public or private), represent one of the pillars of the EU smoking control policy and in recent years have been implemented in most of the EU Member States, while little is still known about their effects on workers' health. Second, they use comparable micro-data for a large number of (European) countries to study the effect of smoking control policies both on exposure to smoke, as well as on direct measures of workers physical health (such as the presence of respiratory problems). In the Table 8 reports the TCS for the two available years, presenting for each EU-15 country both the overall score and the specific score for comprehensive smoking bans. Countries are ranked according to the date of introduction of such smoke-free laws.

Table 8 Comprehensive smoking bans
The empirical strategy exploits variation in the timing and design of smoking control policies, as implemented by various countries, to assess the causal effect of comprehensive smoking regulations on workers perceived health using a quasi-experimental approach (i.e. a 'Diff-in-Diff' estimator). Whether such bans may produce some "unintended" effects within workplaces beyond those expected on risk exposure and workers smoke-related health is investigated.
Lucifora C. and F. Origo, (2010) provide some details on the diffusion of comprehensive smoking bans in Europe. The authors outline a methodology to measure and compare tobacco control policies across EU countries. Then they present new evidence on the effects of comprehensive smoking bans in European workplaces, considering not only exposure to smoke, but also measures of workers' perceived health such as the presence of work related respiratory problems that should be directly affected by these policies, as shown in the Table 9. They finally discuss whether
Table 9: Share of workers reporting work related anxiety or irritability
comprehensive smoke-free laws may have "other" effects both within and outside the workplace which may partly offset the (positive) effects on smoking behaviour and health.
Böckerman Petri, Pekka Ilmakunnas and Edvard Johansson, (2011) argue that creative destruction is a key process underlying so-called firm dynamics, has been assessed to contribute significantly to the productivity growth observed in economies around the world. The creative destruction process entails simultaneous job creation and destruction and, consequently, induces worker flows. One can easily envisage that a job in an establishment characterized by rapid hiring and firing may be considered to be worse than a job in an establishment characterized by slower worker turnover, as rapid turnover means more uncertainty regarding the future. Increased uncertainty about future job prospects may, in turn, affect employee well-being in the form of reduced job satisfaction. The potential effects of labour market turnover on employee well-being are particularly important, because job dissatisfaction has been shown to be associated with a multitude of 'negative' activities. These include lower job performance, increased absenteeism, more actual and intended job switching, as well as various discretionary activities, like less voluntary overtime, less pro-social activity and less adaptive behaviour. All these are likely to increase the costs of workplaces. Job dissatisfaction is costly also from the society's point of view, if it leads to early retirement or withdrawal from the labour market. The study examines empirically whether a faster pace of creative destruction tends to negatively influence job satisfaction, using a unique dataset for Finland constructed from the European Community Household Panel (ECHP) for the years 1996 2001, and the Finnish Linked Employer Employee Data (FLEED). The results show that if the wage can fully compensate for potentially negative effects of uncertainty in establishments or industries with a high turnover of employees, then this uncertainty should have no effect whatsoever on job satisfaction. In contrast, if this unfavourable job characteristic is not fully compensated for in the form of higher wages, then the uncertainty induced by increased job and worker flows can be expected to exert a negative influence on job satisfaction. The results indicate that employees do obtain compensation for intensive restructuring at the establishment level, whereas the impact of this uncertainty on employee job satisfaction is found to be negligible. The significant effect of a high pace of turnover on real wages and insignificant effect on job satisfaction give consistent support for the existence of compensating wage differentials for uncertainties at the establishment level. At the industry level, high labour turnover seems to have no effect on real wages or on job satisfaction. Hence, compensation for uncertainty occurs at the establishment rather than at the industry level. The establishment- rather than industry-level volatility is the relevant source of uncertainty among the employees. However, the results also indicate that the relationship between job and work flows, and wages or job satisfaction involves important elements of non-linearity.
Mazzolini G., (2011) argues that accidents at work are unforeseeable random shocks that could result in serious health problems or deaths for workers involved. The role of workplace conditions and safety at work in reducing the probability of accidents at work and the duration of absence following an injury is investigated using the EWCS data. Following Lanoie (1991), the endogeneity in the provision of safety at work is accounted for. Endogenous selection arises from two different sources. Differences in risk aversion may influence workers' career preferences entailing a selection effect. Employers' behaviour may also affect safety at work when they choose safety expenditures national laws and regulations and for achieving quality management certifications. Thus controls for cross-country differences in occupational health and safety regulations and for innovations in organizational practices are included. The main results show evidence of an inverse relation between safety at work and accidents. An increase of 1 percent of safety at work standards may reduce the probability of an accident of 2.7 percent and the duration of absence of 46.2 percent. When separate analyses are performed by personal characteristics, firm attributes and working conditions, we show that the effect of safety at work in reducing the probability of an accident is statistically significant when a job is characterized by an intrinsic risk at work. Increasing safety at work standards does not produce effects on workplace accident rates, where occupational risks are limited and workers are almost never exposed to dangerous agents.
Schneider, and Beblo (2010) argue that the state of occupational safety and health, Germany has achieved a high level of safety and health at work (BAMS and BAuA 2009), decreasing numbers of occupational diseases and accidents. Working conditions and the structure of the labour force have been subject to rapid changes which have come along with new challenges for safety and health at work. These labour market changes affect the mental health of employees as work-related mental problems are often associated with poor working conditions and non-standard employment OECD (2008). Academic re-search on the determinants of OSH indicators focuses mainly on sickness absence and early retirement. Most studies are based on the GSOEP, a rich micro-data panel, albeit without information on health. The empirical analyses are not always convincingly addressing or correcting for methodological problems such as reverse causality, unobserved heterogeneity or measurement error. These types of problems may cause standard regression methods to produce biased and/or inconsistent estimates that cannot be interpreted unambiguously. Furthermore, the causal impact of working conditions on health outcomes is not clearly established and measurement errors are likely to occur.
Latreille, Jones and Sloane, (2012) use a unique set of questions included as part of the 2001 British Social Attitudes Survey to consider the determinants of individual employee perceptions of a global measure of workplace health and safety that abstracts from the risk and other potential biases associated with more conventional measures such as perceived risk. A substantial literature considers the factors associated with perceptions of individual risk, while a related body of work has focused on how risk might be compensated for with appropriate wage differentials. The former potentially suffers from well-established cognitive biases which leads individuals to understate their actual risk, while aggregation issues beset the latter (Pouliakas and Theodossiou. 2011). Among psychologists and practitioners, a focus has been on conceptualising and measuring so-called 'safety climate' and its relationship with safety performance. The former is usually constructed using factor analysis of questions, with the elements often varying among studies according to the survey instrument used. This is typically then linked to measures of safety performance such as workplace accident rates using structural equation modelling with various mediating and moderating factors posited. The current study seeks to avoid such problems by considering a simple, single, global measure of perceived workplace health and safety, essentially treating this as a measure of workplace well-being for a representative sample of employees and former employees drawn from a cross-section of occupations and industries. The results show that even after controlling for a wide range of individual, job and workplace characteristics, individuals with previous experience of accidents at work and of work-related ill health, regard their workplaces as less safe and healthy than those without such exposure, as do those who were aware of instances affecting others at the workplace. The impact of individual experience is found to be greater the more recent was the accident or health problem and if with the current (or for those no longer working, last) employer. Conversely, workplace health and safety is found to be higher in workplaces with a safety officer, with regular consultation on health and safety issues, and where the organisation complies with the requirements around safety posters/leaflets.
Dodd, Drakopoulou, (2011), argues that the family firm organizational form is an important element of even the most modern industrial landscape. Dramatic examples of the workplace risks run by family business managers are staple ingredients of the news media. Although little empirical data is available specifically studying this labour market group, nevertheless "anecdotal evidence has long suggested a high degree of chronic stress among the self-employed" (Jamal, 1997: see also Akande 1994; Boyd and Gumpert 1983; Yusuf 1995). Jamal (1997) found that self-employed workers reported higher levels of stress and psychosomatic health problems than salaried employees. Whilst comprising just 7% of the US workforce in 2004, the self-employed made up 20% of workplace fatalities (Pegula, 2004). It thus appears that (small) business owner-managers are an especially vulnerable group, in terms of occupational health and safety. One of the principle elements within the SME environment is that of kin involvement in business ownership and management. The family firm remains a significant economic form across Europe, and beyond, and is not only to be found within the SME sector - which it dominates. Many very large businesses indeed exhibit this ownership and management structure. The interactions between family and enterprise create very specific health, safety and stress factors, some of which may be seen to be generally positive, and others quite the reverse. Characteristics of the 3 family firm which can be anticipated to reduce exposure to health and safety risks include strong organizational identity and engagement, and heightened social support within the workplace. Family-firm specific dangers comprise, inter alia, the transfer of domestic conflict to the working environment, kin rivalry and the taking of heightened risks to support one's family. It is perhaps surprising that this topic has received very little attention indeed, in spite of its clear importance. The aim of this study is to develop an initial, tentative conceptualization of health and safety at work, within the family firm context. The methodology utilized will combine an extensive literature review with the development of a basic conceptual framework, and use findings from family firm research to illustrate and underpin this model. In line with common practice a family firm is defined as a business enterprise where ownership control is held by two or more members of a family (or a partnership of families); there is clear strategic influence by family members on the management of the firm; there is a real concern for family relationships, and where we can observe the dream (or the possibility) of continuity across generations.
(v) Research on Demographic Effects and OSH
Sloane, Jones, Latreille, and Staneva, (2011) use the fourth EWCS (2005) to examine the impact of age on work-related self-reported health outcomes. As the proportion of the workforce aged 55 and over continues to grow a trend given additional impetus by recent or proposed raises in the (statutory) retirement age in several European countries including the UK, France, Germany, Spain, Italy and Greece closer examination of the work-related health of older workers is imperative. However, most research into the relationships among work conditions and health has not explored age as a variable of explicit interest, instead treating it as a potential confounding factor, either 'partialled out' statistically or simply ignored (Griffiths, 2000). The aim of the current study is therefore to investigate age differences in a battery of self-reported work-related health measures using cross-country data from the 2005 European Working Conditions Survey (EWCS). Five work-related 'outcomes' are used to assess overall work-related health and well-being: health and safety risk perception; mental and physical ill health; sickness absence; injury rates; and work-related fatigue. However, since the EWCS contains information on working respondents only, a key issue for the analysis is to account for a potential 'healthy worker effect'. This effect arises if older workers still in the labour market have better underlying health than those who leave employment, a source of selection bias (Li and Sung, 1999). We address this issue by making use of an external data source the European Social Survey (ESS) in order to account for labour market non-participation, adapting the Wolinsky et al., (2009) re-weighting approach based on propensity scores for being in employment. In particular, the employment model estimated with 2004/05 ESS data is used to predict probability weights, and the inverse of these are then used to weight the observations in the main EWCS-based data models. In summary, the results which do not adjust for selection or endogeneity find that older workers are no more likely to report work-related health problems than those of prime age. Controlling for the endogeneity of working conditions does not affect this qualitative pattern of results. However, once selection bias is accounted for, those aged 55-65 years tend to be at higher health and safety risk. Accounting for the 'healthy worker effect' thus has a substantial and demonstrable impact on the results for older workers.
Economou, and Theodossiou, (2011a) examine the relationship between recessions (approximated by country unemployment rates) and fatal and non-fatal work accidents. The literature provides ambiguous conclusions; the empirical findings are sensitive to the choice of countries or occupational sectors and to the time period examined. The present study revisits the issue of the work accidents/unemployment rate relationship using a panel of thirteen European Union countries (Austria, Belgium, Cyprus, Denmark, Finland, France, Greece, Ireland, Italy, Portugal, Spain, Sweden, UK) for the time period 1980-2006 and disaggregated by industry fatal and non-fatal work injury rates. The first step in investigating the unemployment rate/accidents at work relationship is the estimation of the Fixed Effects and Random Effects specification model for the fatal and non-fatal work accidents. The results show that after controlling for the country level of GNP per capita, a negative and statistically significant relationship between national unemployment rates and both fatal and non-fatal work injury rates is observed. The study takes into account cross-panel correlations and groupwise heteroskedasticity and distinguishes between the temporary and permanent effects of unemployment on both fatal and non-fatal work accidents. In doing so, the Mundlak (1978) decomposition methodology is utilized and the following regression models are estimated:

The term expresses deviations from the country average unemployment rate and measures the temporary effects. The term expresses the difference in the country average unemployment rates and measures the permanent effects.
Table 10 shows that there is a transitory negative relationship between the unemployment rate and the fatal work accidents. There appears to be a significant positive permanent effect. The pattern is similar when the effect of recession on non-fatal work accident rates is examined. The transitory effect of recessions on non fatal work accidents is negative and the permanent effect of recessions on non fatal work accidents is strongly positive. The above mentioned results are uniformly consistent with the results obtained after the industrial disaggregation, with two exceptions namely the manufacturing and agriculture sectors where there is evidence of a transitory underreporting of non fatal work accidents and injuries. The results may imply that at the initial stages of the economic downturn work accidents tend to decrease as the slowdown of production reduces the number of job shifts, ease the work intensity and the proportion of newly hired and less inexperienced workers decrease as last-in-first-out dismissal practices are implemented by employers who are keen to retain their most experienced workers. Furthermore, the job creation rate greatly declines and this further reduces the proportion of the inexperienced workforce. This entails a reduction in the incidence of work accidents and injuries. Hence, the transitory effect of the unemployment rate on the incidence of fatal and non-fatal work accidents is negative.

Table 10: Short term and long term effects of unemployment on OSH
However, at later stages of the recession, cost cutting practices by employers may affect the OHS investments as firms reduce their expenditure on training and safety equipment and/or workers in employment, facing an increased risk of job loss, tend to undertake far riskier job tasks in a world of scarce employment opportunities. Furthermore, high unemployment severely weakens the ability of the trade unions to protect their rank and file from the degradation of the health and safety standards or succeed in improving the OHS procedures. Hence, the permanent effect of unemployment on work accidents becomes positive. Conversely, at the initial stages of economic expansions, an increase in work intensity occurs as firms react to the pressure of increasing aggregate demand. This effect is reinforced by the reluctance of employers to hire new personnel before they are confident that the increasing demand is long lasting. Hence, there is an increase to hours of work and an increase to the pressure for increased productivity of the workers which might induce them to be less careful with safety procedures. With the passage of time as the economy adapts to the improved state of affairs, productivity improvements during economic expansions encourage firms to invest in occupational health and safety capital. As the bargaining power of labour unions improves during economic expansions, their increased concerns and pressures towards employers to increase workplace safety contributes to the decrease of workplace injuries. Thus the permanent effect of unemployment rate on work accidents turns out to be positive.
Economou and Theodossiou, (2011) note that the empirical evidence on therelationship between unionization and workplace injuries is conflicting most likely due to endogeneity. Indeed, workplace injuries are affected by increased unionization but at the same time, workplace injuries may also affect the degree of unionisation since high accidents rates may motivate workers to organise in unions in order to protect themselves from the hazardous working conditions. The present study investigates the effect of unionisation on work-related injury rates using a panel of ten European Union countries (Austria, Denmark, Finland, France, Ireland, Italy, Portugal, Spain, Sweden, UK) during the period 1982-2006. The degree of unionisation is approximated by the union density index. In order to take into account the time persistence in work injuries and the endogenous nature of the work injuries unionisation relationship, system GMM regression techniques are utilised. The main variables of interest are "Total fatal injury rates per 100,000 employees" and "Total non-fatal injury rates per 100,000 employees". The independent variables of interest are "Union density" and "GDP per capita (in PPP)". The instrumental variable used to control for the endogenous relationship between work injuries and union density, is the "days lost due to strikes and lockouts per 100,000 employees". Two methodological shortcomings are addressed; the dynamic nature of work-related injuries, since the experience of a work related health problem at present may be affected by work related health condition suffered in the past, and the endogenous nature of the relationship between trade union density and work injuries by the use of system GMM models. The results are presented in Table 11.

Table 11: The effects of unionisation on OSH
Union density appears to be a strong determinant of work-place fatalities. An increase in union density is associated with a lower rate of fatal work injuries, indicating that increasing union density helps unions to achieve better outcomes on occupational health and safety conditions. The findings are similar when one considers the effect of union density upon non-fatal work injuries. Non-fatal work injury rates move procyclically to economic conditions as approximated to country GDP. These findings suggest that during economic expansions non-fatal work injuries are decreased and this may be attributed to the higher investment on occupational health and safety issues undertaken by employees at periods of economic booms.
Drakopoulos and Theodossiou (2011) point out that many studies on individual beliefs concerning risk have shown that people often underestimate and/or overestimate risks. For instance Kahnemann and Tversky (2000) suggest that in decision making people consistently underestimate outcomes that are merely probable, in comparison with outcomes that can occur with certainty. They also argue that individuals are unlikely to perform the operation of subtracting the cost from the outcomes in deciding whether to buy a gamble. In addition, there is ample empirical evidence concerning similar behaviour towards risk in the context of occupational environment. Much specialist scholarship indicates that workers constantly underestimate their exposure to work risk for work accidents, since perceptions of risk are influenced by pre-existing, recent or readily available experiences (Thaler and Sunstein, 2008) or overestimation of personal immunity from harm (Weinstein, 1989). This paper studies the repercussions of the above theoretical developments in the occupational safety and health framework (OSH) when the workers systematically underestimate job risks. In doing so the paper first briefly reviews the standard debate relating to the implementation of OSH. Although, it is generally accepted that there is a need for regulation in the case of job risk underestimation, there is no much work regarding the type of appropriate regulation. The paper provides a discussion of the types and impact of regulation on health and safety effort in a simple framework, in which workers beliefs concerning accident risks at workplace interact with the behaviour of Health and Safety regulator. One can investigate the above in a simple framework utilising research on risk perceptions. A discussion of the general case of the difference in beliefs concerning risk perceptions and regulation can be found in Viscusi, (1998); Salanie and Treich, (2009). In the specific context of safety at work, one can assume that there is a difference concerning the assessment of a job risk between workers and the work safety regulator, and that there is agreement between the two parties concerning all other preferences. Contrary to the standard approach, the health and safety regulator does not impose a maximum level of job risk but instead he/she adjusts the level of effort to work safety. The paper shows that a particular type of regulatory intervention is necessary for the risk underestimating workers not to suffer a welfare loss.
Ferrer-i-Carbonell, Haafkens and Theodossiou (2012) investigate the preferences of the employers and employees regarding aspects of occupational safety and health (OSH). The provision of OSH is determined by the interplay of preferences of workers and employers. Hence, the need for a though understanding of the determination of the preferences of both workers and employers on OSH is of critical importance in order to identify areas of required action and to set priorities for policy initiatives on improving health and safety at work. The need to focus on health and safety is important since the interplay of preferences of workers and employers need to be completed by OSH strategies designed to address the consequences of a continuous adaptation. From the employers' side a sufficient condition for profit-maximising is that optimal OSH level should equalise the OSH costs and OSH benefits at the margin. The outlays associated with improvement in workplace conditions (preventative practices, OSH training and the like) rise as the level of health and safety increases. The benefits are associated with having to pay lower ('compensating') wages and with the reduction of workplace injuries and illnesses (lower level of sickness absence and sick pay and the like). It is therefore clear that firms differ in their ability to trade-off the costs of providing a safer working environment with the expected benefits (particularly, a lower wage bill). For the workers' side it is assumed that rational workers, who possess perfect information and exhibit a high degree of labour market mobility, demand a wage premium as compensation for their willingness to assume elevated occupational health and safety risk. In order to maintain expected utility constant, a given worker will therefore demand higher levels of wage compensation for higher degree of job risk. Inefficiencies are likely to arise since competitive markets require the existence of full information between the two sides of an employment relationship regarding workplace risks. Recent research has employed the stated preferences methodology of estimating the "willingness-to-pay" (WTP) of individuals for OSH. Subjective WTP estimates typically utilise survey methods in order to elicit the magnitude of the rate at which individuals (employers or workers) are willing to forgo income in exchange for additional OSH, which would nevertheless keep their utility unchanged:
where in a perfectly competitive labour market and applicable to a range of risk increments r for income w. Eliciting subjective evaluations of by survey respondents is a useful tool for the understanding of preferences. This paper is a first attempt to investigate the workers' OSH preferences and to compare them with those of the employers after controlling for a host of personal, occupational and firm characteristics. The study uses conjoint analysis. This involves choices or evaluation responses by an individual concerning various OSH initiatives. This study contributes to the growing literature on OSH. At this stage of the research agenda, it is crucial to understand and operationalize the WTP tool for the two participants in the labour market in order to understand their motivations regarding economic decisions for OSH
The data for this study are derived from the identical HEALTHatWORK project surveys, among employers/human resourse managers and manual workers in the Netherlands, and the U.K. via an Internet survey. The sample consists of firms or employes/ Human resourse managers or non administrative workers. Individual respondents are obtained from the databases which are maintained by the interviewing company (SYNOVATE, NL).
The estimated coefficients show that he most relevant OSH attributes for the employees seem to be those related to accidents and incidents and the risk assessment. These are followed by work illness investigation and whether there is safety and health training offered by the employer. The least important OSH attributes are those relating to caring for the injured and how the company plans and facilities their return. The results for the employers or human resource managers indicate that the most relevant OSH attributes for the employers or human resource managers are, as in the case of employees, those related to accidents, incidents and the risk assessment. In addition, employers also value safety and health training (more than employees do). In contrast, the employers or human resource managers appear to value work illness investigation less important (in contrast to the employees). It appears that respondents reply strategically flagging that they do not care about costs but only about the OSH improvements. This issue calls for further careful research in the appropriate survey methodology to circumvent this strategic behavior from the part of the employers or human recourse managers.
P?ci??o et al (2012) use a willingness-to-pay of company executives for occupational safety and health (OSH) as an attempt to specify whether and how much, company executives value the health and safety of company workers. The study was performed as part of "HEALTHatWORK" research programme under 7th Framework Programme of the European Commission (FP7). The willingness to pay data has been collected among 147 representatives of managers from 25 medium- and big-sized enterprises in Poland with use of a questionnaire. Findings obtained show that 70-80% of the managers interviewed are willing to increase expenditures on different activities aimed at improving occupational safety and health performance in their companies. The research output allows the evaluation of company managers' willingness-to-pay to increase the OSH costs. The analyses results show that the managers are ready to increase the spending by an average of about 30% in respect of measures to reduce accidents at work. The important finding is that managers believe that increasing workers engagement in this type of activities is vital because they declared the willingness to increase the spending on this account by about 37% on the average, and to augment the cost of OHS training and information dissemination by 34%. The relatively largest spending increase, by about 38% on the average, was declared regarding the measures aimed at reduction of work related stress. It needs to be reminded that the declarations concerned the increasing of spending to be incurred in the future as compared with the current level of spending for a given purpose. Majority of respondents did not specify the current spending amounts, which could also affect their tendency to increase spending. Given the fact that the respondents could not estimate the amounts of the OSH costs incurred in their companies (only two respondents answered a question about such costs), the study did not allow to assess directly the willingness of the company management to incur the costs of occupational health and safety in monetary terms.

Potential Impact:
PLEASE ALSO SEE THE ATTACHED PDF FILE - Potential Impact and Policy Implications_2012x.pdf

The potential impact (including the socio-economic impact and the wider societal implications of the project so far) and the main dissemination activities and exploitation of results

Potential Impact and Policy Implications
(a) On Absenteeism
Barmby (2012) research has the potential to allow a more accurate interpretation of empirical absence studies, and gives a new way of thinking about the effect of policy in this area. In Ercolani (2012), the long-run socio-economics patterns of sickness absence brought to light by the research are either ones which verify existing ideas or are ones that seem self-evident when they are first shown. It is actually unusual to have statistical research where the results are so consistent with expectations and the reason may be the use of a large, four-million individuals, sample size. These results are useful because they provide information that can be interpreted by interested non-specialists. The results show that, unsurprisingly, rates of sickness absence seem to fall as hourly earnings increase. There are regional variations in the rate of sickness absence, though there is no single obvious explanation for this. Rates of sickness absence also vary by industry and occupation, and these seem to be connected to the status of the occupation and the nature of the work undertaken. Two particularly strong results suggest that rates of sickness absence increase with time in tenure with an employer and with the number of employees at the workplace. Employees in the first few months of tenure have very low absence rates and employees who have smaller numbers of co-workers at the workplace also tend have lower absence rates.
The findings of Ercolani (2012b) inform policy makers as to the formulation of regional policies that aim to improve the health of the workforce and diminish regional differences in workplace absenteeism. We find that half of all the explainable regional difference is due to the higher hourly wages offered in the South of the UK. Therefore much of the regional difference in absenteeism is probably due to two factors. Firstly is the incentives that workers face in attending an extra hours of work either because of the direct wage remuneration or because better attendance translates into higher wage rates in the future. Second are the direct health effects that a higher income can provide in terms of a healthier life. We can say little of the latter but with respect to the former we can say that this is a straightforward matter of incentives. The prediction turns out to be an obvious one but backed by statistical evidence; relative wage increases in the North would drive a drop in absenteeism to the same proportion as they would in the South. The other characteristics that are associated with higher sickness absence are being female, holding a second job and working in the public sector. These latter results are driven by females still being typically responsible for what is generically called "home production", holding a second job evidently imposes an opportunity cost on the main job and being in the public sector may infer a degree of job security that permits greater sickness absence. Half of the solution to the regional differences in absenteeism is therefore simply a matter of economic development and creating opportunities for regional labour markets. However, simply reducing absenteeism should not the only target because increasing output is only an intermediate target to achieving our real target of increasing welfare (or utility in the economists' jargon). We already know that achieving a work-life balance is an important objective but simply increasing incentives for nobody to ever be absent should not be the target nor is it achievable. We would not wish workers to continue to attend work when genuinely ill because they may pass on transmittable their illness to others and also because this may harm their long-term health status.
Beblo (2012) shows the statistical relationship between absenteeism and family obligations has practical implications for personnel policy as well as social policy: if in the lack of public care infrastructure women are considered responsible for household and care work, employers will expect women (with children) to be absent more often than men. Statistical discrimination against women or individuals with household and care responsibilities might be the frequently discussed consequence. The existence of different societal norms regarding the absences of female and male employees is demonstrated in Patton Eric and Johns Gary's (2007) analysis: higher absences of women are not only expected but also legitimized. However, our findings indicate that this stereotype does not represent actual behaviour of German employees any longer. Nonetheless, in order to avoid a self-fulfilling prophecy of an employer's expectations of a care-related higher absence rate of a female employee, resulting in less pleasant job characteristics and working conditions, social policy should provide an infrastructure of care facilities that would ease the link between absenteeism and family obligations.
Schreiber (2011) give a first illustration of our approach (of using Google symptom search data as a proxy for actual sickness) and present preliminary findings with German data. The predominant problem of our analysis is the very short sample size: the German diagnosis-specific sickleave data was only available until 2009, apart from the fact that it is only published at an annual frequency. The weekly internet search data is downloadable from Google for the years starting in 2004. This time span does not really represent a "longer run" in the usual sense. Thus our empirical results should be viewed as an illustration of our suggested methodology, not as robust evidence, and be followed up in the future (or for other countries) when more data become available.
Böckerman, Petri, Johansson and Kauhanen, (2011) state that the finding of minor, if any, effects of workplace innovations on sickness absence and accidents at work stands in sharp contrast with previous results reporting a positive relationship between workplace innovation systems and sickness absence. However, these previous studies have, as noted above, mainly considered cumulative disorders and other specific injuries or illnesses, not the overall impact on employee health. This implies that even in the case of non-negligible effects on single health aspects, the overall impact of the adoption of innovative workplace practices may, nonetheless, be neutral or even health improving. There might, however, also be other important aspects influencing the results obtained. In particular, the tradition of close cooperation between employers and employees characterizing Finnish labour market institutions may well affect the results. This would suggest that improved cooperation between employees and employers could help in reaping the benefits to be gained from wider use of innovative work practices.
Drakopoulos and Grimani, (2011) highlight that in recent years, there has been an increasing interest concerning the economics of absenteeism, spurred by a growing awareness that the economic and social costs of the phenomenon are quite substantial. Thus, it is not surprising that many studies have concentrated on the determinants of absenteeism in an attempt to discover ways and policies to reduce it. In contrast to other European countries, the issue of absenteeism in Greece has not been the subject of systematic investigation. This paper utilized a large sample to test the issue of injury absenteeism job satisfaction relationship. There is a lack of attention to this specific type of absenteeism and its relationship with job satisfaction. Hence, given the limited empirical research based on Greek data, this study also attempted to fill this particular gap. However, more similar research and inter country comparisons of absenteeism similarities and differences are needed if appropriate policy recommendations can be identified.
Blazquez (2012) states that in order to reduce the incidence of absenteeism due to sickness leave it is vital to understand in more detail the reasons behind this. This is the main purpose of the present paper. Using quarterly data from the Spanish Labor Force Survey for the period 1996-2004, it examines which demographic and workplace characteristics are more likely to increase the rates of sickness absence. The sickness absence rate varies between 2% and 2.5%, and the values have remained stable along the period under consideration. Both the descriptive and the econometric analysis reveal that females are more likely to be absent from work due to sickness than their male counterparts. Furthermore, gender difference are found as regards the effects of marital status on absenteeism, with married females being more likely to be absent from work than their single counterparts, while the opposite is observed among males. Age also plays a role in explaining absenteeism. In particular, the analysis confirms previous findings on increasing prevalence of sickness absence among older workers. Public sector absence rates exceed private sector ones. The disparity in these figures is often seized upon by conservative commentators as proof that the public sector has a "relaxed attitude" towards employees taking time off work when compared to the private sector. This result, then, claims for stronger effort to prevent absenteeism in the public sector. The lower rates of sickness absence observed among temporary workers reveal that absenteeism is also related with job insecurity. Finally, the results point out that socio-economic status play a role in explaining the risk of sickness absence. This conclusion is extracted from the potential impact that occupation and education ? closely related to socio-economic status ? exert on absenteeism.
(b) On exposure to job strain
Sultan-Taïeb et al. (2011) study provides a comparison of job strain exposure covering a large number of European countries based on harmonized data. We showed that differences between countries persisted when individual characteristics (age, gender), economic activity and occupations are controlled for. Greece, Cyprus and Slovenia had the highest job strain prevalence, whereas Sweden, Denmark and Latvia had the lowest. Differences in institutional frameworks between countries, such as welfare state regimes, showed that working in a Southern or Eastern European or Bismarkian regime was associated with a higher risk of exposure than working in a Scandinavian regime. This study may help enhance our knowledge of job strain exposures and therefore assist decision-makers in orienting prevention policies in order to improve working conditions at European level.
The results of Sauze et al. (2010) invite an investigation into the reasons why Sweden succeeded in reducing exposure of their workers to psychosocial risks. Prevention policies like a systematic analysis of the consequences of organizational changes on psychosocial risks could be part of the explanation. More research has to be undertaken about national prevention policies to validate this assumption.
Jones et al. (2011) state the evidence presented in their paper suggests that employers are able to influence the levels of job anxiety amongst their employees by matching job demands to skills, giving employees more discretion or control over how they do their work and providing a working environment that is perceived (by employees) to be supportive and fair. Further, there is evidence that employee reported measures of job anxiety contain valuable information which is correlated with management reports of work-related stress. The evidence also suggests that workforce stress and anxiety may be detrimental to workplace performance. However, employers need to consider the costs associated with improving psychological health before it is possible to conclude that there is necessarily a business case for improving employee psychological health.
Haafkens et al. (2011) identify that chronic diseases are increasingly prevalent among people of working age in Europe. In the Netherlands, about one third of the working population (37%) has a chronic disease. Many chronically ill workers (30%-60%) feel more or less restricted by their condition at work, and they are more likely to exit their job prematurely than their healthy colleagues. Early job discontinuation can lead to negative health and socioeconomic outcomes for the individual and increased costs to employers and society. To reverse these problems, recent EU and national social and employment policies are encouraging employers to facilitate continued employment for employees with long-term health problems and disabilities. In recent years Dutch employers have improved absence management programs for employees who are ill. The available statistics indicate that this has led to an over-all decrease in absenteeism and work disability rates in the Netherlands since 2004. Yet, an increase of work disability rates has been observed among employees with some chronic conditions. In the past occupational physicians and occupational health services were mainly responsible for the development and implementation health and safety programs in Dutch and European companies. Today, line managers (LMs) and human resource managers (HRM) are playing an increasingly important role in workplace health management. This concept mapping study provided insight into the perspectives of Dutch LMs and HRM on what is needed to facilitate continued employment for chronically ill employees. The extent to which these perspectives are shared by LMs and HRM working in other countries or organizational settings remains to be determined through other studies. Despite this caveat, the study provides important new information about topics that occupational health researchers and planners should take into consideration when developing job retention programs for chronically ill workers.
The conclusion of Haafkens, Kopnina, Meerman, van Dijk, (2011) (2009) is that it appears that there is a gap between the formal structures promoting employment and actual experiences of the stakeholders within organizations. Flexible, supportive and open organizational culture emerged as one of the key factors for optimal functioning of the chronically ill employees.
Barnay et al. (2010) and Serrier et al. (2009) have similar policy and societal impacts and will be addressed simultaneously.The economic evaluation of occupational and safety health interventions could play an important role in the orientation of prevention policies. Health economics provides an interesting set of methods implemented in cost-benefit analyses, which could be used in this specific field. Moreover, the evaluation of the costs imputable to occupational exposures, such as psychosocial risk factors, underlines the socio-economic burden of occupational risks. A lot is at stake since the implementation of prevention policies might reduce this burden for society as a whole as well as for employers. Comparative analyses of different prevention interventions might be useful for establishing priorities and trade-offs, based on a cost-benefit criterion. A more in-depth analysis of occupational exposures in the different European countries allows to identify the most at-risk groups of workers and to orientate prevention policies towards such groups. It is also a way to underline the distribution of costs among the different diseases, types of exposures and economic sectors in order to focus interventions on these specific issues.
Serrier (2011) highlights the economic impact of respiratory cancers imputable to occupational exposures. This may help policy makers prioritize resources for prevention policies, given the high burden these exposures represent for society as well as for employers and social security systems.
(c) On working conditions and Deprivation
Cottini (2011) says that knowing how health depends on work environment and employment arrangements is clearly of policy relevance as it provides key equity considerations to complement the efficiency argument advocated by employers. Moreover, the extent of these problems seems to vary across countries according to the legal and social protection for worker's health and safety.
Blasquez (2012) shows that low relative income may contribute to socioeconomic disparities in health. Efforts to eradicate socioeconomic differentials should take into account psychological and physical perceptions and self-esteem in addition to absolute material resources.
Mertens and Beblo (2011) conclude that the costs of the financial and economic crisis have not been limited to the well-documented fall in the domestic products but include psychological costs such as lower satisfaction with life in general as well as job and health in particular especially of West Germans. Despite Germany's subsidized short-hour work programs to keep people in employment and despite higher employment protection in Germany than in the UK, only Germans showed relatively strong emotional reactions during the crisis years particularly on reported job and health satisfaction.
Blazquez et al. (2012) observe that research in social sciences has increasingly paid attention to the controversial relationship between income, relative deprivation and health, but findings are ambiguous mostly because of lack of proper data that are able to follow individuals longitudinally. To the extent that individual deprivation frequently extends to different domains ? apart from income ? our results warn that their consequences on individual health may be overwhelming. We suggest that other life domains should not be neglected and claim for the necessity of a renewed approach where more emphasis should be directed to other dimensions reflecting some minimum standards of living. This may be potentially relevant for policy makers in the design of the most appropriated policies and the most effective targeting of resources intended to improve the health of citizens, and that so far have been almost exclusively focused on income support.
The policy implications of Cottini and Lucifora (2010) for health expenditures, work related disability benefits and, more generally, for worker's overall well being are also of some interest. Policy measures should combine renewed efforts to monitor working conditions at workplaces, help firms to improve the overall quality of work, as well as regulating work-related physical hazards and psychological job stressors.
Cottini and Ghinetti (2011) make the following observations and recommendations. From a policy perspective, the result that drinking has a negative impact on mental health is a novel and interesting result, especially for Denmark. Indeed, a report commissioned by the European Union concludes that: however much the continent associates alcohol with Ireland, much of the EU has a serious drinking problem .with Denmark being something of a standout., while, for example, there is now less concern for the consequences of smoking, which is decreasing. This suggest that the interventions aimed at promoting good lifestyle practices and better working conditions should be particularly targeted to specific behaviours (drinking) or conditions (job security), and that they may be particularly effective on those health component (the mental ones) that are increasing in their importance in modern societies and workplaces.
Cottini et al. (2011) observe that voluntary turnover appears to be a rational worker response to adverse workplace conditions, and unless the firm alleviates its adverse workplace conditions directly or mitigates their effects on voluntary turnover through HIWPs, workers exposed to adverse workplace conditions will likely continue to take the exit option. Granted the implementation of HIWPs is not free, however, we are not advocating the implementation of HIWPs without conducting a careful cost benefit analysis of HWIPs. Since we do not have adequate data to estimate the cost of implementing HIWPs in Denmark, we are unable to conduct a careful cost benefit analysis. Nonetheless, considering that the cost of voluntary turnover is estimated at close to 30% of annual salary for the average Danish worker, the benefit of HIWPs as a mechanism to mediate the quit-inducing effect of workplace hazards may be considerable.
Cottini (2011) state that policy makers should then make efforts to evaluate the cost, both at the economic and social level, of health problems deriving from an adverse work environment, focusing on the role played by new risk factors such as psychosocial hazards.
Cottini and Buhai (2011)observe that in situations in which the risks are not known to workers, as in the case of health hazards or situations in which the labour market is not competitive, market forces might not operate effectively to internalize the risk. These cases provide an opportunity for cost effective government intervention.
Lucifora and F. Origo (2010) find empirical evidence which confirms that comprehensive smoking bans are an effective policy to fight exposure to tobacco smoke: compared with country rankings in terms of the Tobacco Control Scale indicator, our empirical results actually provide additional support to the effectiveness of comprehensive smoking bans in curbing exposure to tobacco smoke and work-related respiratory problem.
In a similar Lucifora and Origo (2010) conclude that comprehensive smoking bans are an effective policy to fight exposure to tobacco smoke, but more effort is needed in trying to identify and measure their potential "side" effects in order to implement the proper policy mix. For example, given our evidence on the unintended increase of mental distress, their introduction should be accompanied by psychological counselling and/or treatment to help those workers quit who used to smoke at the workplace.
The results reported in Böckerman et al. (2011) imply that changes in real wages may well counterbalance negative direct effects of turnover on employee well-being in the form of reduced job satisfaction. While the evidence concerns Finland, a country characterized by substantial wage compression, this counterbalancing effect can be expected to be even stronger in countries with a less compressed wage structure.
Kopnina and Haafkens (2009) discovered that the topic of the chronically ill employees is largely invisible to HRM practitioners, line managers and diversity specialists who do not always have the right instruments for implementation of the European or national frameworks. Line managers and human resource managers use their personal opinions for handling the chronically ill employees. Most practitioners are unaware of the impact of chronic illness in their organizations and in employees work life. Organizations vary in the ways in which information about the chronically ill is disseminated, processed and how the decisions are made. When our database is expanded with a wider set of case studies, we may be able to classify the organizations in accordance to different institutional or cultural characteristics and deduce which arrangements work better in enabling the chronically ill employee to continue optimally functioning at work. The group of chronically ill employees does have some common characteristics, including the long duration of chronic illness; and physical or mental limitations. There are also common needs that most group members can identify with, such as requirement of ongoing medical care, the need for recognition and acceptance, and often the need for more flexible, open and humane organizational culture. Approaching the chronically ill employees as a group may be constructive in both serving the employers' need to retain good workers and the employees' desire to retain job and the need to be treated equally. Greater efforts are needed to translate EU policy as regards to chronically ill employees into HRM policies and practice.
In investigating the determinants of workplace accidents, Mazzolini (2011) pays particular attention to highlight the policy implications of our findings. Referring to preventing accidents at work, we find that more stringent duties imposed on employers and workers by occupational health and safety regulations may often lead to an increase of costs in providing occupational health services that is not compensated by a significant reduction of workplace accident rates. Introducing laws in favour of the Workplace Health Promotion (WHP) may provide an alternative policy strategy to incentivize the employers in guaranteeing more safety at work. For instance, a financial aid for the employers who provide innovations in work organization practices may lead to lower probability of an accident and lower costs, for employers, in setting and organizing occupational health services.
Despite the limitations, Schneider and Beblo's (2010) diagnosis for health research in Germany is rather optimistic. Many datasets of high quality with a large number of health-related indicators are available. It seems that at least economic research has not yet exploited them to their full potential.
The findings of Latreille et al. (2012) reveal that employees' perceptions of workplace health and safety are adversely affected by previous individual or workplace experiences. To the extent this may be linked to employee well-being and to job satisfaction, employee engagement and ultimately performance, employers need to ensure that they are seen to respond when such issues arise, taking appropriate action to reduce this effect, and possibly providing appropriate support to victims. The evidence also shows that, in conformity with the safety climate literature, measures such as appointing a safety representative, good communication around health and safety and compliance with even cheap and simple regulatory requirements such as displaying a health and safety poster or providing leaflets enhance how safe and healthy employees consider their workplaces.
Drakopoulos et al. (2012) identify that although the legal framework in Greece is quite adequate, there is a need for both prevention strategies and enforcement of the existing safety regulations. The need for more intense and systematic inspections at workplaces, as well as the need in training and OSH education of the labour force is emphasized by the majority of the relevant surveys presented in this study. OSH education should be a priority target so that employees can acquire adequate knowledge on both the risks and the prevention strategies of work-related injuries (Alexe et al., 2003). On the other hand, enrichment of worker protection legislation over the last decade should be combined with stricter monitoring measures by local SEPE offices (Lamprousaki, 2009; Soumeli, 1998) who also have to be reinforced given the significant shortage of staff (Bazas, 2001). While OSH is an issue of interest in current research at the international level, a substantial research gap is observed in Greece. Studies that examine work-related accidents and diseases are quite limited in number and they have the disadvantage that they are of a cross-sectional character. Furthermore, the existing research results should be treated with caution given that the majority of the studies draw information from small samples. In addition, more systematic research should be done on the determinants of injuries and on their effects on job participation and productivity. Priority should also be given to the population groups that report systematically higher incidence of work-related injuries (e.g. migrants, males, middle-aged employees) and to the blue-collar jobs which seem to record the higher incidence of work-related injuries and diseases.
The aim of Drakopoulou Dodd (2011) was to develop an initial, tentative conceptualization of health and safety at work, within the family firm context. Special attention has been paid to the family firm owner-manager. The significance of family firms within even the most modern industrial environment was established, and their special importance within the SME grouping underlined. A simple model of occupational stress and health was presented, and utilised as the framework for this analysis. Family firm data and theory was utilised wherever possible, supplemented by material relating to the self-employed, or the entrepreneur, where necessary. Figure 2 illustrates the findings of this review, within the frame of the organizational stress model.

With regard to stressors, role conflict appears to be low for entrepreneurs, whilst role ambiguity may be higher than for comparable managers. However, family firm entrepreneurs, by contrast, appear very clear as to their roles, in spite of the multi-function nature of their decision authority. The farming sector, symbolic "home" of the family firm, showed that psychological demands on attention and time can be correlated with greater chance of workplace injury, and that economic anxieties represented a special danger. The high task-control and decision autonomy of the self-employed reduce the incidence of psychosocial stressors related to the lack of such freedom in the workplace. Indeed, family-firm member-employees reported greater autonomy, and the challenge of higher performance demands, as well as greater job satisfaction which these created. Arguments from evolutionary psychology were utilised to propose that the family firm represents an environment to which we are particularly suited, or "fit". The "stewardship" culture of family firms inhibits generates a comforting continuity, whilst entrepreneurial culture promotes an engaging and challenging flexibility. Thus the stressors generated by organizational change are less likely to be experienced in a negative fashion within the family firm environment, especially given the family's control over decision relating to the change process. Although entrepreneurs work very long hours, this does not seem to impact negatively upon their health, rather acting as an accumulation of rich experiences which generates overall wellbeing. It is within the sphere of interrole relationships that family firm specificities are the most pronounced. The dramatic conflict which is the dark side of family firms is generally related to some form of family emotional crisis. Outcomes of such conflict can be violent, even fatal. However, there appears to be no overall greater tendency for family-work interrole conflict within the family firms, and, indeed, positive spill-overs between environments have been argued to create virtuous cycles of support, positive affect, and resilience. In terms of individual differences, the genetic lottery seems to generate more divergence between family-firm leading kin groups than is the case where choice drives managerial selection. Coping strategies of special relevance for family-owned businesses include social support, for which family firms are especially well-suited, and detachment, for which they are not. Exercise appears another very suitable coping strategy for family firm managers. Family firm owner-managers can be argued to exhibit heightened exposure to accidents and fatality due to a variety of causes such as industrial sector, longer working hours, greater age, and higher self-selected exposure to risk. Whilst entrepreneurs in general report worse physiological health than other control groups, this is not so for psychological health. However, one study has shown that the greater the strain experienced by the entrepreneur (measured as occurrence of psychosomatic complaints), the higher the chances of long-term survival for their business (Rausch et al 2007). Perhaps some trade-off between personal and business health is accepted by the entrepreneur, particularly if kin-altruism is motivating such sacrifice within the context of the family firm. The study has indicated that the family firm is a rather idiosyncratic workplace, from an OHS perspective. It exhibits special dangers such as heightened risk of accident and poor physiological health and special benefits such as higher task control, social support, evolutionary-fitness, and a richer accumulation of experiences. Given these family firm-specific elements, and the overall importance of family firms to the wider environment, it seems clear that much further research in this area is merited.
The study by Drakopoulos and Theodossiou (2011) indicates that a particular type of regulation is necessary for an economy to attain efficient and equitable levels of OSH. The basis of this type of regulation is the actual probability of occupational risk. In this sense, the regulatory activity should take into account the objective probability of accident or illness at work for the particular occupation, and intervene by setting appropriate health and safety standards. Monitoring and enforcing these standards (by the imposition of financial penalties or prosecution to non-compliers) seems to be necessary given the observed tendency of workers' job risk underestimation.
The research by Sloane et al. (2011) demonstrates that evidence based purely on currently employed workers in the EWCS may be biased since it ignores an important 'healthy worker' selection effect that may operate. This is particularly acute for analysis of older workers. The selection adjusted findings have important implications for attempts in Europe and elsewhere to retain older workers in employment to offset the effects of population ageing. Workers who remain in employment are typically those in better health; facilitating employment for those who may otherwise may choose not to work is likely to be associated with deterioration in work-related health outcomes, with potential costs both to individuals and employers, as well as potentially the state in providing health services. This highlights the importance of finding ways to (further) reduce the risks associated with work, as the analysis confirms the vital role of physical, ergonomic and psychosocial exposures.
Economou and Theodossiou (2011a) investigates the effect of recessions as mirrored in the national unemployment rates on fatal and non-fatal work accidents, disaggregating into nine industrial sectors, in a panel sample of 13 European Union countries, for the period 1980-2006. The results suggest that the effect of recessions on the incidence of work accidents are more complicated than the patterns detailed in the literature. For a more accurate picture of the effect of unemployment upon work injuries, the decomposition of unemployment rates into temporary and permanent effects is crucial, since they exert a differential effect upon work-related accident rates. It seems that during the later stages of an economic downturn (or similarly, at the initial stages of an economic downturn) work-related injury rates tend to increase. Therefore, public labour market measures that aim to enhance occupational health and safety (OHS) should pay particular attention to these specific phases of the economic cycles. For example, such measures could be the stricter monitoring of the OHS policy implementation during economic downturns and at the initial stages of economic upturns; the provision of financial incentives to employers during economic recessions to undertake OHS investments, and; the dissemination of information to employees regarding safety measures and individual protection at the workplace. Furthermore, the results imply that cross panel correlation is important. This means that the European Commission directives and legislation, as well as OHS related labour market policies at the European level affect working conditions across member countries. All in all, in view of the evidence provided by this study, policy at the national and the European Union level which aims to improve working conditions and the decrease work accidents should be designed in combination with the macroeconomic policies aimed at evening out the macroeconomic fluctuations.
Economou and Theodossiou (2011b) investigate the effect of unionisation on fatal and non-fatal work accidents after controlling for the country GDP, using a panel sample of 10 European Union countries, for the period 1982-2006. The study takes into account the time persistence in work injuries and the endogenous nature of the work injuries union density relationship, by using GMM regression models. After controlling for endogeneity, both fatal and non-fatal work injuries decrease as union density increases. The empirical findings indicate that endogeneity is a shortcoming that affects the union density-work accidents relationship and should be taken into account in relevant studies. The paper highlights significant policy implications regarding the reduction of work-place injuries. Overall, the results imply that union density is conducive to reducing work place injuries. Union power seems to be an important determinant for the success of unions in occupational health and safety negotiations. Increased membership improves the ability of the unions to be effective in achieving improvements on occupational health and safety and hence the improvement of working conditions. All in all, it seems that unions act towards the improvement of workplace safety and policy makers should facilitate this effort and encourage union activities, under the scope of the improvement of working conditions.
Ferrer-i-Carbonell and Theodossiou (2012) investigate the preferences of the employers and employees regarding aspects of occupational safety and health (OSH). The provision of OSH is determined by the interplay of preferences of workers and employers. Hence, the need for a though understanding of the determination of the preferences of both workers and employers on OSH is of critical importance in order to identify areas of required action and to set priorities for policy initiatives on improving health and safety at work. The need to focus on health and safety is important since the interplay of preferences of workers and employers need to be completed by OSH strategies designed to address the consequences of a continuous adaptation.
Bender and Theodossiou (2011) Further work investigated potential pathways in which this linkage between performance pay and health might lay. While performance pay is found to increase work time, it does not seem to be correlated with a limited number of potentially healthy leisure activities, except for increases in drinking and restaurant meals. A more consistent pattern was found with a series of measures of stress, where duration models found that increases in the time spent in performance pay increase the hazard of five measures of stress. Performance pay can generate a variety of efficient labor market outcomes. However, the findings here are firmly in the camp of a potential unintended consequence of performance pay. Like other research that finds that performance pay can lead to workers, for example, focusing on quantity rather than quantity or overusing physical capital, long term exposure to performance pay is related to worse health, suggesting that firms may face increased health insurance or workers compensation costs (e.g. Freeman and Kleiner, 2005). Perhaps to mitigate these increases in costs, future research, using more detailed health data, should focus on identifying the pathways that performance pay can affect health. It may well be through increased stress, as suggested in the results here, but this should be more explicitly examined with more detailed data.

Main dissemination activities and exploitation of results
The project results are presented in meetings, conferences, and symposia; press releases and project web pages, communications to relevant mailing lists on the internet; newsletters are regularly posted to a mailing list of end-users established during the early meetings of the project providing information on the progress of the project and summarising and highlighting major reports and published papers; and a final conference (January 2012) and a reports with recommendations. Some of the results produced are written in form suitable for publication in the mainstream scientific literature, for example, peer-reviewed journals.

The results of the project were also presented in the Final Conference. They were available to a wide range of end-users and stakeholders such, national policy makers and regulatory authorities. One of the major functions of the final conference was highlight the findings of the project. Representatives from the national governments of the partners involved and European Union officials, as well as other interested parties including human resource managers, representatives of employer associations and trade unions, were invited to the final conference. Particular emphasis was given to secure the participation of employers (or their representatives) of SMEs. The results are of direct interest to those partner institutes involved in the implementation of European and national programmes on health and safety at work initiatives. The results produced provide useful information for other projects involved with similar issues in many European economies and other studies aiming at enhancing the health and safety at work. The database and the CIS outcomes of the project are a useful resource for future research on health and safety at work and they are available for consultation with free access through the Internet.

The co-ordinator maintains a web page providing information to a wider audience, including the general public, the research and health and safety at work management communities. Regular targeted press releases highlighted the findings of the research and the activities of the Project. The project adopted a comprehensive dissemination strategy aiming at targeting all the potential end-users including:

Reports to the EU
Publications in international peer-reviewed journals.
Institutional reports and working papers
Presentations of results in meetings, conferences, and symposia where end users and stakeholders will be invited
Press releases and newsletters to inform the general public and the end users and stakeholders
Project web page with open access to the general public and the end users and stakeholders
Project CD-ROMs
A final conference where end users and stakeholders had active participation

All data, tools and concepts developed by this Projectare considered public goods and are made available through the HEALTHatWORK web portal. Members of HEALTHatWORK have archived data and make tools available beyond the duration of this project. Public participation is mainly through the HEALTHatWORK web portal where interested persons can review the project, attach their own web sites if relevant to HEALTHatWORK, or discuss issues with experts in the relevant forum. Public awareness is raised through traditional means such as press releases demonstrating success stories and guiding lay persons to the HEALTHatWORK web portal.

Major tools of the dissemination are:

i) Direct contacst
A list of contacts is prepared from the beginning of the project and regularly updated. Policy makers (National government and EU officials or other interested officials are identified as well as representatives from the employer's (with particular reference to SMEs) associations and trade unions.

ii) The Web Site:
The web site is a central tool for communication. This web site acts an information platform on health and safety at work practices and indicators. It includes the results of the project reports. It allows consulting the most relevant publications on health and safety at work

iii) Press releases
Press releases and Newsletters are published at regular intervals. They are distributed to the research community and other EC project researchers.

iv) Final Conference
The final results were disseminated at the final conference where policy makers, national governments and EU officials, scientific communities on issues of health and safety at work, human resource managers, representatives of employer associations (including SMEs) and trade unions and other interested end users were invited.

List of Websites:

Grant Agreement number: 200716
Project acronym: HEALTHatWORK
Project title:
An Inquiry into the Health and Safety at Work; a European Union Perspective
Funding Scheme: Coordination and Support Action (Supporting)
Period covered: from 28 August 2008 to 28 February 2012

Name of the scientific representative of the project's co-ordinator , Ioannis Theodossiou
Title and Organisation: Professor, University of Aberdeen Business School, University of Aberdeen
Tel: +44 1224 272183
Fax: +44 1224 272181
Project website address: http/