Final Report Summary - CEWH (The Causal Effect of Work on Health)
BACKGROUND
The nature of work has changed dramatically in developed countries during the past decades, which implies new risks and challenges for workers. Trends like globalization and increasing world competition force firms to increase production efficiency in order to minimize costs. As a result, many workers complain about work stress and working overtime hours. Other important changes include the continuously changing technology, the increased female labor market participation rate, and an increased need for flexibility. This requires workers to adapt to less physically and more mentally intensive tasks, and to find a new balance between work and life while working flexible hours. The question is: how has this affected the health of workers?
SOCIETAL RELEVANCE OF EMPIRICAL STUDIES OF THE WORK-HEALTH RELATIONSHIP
Being in good health is an important asset for individuals to live a happy life. Firms also care about workers’ health as it ensures a continuous production process. But in addition to these private gains from having a good health, work-related sickness is a burden to society due to the high health care costs. Since 1995 the total costs of work-related illness in the Netherlands have increased with 56 percent to 1768 Euro per worker in 2001. This amounts to 3 percent of the gross national product (GNP). Because work-related sickness is very costly both in human and in economic terms, the European Union launched the Strategy on Health and Safety at Work (2007-2012), aiming to reduce the total incidence of work-related illness by 25 percent in 2012. For this type of initiatives to succeed it is essential to know how working affects health, and which conditions are causing poor health and which working conditions can improve health. From a theoretical perspective the effects of working on health are ambiguous. Working could benefit health due to a social status effect, the daily routine, and the social interactions. However, working could adversely affect health if the job is very demanding, if it limits time for health-improving leisure activities (e.g. exercising, home cooking), or if the labor earnings increase risky health behavior (e.g. drinking, smoking). Since the theoretical predictions are ambiguous and also the size of a potential effect is unknown, empirical studies are needed to identify the causal relationship of working on health.
PREVIOUS WORK
The main aim of this project was to investigate the causal effects of working – and different dimensions of work – on health. Credible estimates of the effect of work on health are very important because working is an important daily activity for many people. In addition, learning how work affects health is essential in the light of the current political debate in several EU countries on changing the retirement age.
Disciplines like epidemiology and occupational health have a long tradition of studying the association between work and health. For example, the literature in these fields has shown that hazardous work and difficulties in balancing work and life are correlated with poor health. More generally, work which is perceived to be unsatisfactory is deleterious to health. Also, social status seems to be related to health. People with a good job seem to have better health. Furthermore, individuals start to feel better when they are no longer unemployed. Later, economists also became interested in the effects of working on health as they became more aware that economic decision making might have important consequences on (non-economic) life outcomes. An important ambition in the economic literature is the aim to identify causal effects, rather than presenting mere correlations between work and health. The causal relationship between work and health is difficult to establish as people decide about work and health simultaneously. The main difficulty is that, on average, people in poor health work less, but it is not clear if their poor health is caused by non-employment, or if these people would have been in worse health even if they were working. The majority of the studies in this economic literature has focussed on the causal effect of health on labor market outcomes, showing that a health deterioration leads to a significant reduction in the probability of working and an increase in the probability of being unemployed or out of the labor force. However, the reverse causal relationship – from work on future health – is not yet well established. Some studies have focussed on recessions or job loss due to plant closure as an unanticipated shock to work behavior to identify the health effects this had. These studies have shown that job displacement leads to unhappines and mental problems and is found to increase mortality, even 20 years later. The seemingly contradictory finding that recessions are good for your health, can apparently be explained by the fact that the negative health consequences of job loss for those affected are compensated by better health of others who can now afford to buy cheap health care. However, work is characterized by many other aspects than simply job loss, e.g. the number of hours worked or the quality of the job.
Although these dimensions are important aspects of work, the health effects of these other dimensions of work are not yet known.
OBJECTIVES OF THE PROJECT AND WORK CARRIED OUT
The main objectives of this project were as follows.
First, it aimed to provide a clear and unambiguous identification of the causal effect of work on health, by exploiting information from natural experiments applying the state-of-the-art regression Discontinuity (RD) methodology. Detailed objective health information – administrative data on doctor and hospital visits, as well as mortality information – has been used to achieve this, enabling in depth investigation of the differential effect on mental and physical health. This distinction is particularly relevant since the number of individuals suffering from mental problems has been increasing substantially over the years. Furthermore, it enabled the identification of both short-run and long-run health effects.
Secondly, this project focused on various dimensions of work to answer the question of how work affects future health. Aspects that were taken into account included: (i) the extensive margin of work (i.e. the decision of working versus nonemployment), (ii) the intensive margin of work (i.e. the decision to work many versus few working hours per week), and (iii) the quality of the job (i.e. does the job fit the workers qualifications, interests and preferences). Examining various angles of work is important because it provides both a deeper and broader understanding of the effect of work on health.
More specifically, the following steps were taken to achieve the project's objectives.
After a refinement of the literature study three subprojects were performed, each one following a similar sequence of data preparation, models estimation and interpretation of results. The subprojects dealt with the causal effects of working on health, the causal effects of working hours on health and the causal effect of job quality on health.
Progress was in accordance with the original plan, except for the suspension at the beginning due to the maternity leave of the fellow. The researcher performed the three studies as introduced above and produced three papers on health effects of containing moral hazards, prenatal testosterone and the earnings of men and women and the relationship between happiness and employment. As part of the training/knowledge transfer activities she also participated in international conferences and presented papers there, wrote policy reports, magazine papers targeting practitioners and created blogs. Besides the maternity leave no major deviations occurred. Objectives were achieved and resources were used according to plan.
RESULTS, CONCLUSIONS AND POTENTIAL IMPACT
Main findings from the first study were as follows.
An unexpected shock to welfare participation (and employment participation as such) may have severe adverse health effects for individuals. In particular, women are more likely to be hospitalized immediately after the shock, and even have higher mortality rates about 15 years later. This important finding implies that a policy reform not only entails direct costs and benefits for the government, but that there are also long-term (health) costs that should be taken in to account.
Main findings from the second study were as follows.
Biological factors explain an important part of the gender wage gap. This makes an important contribution to the literature on gender wage gaps and discrimination, as before the degree of discrimination was over-estimated by not taking into account biological differences between men and women.
Main findings from the third study were as follows.
Although quite a bit is known about the relationship between employment and mental health already, the project contributed to the body of knowledge by showing that there is a lot of heterogeneity in this relationship. While some individuals are incredibly unhappy while unemployed, others can cope with unemployment much better. Since this has a major impact on later life employment outcomes, this heterogeneity is important and should deserve more attention from policy makers.
The conclusions bear relevance for policy makers and health care providers around the world. The researcher and the group where she is employed already had good working relationships with relevant government ministries in the Netherlands and in several other countries, but also with international organizations like the World Bank, IMF and Unicef. These target communities have been addressed through non-academic publications, conference presentations and dedicated talks. Interest in the results and in subsequent research from this community is high. Health care costs constitute a substantial part of a government's budget in many countries around the world and they are expected to grow in the coming years, while many related questions still deserve academic scrutiny. The project leveraged the opportunity for the fellow to establish herself as a though leader in this dynamic area of study. Partly as recognition for this success of the fellow received tenure at the end of the project and became an associate professor at Erasmus University.
The nature of work has changed dramatically in developed countries during the past decades, which implies new risks and challenges for workers. Trends like globalization and increasing world competition force firms to increase production efficiency in order to minimize costs. As a result, many workers complain about work stress and working overtime hours. Other important changes include the continuously changing technology, the increased female labor market participation rate, and an increased need for flexibility. This requires workers to adapt to less physically and more mentally intensive tasks, and to find a new balance between work and life while working flexible hours. The question is: how has this affected the health of workers?
SOCIETAL RELEVANCE OF EMPIRICAL STUDIES OF THE WORK-HEALTH RELATIONSHIP
Being in good health is an important asset for individuals to live a happy life. Firms also care about workers’ health as it ensures a continuous production process. But in addition to these private gains from having a good health, work-related sickness is a burden to society due to the high health care costs. Since 1995 the total costs of work-related illness in the Netherlands have increased with 56 percent to 1768 Euro per worker in 2001. This amounts to 3 percent of the gross national product (GNP). Because work-related sickness is very costly both in human and in economic terms, the European Union launched the Strategy on Health and Safety at Work (2007-2012), aiming to reduce the total incidence of work-related illness by 25 percent in 2012. For this type of initiatives to succeed it is essential to know how working affects health, and which conditions are causing poor health and which working conditions can improve health. From a theoretical perspective the effects of working on health are ambiguous. Working could benefit health due to a social status effect, the daily routine, and the social interactions. However, working could adversely affect health if the job is very demanding, if it limits time for health-improving leisure activities (e.g. exercising, home cooking), or if the labor earnings increase risky health behavior (e.g. drinking, smoking). Since the theoretical predictions are ambiguous and also the size of a potential effect is unknown, empirical studies are needed to identify the causal relationship of working on health.
PREVIOUS WORK
The main aim of this project was to investigate the causal effects of working – and different dimensions of work – on health. Credible estimates of the effect of work on health are very important because working is an important daily activity for many people. In addition, learning how work affects health is essential in the light of the current political debate in several EU countries on changing the retirement age.
Disciplines like epidemiology and occupational health have a long tradition of studying the association between work and health. For example, the literature in these fields has shown that hazardous work and difficulties in balancing work and life are correlated with poor health. More generally, work which is perceived to be unsatisfactory is deleterious to health. Also, social status seems to be related to health. People with a good job seem to have better health. Furthermore, individuals start to feel better when they are no longer unemployed. Later, economists also became interested in the effects of working on health as they became more aware that economic decision making might have important consequences on (non-economic) life outcomes. An important ambition in the economic literature is the aim to identify causal effects, rather than presenting mere correlations between work and health. The causal relationship between work and health is difficult to establish as people decide about work and health simultaneously. The main difficulty is that, on average, people in poor health work less, but it is not clear if their poor health is caused by non-employment, or if these people would have been in worse health even if they were working. The majority of the studies in this economic literature has focussed on the causal effect of health on labor market outcomes, showing that a health deterioration leads to a significant reduction in the probability of working and an increase in the probability of being unemployed or out of the labor force. However, the reverse causal relationship – from work on future health – is not yet well established. Some studies have focussed on recessions or job loss due to plant closure as an unanticipated shock to work behavior to identify the health effects this had. These studies have shown that job displacement leads to unhappines and mental problems and is found to increase mortality, even 20 years later. The seemingly contradictory finding that recessions are good for your health, can apparently be explained by the fact that the negative health consequences of job loss for those affected are compensated by better health of others who can now afford to buy cheap health care. However, work is characterized by many other aspects than simply job loss, e.g. the number of hours worked or the quality of the job.
Although these dimensions are important aspects of work, the health effects of these other dimensions of work are not yet known.
OBJECTIVES OF THE PROJECT AND WORK CARRIED OUT
The main objectives of this project were as follows.
First, it aimed to provide a clear and unambiguous identification of the causal effect of work on health, by exploiting information from natural experiments applying the state-of-the-art regression Discontinuity (RD) methodology. Detailed objective health information – administrative data on doctor and hospital visits, as well as mortality information – has been used to achieve this, enabling in depth investigation of the differential effect on mental and physical health. This distinction is particularly relevant since the number of individuals suffering from mental problems has been increasing substantially over the years. Furthermore, it enabled the identification of both short-run and long-run health effects.
Secondly, this project focused on various dimensions of work to answer the question of how work affects future health. Aspects that were taken into account included: (i) the extensive margin of work (i.e. the decision of working versus nonemployment), (ii) the intensive margin of work (i.e. the decision to work many versus few working hours per week), and (iii) the quality of the job (i.e. does the job fit the workers qualifications, interests and preferences). Examining various angles of work is important because it provides both a deeper and broader understanding of the effect of work on health.
More specifically, the following steps were taken to achieve the project's objectives.
After a refinement of the literature study three subprojects were performed, each one following a similar sequence of data preparation, models estimation and interpretation of results. The subprojects dealt with the causal effects of working on health, the causal effects of working hours on health and the causal effect of job quality on health.
Progress was in accordance with the original plan, except for the suspension at the beginning due to the maternity leave of the fellow. The researcher performed the three studies as introduced above and produced three papers on health effects of containing moral hazards, prenatal testosterone and the earnings of men and women and the relationship between happiness and employment. As part of the training/knowledge transfer activities she also participated in international conferences and presented papers there, wrote policy reports, magazine papers targeting practitioners and created blogs. Besides the maternity leave no major deviations occurred. Objectives were achieved and resources were used according to plan.
RESULTS, CONCLUSIONS AND POTENTIAL IMPACT
Main findings from the first study were as follows.
An unexpected shock to welfare participation (and employment participation as such) may have severe adverse health effects for individuals. In particular, women are more likely to be hospitalized immediately after the shock, and even have higher mortality rates about 15 years later. This important finding implies that a policy reform not only entails direct costs and benefits for the government, but that there are also long-term (health) costs that should be taken in to account.
Main findings from the second study were as follows.
Biological factors explain an important part of the gender wage gap. This makes an important contribution to the literature on gender wage gaps and discrimination, as before the degree of discrimination was over-estimated by not taking into account biological differences between men and women.
Main findings from the third study were as follows.
Although quite a bit is known about the relationship between employment and mental health already, the project contributed to the body of knowledge by showing that there is a lot of heterogeneity in this relationship. While some individuals are incredibly unhappy while unemployed, others can cope with unemployment much better. Since this has a major impact on later life employment outcomes, this heterogeneity is important and should deserve more attention from policy makers.
The conclusions bear relevance for policy makers and health care providers around the world. The researcher and the group where she is employed already had good working relationships with relevant government ministries in the Netherlands and in several other countries, but also with international organizations like the World Bank, IMF and Unicef. These target communities have been addressed through non-academic publications, conference presentations and dedicated talks. Interest in the results and in subsequent research from this community is high. Health care costs constitute a substantial part of a government's budget in many countries around the world and they are expected to grow in the coming years, while many related questions still deserve academic scrutiny. The project leveraged the opportunity for the fellow to establish herself as a though leader in this dynamic area of study. Partly as recognition for this success of the fellow received tenure at the end of the project and became an associate professor at Erasmus University.