Unemployment among young people and mental health: A systematic review

Abstract

Aim: The aim of this systematic review is to obtain a better understanding of the association between unemployment among young people and mental health. Methods: After screening the title and abstract of 794 articles drawn from four electronic databases, 52 articles remained for full-text reading. Of these, 20 studies met the inclusion criteria and were assessed on methodological quality. All steps were performed independently by two reviewers. Finally, a total of 17 articles were included in the systematic review. Results: Analysis of cross-sectional studies (N = 5) showed an association between unemployment among young people and mental health. An effect of unemployment on mental health was found when considering cohort studies (N = 12) that did not control for confounders (7/7). When controlling for confounders except mental health at baseline, this effect decreased in most studies leading to mixed results, although the majority (6/8) still found an effect. However, when taking mental health at baseline into account as one of the confounders, only a minority of studies (3/8) found a significant effect of unemployment on mental health. Conclusions: This systematic review showed an association between unemployment among young people and mental health. However, whether there is a causal relationship is less clear. More evidence from, for example, natural experiments and longitudinal studies that control for confounding variables, especially mental health at baseline, is required to better understand the association and potential causation between unemployment among young people and mental health.

Keywords

Unemployment youth young people mental health systematic review depression anxiety

Background

The life-period of ‘emerging adulthood’ is characterized by high instability in the areas of education, work, love and habitation [1]. This instability in work is also reflected by the percentage of unemployed 15- to 29-year-olds, which was about 15% in the European Union in 2016 [2]. Unemployment in this age group is often referred to as ‘unemployment among young people’ [3] and has been described as having serious consequences for society and for the individual. For instance, previous research has shown increased financial costs for society due to unemployment benefits, underuse of production potential [4] and foregone tax payments [5]. On the individual level, several studies have suggested that unemployed young individuals are more likely to have poor physical health [6–7], increased risk of cardiovascular disease [8], engage more frequently in criminal behaviour [9] and express poorer lifestyle behaviours such as smoking [10], increased alcohol consumption and substance abuse [9, 11]. Furthermore, unemployment in this age group has been associated with higher mortality rates due to suicide [9, 12], alcohol-related mortality [13], lower income and poorer quality of employment later in life [14] and is related to a wide spectrum of mental health problems such as increased stress [4] and psychosomatic symptoms [15], decreased mental wellbeing [16], lower life satisfaction [17], increased risk of depression [18], anxiety [19], other psychiatric disorders [20] and decreased mental health later in life [21].

Despite this growing body of literature that focuses on the relationship between unemployment among young people and mental health, to the best of our knowledge no systematic review has been conducted to summarize and interpret these results. Previous systematic reviews and meta-analyses have focused on age-mixed populations, with three reviews published in 1999, 2005, and 2009 finding evidence for an association between unemployment and mental health and supporting a possible causal relationship [22 –24].

Murphy and Athanasou, who included 16 longitudinal studies in their meta-analysis, concluded that unemployment had negative effects on mental health. Moving from unemployment to employment was associated with an improvement in mental health with a weighted average effect size of Cohen’s d = 0.54 (based on seven studies). Alternatively, the effect of moving from employment to unemployment was less clear due to the small number of studies included (five studies with 616 participants in total) and showed an average effect size of Cohen’s d = 0.36. The results of this meta-analysis were based on studies that covered both youth and adulthood and only two of the 16 studies included in the meta-analysis controlled for mental health at baseline [24].

In another meta-analysis, McKee-Ryan et al. examined the effects of unemployment on wellbeing across 104 empirical studies. Although the majority of these included studies concerned adult unemployment, no results for specific age groups were reported. Based on 52 cross-sectional studies included in the meta-analysis, it was found that unemployed individuals had lower mental health compared to employed individuals, with a weighted mean effect size of Cohen’s d = -0.57. In addition, findings from 15 longitudinal studies showed that reemployment improved mental health (d = -0.89) whereas eight longitudinal studies indicated that becoming unemployed decreased mental health (d = -0.38). Based on the performed analyses, it was concluded that there is strong support for a causal relationship between unemployment and mental health [23]. Unfortunately, no information was given on how confounders, including mental health at baseline, were controlled for.

Thirdly, Paul and Moser compared the mental health of unemployed and employed persons using the results of 237 cross-sectional and 87 longitudinal studies. Results from the cross-sectional studies showed that unemployed individuals had higher levels of distress and lower wellbeing than employed individuals, with an overall average effect size of Cohen’s d = 0.54. Longitudinal studies indicated a significant increase in distress for individuals that became unemployed (d = 0.19), whereas a decrease in distress was observed after reemployment (d = -0.35). Notably, for youths that became unemployed right after school, no significant increase in distress was observed. From this, it was concluded that unemployment has a negative effect on overall mental health with Cohen’s d = 0.5122 [22]. Unfortunately, no separate analysis was provided for studies that controlled for confounding variables including mental health at baseline and results were not reported for the different age groups.

Within the literature available on the relationship between unemployment and mental health, one of the key concerns is the distinction between selection and exposure effects of unemployment. An exposure effect would mean mental health is the result of unemployment and would thus propose a causal relationship, whereas a selection effect would assume unemployment is the result of mental ill health and the association between unemployment and mental health occurs because individuals with lower mental health have more difficulties in finding a job and have a higher chance to become unemployed [22]. The problem of causality is impossible to solve using evidence from cross-sectional studies, because these data do not contain any information regarding the direction of the association [26]. From a research point of view, an experimental design where individuals are randomized into either the unemployment or the employment condition would provide the best evidence for a possible causal relationship between unemployment and mental health. Obviously this is not possible due to ethical reasons; however, cohort studies that examine the relationship between unemployment and mental health are widely available. Although ultimately experimental designs will be necessary to find evidence for a causal relationship between unemployment and mental health, these longitudinal cohort studies that control for confounders including mental health at baseline could provide support for a causal hypothesis [27].

Aim

The aim of this systematic review is to better understand the association between unemployment among young people and mental health. The following research question will be answered: what effect does unemployment between the ages of 16 and 30 have on the mental health of individuals in high-income countries?

Methods

Terms and definitions

To operationalize mental health the World Health Organization (WHO) definition was used, which explains mental health as ‘a state of wellbeing in which the individual realizes his or her own abilities, can cope with normal stresses of life, can work productively and fruitfully, and is able to make a contribution to his or her community’ and is ‘not just the absence of mental disorders’ [28]. In addition, anxiety and depression [29] were used as these are the two most common mental health disorders among youth. Furthermore, to have a broad scope for studies, mental ill health was broadly operationalized by decreased mental wellbeing, feelings of stress or distress, feelings of depression or anxiety, or diagnosed depression or anxiety disorders.

‘Unemployment among young people’ was operationalized as a self-reported or registered period of not working in paid employment between the ages of 16 and 30 years.

Because the economic situation of a country can influence the magnitude of consequences of unemployment, for example due to differences in unemployment protection systems and equality of income distribution [22], only high-income countries were compared in this systematic review. The World Bank classification system [30] was used to define high-income countries.

Search strategy

Three electronic databases (PubMed, Web of Science and PsycInfo) were used to systematically search for articles that assess the association between unemployment among young people and mental health (see Table I for the search strategies of PubMed, Web of Science, PsycInfo). Google Scholar was searched as a supplementary source using the search terms ‘youth unemployment’ and ‘mental health’. The first four pages of hits were compared with the articles found through the other electronic databases. Only unique articles were included. The search was limited to English language articles published in a peer-reviewed journal between 1 January 2000 and 29 January 2018. Additional articles were searched through reference lists of included articles.

Table I.

Search strategies for PubMed, Web of Science and PyscInfo (29 January 2018).

Search	Keywords	Limitations	Number of hits
PubMed
#1	Unemployment[MeSH]		6159
#2	Mental health[MeSH] OR mental disorders[MeSH] OR depression[MeSH] OR anxiety[MeSH]		1,213,839
#3	Adolescent[MeSH] OR young adult[MeSH]		2,125,602
#4	#1 AND #2 AND #3		642
#5	#4	Publication date 2000–	436
#6	#5	English language	413
Web of Science
#1	Unemployment		35,580
#2	Anxiety OR depress* OR stress OR “mental ill-health” OR “mental health” OR “mental disorder*”		1,760,615
#3	Adolescen* OR “young adult” OR youth		463,664
#4	#1 AND #2 AND #3		457
#5	#4	Publication date 2000–	383
#6	#5	English language	361
PsycInfo
#1	Unemployment[SubjectHeadings]		3897
#2	Mental Health[SubjectHeadings] OR Well Being[SubjectHeadings] OR Mental Disorders[SubjectHeadings] OR Behavior Disorders[SubjectHeadings] OR Anxiety[SubjectHeadings] OR Stress[SubjectHeadings]		868,130
#3	Adolescen* OR young OR youth OR “emerging adulthood”		431,193
#4	#1 AND #2 AND #3		201
#5	#4	Publication date 2000–	108
#6	#5	English language	101

Inclusion and exclusion criteria

Articles were included if they described the association of unemployment between the ages of 16 and 30 years and were performed in high-income countries. To assess whether the study was performed in a high-income country, the World Bank classification [30] was used. Articles were excluded when (a) they were not available in full text; (b) p values were not provided in the results; (c) the exposure measure was not defined as a period of unemployment between 16 and 30 years old; (d) the outcome measure did not consist of mental health problems (operationalized as low mental wellbeing, stress or distress, feelings of depression or anxiety, or diagnosed depression or anxiety disorders); or (e) mental health and unemployment were not assessed within the same individual.

Study selection and evaluation

Two reviewers (VB and KZY) first individually screened titles and abstracts for suitable articles and then discussed the results and solved disagreements by consensus. Articles eligible for full-text assessment were handled in the same manner and were first independently assessed based on the inclusion and exclusion criteria by the same reviewers, then discussed to reach consensus on which articles to include. After this, the methodological quality of the included articles was independently assessed by the same reviewers using the Effective Public Health Practice Project (EPHPP) quality assessment tool [31]. The EPHPP tool has shown to have construct and content validity [32] and is proven to be reliable [33]. After reaching consensus on the quality of the articles between the reviewers, studies with a weak global rating were excluded. The results of the quality assessment can be found in Table II. Data of included studies were extracted by the first author (VB) using a data extraction sheet in Excel. Information of the following data characteristics was extracted: authors, title, year, aim, study design, name of cohort (if applicable), follow-up time (if applicable), sample size, response rate, age range, sex distribution, study setting, definition of exposure measure, assessment of mental health at baseline, description of outcome measure, type of scientific analysis, measure of effect (mean with standard deviations or odds ratios with confidence intervals), p values, limitations and strengths and conclusions. Due to the heterogeneity of included studies, it was not possible to do a quantitative synthesis of the results, instead a narrative synthesis was done. The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [34].

Table II.

Quality assessment of the studies (N = 20) using the Effective Public Health Practice Project (EPHPP) quality assessment tool.³¹

Author (year)	Global rating
Aquilar-Palacio et al. (2015)	Moderate
Axelsson and Ejlertsson (2002)	Moderate
Bacikova-Sleckova et al. (2007)	Weak
Baggio et al. (2015)	Strong
Bjarnason and Sigurdardottir (2003)	Moderate
Breslin and Mustard (2003)	Strong
Crowe and Butterworth (2016)	Strong
Fergusson et al. (2001)	Strong
Fergusson et al. (2014)	Strong
Goldman-Mellor et al. (2016)	Strong
Hammarström and Janlert (2002)	Strong
Huegaerts et al. (2017)	Weak
McGee and Thompson (2015)	Moderate
Mossakowski (2011)	Strong
Reine et al. (2004)	Strong
Strandh et al. (2015)	Strong
Strandh et al. (2014)	Strong
Taris (2002)	Weak
Thern et al. (2017)	Strong
Virtanen et al. (2016)	Strong

Global rating is the overall judgment based on the assessment of six domains: strong = no weak ratings, moderate = one weak rating, weak = two or more weak ratings.

Results

Selection of studies

A flow chart describing the article selection process according to the PRISMA guidelines [34] can be found in Figure 1. The initial electronic search provided a total of 878 articles, of which 794 remained after removal of duplicates. Subsequent screening of titles and abstracts further reduced this number to 51 articles that remained for full-text assessment. Based on the reference lists in the selected articles, one more study was included at this point. A total of 52 articles were assessed for eligibility in full text, where after a total of 20 articles remained for quality assessment. During quality assessment, three more articles were excluded. Finally, a total of 17 studies met all criteria and were included in this systematic review. An overview of the distribution of different study characteristics among the included studies can be found in Table III.

Figure 1.

Flow chart of study selection according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Table III.

Characteristics of included studies.

Author (year)	Setting	Study design	Sample	Exposure	Outcome measures	Follow up	Controlled for confounding variables	Key findings
Aquilar-Palacio et al. (2015)³⁷	Spain	Cross-sectional	N = 3701Age: 16-24 years.58% men, 42% women	Self-reported unemployment	Mental health during the past few weeks measured with the General Health Questionnaire (GHQ-12)	No	Family social class, educational level, age	• Unemployed men (< 1 year) showed a higher risk of mental health problems than workers in 2012• No significant differences was found for unemployed men in 2006 or women in 2006 and 2011/12 compared to workers• When both time points were considered: unemployed men (⩾ 1 year) showed a higher risk of mental health problems than workers, but no significant differences were found for women, unemployed men (< 1 year), or men that never worked compared to workers.
Axelsson and Ejlertsson (2002)15	Sweden	Cross-sectional	N = 515Age: 20–25 years.46% men, 54% women	Registered unemployment ⩾ 3 months	Mental health based on items: tearfulness, dysphoria, sleeping disturbance, restlessness, general fatigue, irritability	No	Civil status, living situation, education level	Unemployed men and women showed more mental health problems than employed men/women.
Bjarnason and Sigurdardottir (2003)4	Denmark, Finland, Iceland, Norway, Scotland, Sweden	Cross-sectional	N = 7307Age 18–24 years.54% men, 46% women	Registered unemployment ⩾ 3 months 6 months prior to the study	Psychological distress assessed using Hopkins Symptoms Checklist for anxiety and depression	No	Duration of unemployment, perceived public/personal negative attitudes towards employment, support from parents, economic situation, education level, age, living situation	• Unemployed men and women (⩾ 3 months) that were still unemployed at the time of the study (after 6 months) showed significant higher levels of psychological distress than their counterparts in other labour market outcomes.• Less distress was associated with moving beyond unemployment to: a permanent job, a temporary job, returning to school, or leaving the labour market.
McGee and Thompson, (2015)35	US	Cross-sectional	N = 1525Age: 18–2559% men, 41% women	Self-reported unemployment	Depressive symptoms in previous 2 weeks assessed by the Patient Health Questionnaire (PHQ-8)	No	Disability status, smoking, BMI, health insurance status, sex, race, marital status, education	Unemployed individuals (<1 year and ⩾ 1 year) showed a higher risk of depression than employed individuals.
Goldman-Mellor et al. (2016)36	Great Britain	Cross-sectional/ Cohort	N = 2066.Age (M): 18 years.47% men, 53% women	Self-reported NEET status (not studying, working in paid employment or in apprenticeship training)	• Depression and generalized anxiety disorder over the previous 12 months assessed by Diagnostic Interview Schedule	No	Sex, family socioeconomic status, neighbourhood socioeconomic context, childhood IQ, childhood reading skills, childhood mental health problems	NEET status was significantly associated with major depressive episode and generalized anxiety disorder compared to non-NEET individuals.
Baggio et al. (2015)41	Switzerland	Cohort	N = 5990 Follow-up rate 87.2%. Age (M): 20 at baseline and 21.3 at follow-up.100% men	Self-reported NEET status (not studying, working in paid employment, or training) including participants currently looking for work and not looking for work.	• Mental health status in the previous four weeks assessed using the Short Form Health Survey.• Level of depression in the previous two weeks assessed by the WHO’s Major Depressive Inventory	15 months	Language, age, family income, parents’ education level	No significant effect was found for NEET status at baseline as predictor of mental health or depression at follow up.
Breslin and Mustard (2003)42	Canada	Cohort	N = 1868 Follow-up rate 88.9%. Age 18–30 years.54% men, 46% women	Registered unemployment	• Distress over the past month based on items: everything feels an effort, so sad nothing cheers you up, nervous, restless and fidgety, hopeless, worthless.• Presence of depression in the last 12 months assessed by the Composite International Diagnostic Interview	2 years	Gender, age, marital status, living situation, unusual life events, chronic medical conditions, outcome variable at time 1	No significant effect was found for 18–30 year olds of becoming unemployed and remaining unemployed after 2 years of distress or depression at follow up.
Crowe and Butterworth, (2016)18	Australia	Cohort	N = 2404 Response rate at follow-up 1: 83% and 2: 89%. Age: 20–24 years at baseline. 48% men, 52% women	Self-reported unemployment	Symptoms of depression in the prior 4 weeks assessed by the Goldberg Depression Scale	8 years	Age, gender, marital status, physical health, dependent children, education level	Unemployment and working part time but looking for full-time employment was significantly associated with depression at follow up.
Fergusson et al. (2001)51	New Zealand	Cohort	N = 1053.Age (M) 18 at baseline.Women 52%, men 48%	Self-reported unemployment during 16–18 years old and 18–21 years old	Depression in the past 2 years assessed by the Composite International Diagnostic Interview	2 years	Measures of family social background, measures of family functioning, measures of parental adjustment, individual characteristics, cognitive ability and school achievement, measures of behaviour and peer affiliations, history of psychiatric disorder and suicidal behaviour prior to age 16	No significant association was found for the effect of unemployment on depression at follow up.
Fergusson et al. (2014)44	New Zealand	Cohort	N = 1056. Follow-up rate 80%Age (M) : 18 years at baseline, follow up at 21, 25 and 30. Women 52%, men 48%	Self-reported unemployment during 16–18, 18–21, 21–25, and 25–30 years old	Depression and anxiety disorders in the past 12 months assessed by the Composite International Diagnostic Interview	12 years	Reversed causality and non-observed confoundingvariables	• A significant effect was found for duration of unemployment on major depression• No significant effect was found between duration of unemployment on anxiety disorder.
Hammarström and Janlert (2002)27	Sweden	Cohort	N = 1083 Follow-up rate 96% Age (M): 16 years at baseline, follow-up at 18, 21 and 30.52% men and 48% women	Self-reported unemployment	Psychological symptoms during the previous year assessed by a composite index of nervous symptoms and depressive symptoms	14 years	Health situation at 16, working class background	• Unemployment (> 0.5 years) between 16 and 21 years was significantly associated with increased risk for psychological symptoms at age 30 for men and women.• No significant effect was found for unemployment (>1.5 years) between 22 and 30 years on psychological symptoms at age 30.
Mossakowski (2011)38	US	Cohort	N = 6932 Follow-up rate 89.2%. Age: 14-22 at baseline, 27-35 at first follow-up and 29-37 at second follow up	Self-reported unemployment	Depressive symptoms assessed by the Epidemiological Studies Depression scale (CES-D)	15 years	Gender, age, race, marital status, currently a parent, currently employed/out of the labour force, parental education, parental occupational prestige, prior depressive symptoms, prior self-esteem	Unexpected unemployment at age 19–27 predicted higher levels of depression at age 27–35 and 29–37 compared to expecting to become employed and becoming employed at age 19–27. When controlling for confounding variables, the effect of unexpected unemployment on depression became non- significant.
Reine et al. (2004)39	Sweden	Cohort	N = 1083 Follow-up rate 96%. Age (M): 16 years at baseline and follow-up at age 18, 21 and 30.52% men and 48% women	Self-reported unemployment	Psychological symptoms during the previous year assessed by a composite index of nervous symptoms and depressive symptoms	14 years	Daily smoking, having children, financial situation, working class, control, unemployed relatives, work commitment	• Unemployment (⩾6 months) between age 16–21 was associated with poor psychological health at age 21 for men and women.• Unemployment (⩾6 months) between age 25–30 is associated with poor psychological health at age 30 in men. No significant effect was found for 30-year-old women.
Strandh et al. (2015)43	Sweden	Cohort	N = 1083 Follow-up rate 94%.Age: 18–21 at baseline and follow-up at age 21 and 42. Women: 48%, men: 52%	Self-reported open unemployment (actively seeking for work while not being in employment, education or training)	Psychological symptoms during the previous year assessed by a composite index of nervous symptoms and depressive symptoms	27 years	Gender, health situation at age 16, parental employment situation, social class and health problems at age 16, and living with both parents at age 16	Exposure to open youth unemployment age 18–21 was significantly associated with internalized mental health symptoms at age 21 and at age 43.
Strandh et al. (2014)52	Sweden	Cohort	N = 1083 Follow-up rate 94%. Age (M): 16 years at baseline and follow-up at age 18, 21, 30 and 42.Women: 48%, men: 52%	Self-reported unemployment	Psychological problems during the previous year assessed by the Psychological Problems Index (PPI)	27 years	Gender, parental social class at age 16, parental involuntary unemployment at age 16	• Cumulative unemployment of ⩾ 6 months between age 18–21 had a significant effect on psychological problems at age 21, 30 and 42.• Consecutive unemployment of ⩾ 6 months between age 21–30 had a significant effect on psychological problems at age 30, but no significant effect at age 42.• Exposure to unemployment between age 18–21 and 21–30 had a significant effect on psychological problems at age 30, but not at age 42.
Thern et al. (2017)40	Sweden	Cohort	N = 6410 (crisis cohort) with follow-up rate 89%.N = 8162 (non-crisis cohort) with follow-up rate 90%.Age 17–24 at baseline and 34–35 at follow up	Registered unemployment	Affective (mood) disorders and nervous-stress related disorders diagnosed using the International Classification of Disease on the national hospital discharge register data	17 years	Sex, age and country of birth, prior own mental diagnosis before exposure and assessment by the Labour Force Survey, parental socioeconomic index, parental education and parents’ mental health.	• Effect of unemployment at age 17–24 on nervous and stress-related disorders after 17 years was significant in non-crisis cohort but not in the crisis-cohort.• No significant effects were found for affective (depression, bipolar, anxiety) disorders.
Virtanen et al. (2016)19	Sweden	Cohort	N = 1001 (Cohort 65) with follow-up rate 94%. 52% men, 48% womenN = 686 (Cohort 73) with follow-up rate 86%. 50% men, 50% women.Age 21–25 at baseline and followed-up at age 39/43	Self-reported unemployment	Depression and anxiety were assessed by a composite measure of internalized symptoms	22 years (Cohort 65) and 18 years (Cohort 73)	Gender, parental socioeconomic status, mental health at baseline, recent unemployment at follow up	• For Cohort 65, unemployment (25–183 weeks) at age 21–25 had a significant effect on depression and anxiousness at age 43. There was no significant effect of unemployment (2–24 weeks) on depression and anxiousness at age 43.• For the Cohort 73, unemployment (1–15 weeks /16–126 weeks) had a significant effect on anxiousness and at age 39. No significant effect was found for unemployment on depression.

BMI: body mass index; NEET: Not in Employment, Education or Training; WHO: World Health Organization.

Cross-sectional studies

Four out of five cross-sectional studies reported mainly significant differences or increased risk of mental health problems for unemployed young people compared to employed young people. Both Axelsson and Ejlertsson and Bjarnason and Sigurdardottir found a significantly higher level of mental health problems among youth up to 24 years old that were registered as unemployed for at least 3 months compared to employed individuals [4, 15]. In addition, Bjarnason and Sigurdardottir found in their sample of youths from different European countries that this difference remained significant when comparing unemployment with other labour market outcomes (permanent job, temporary job, returning to school or leaving the labour market) and that moving beyond unemployment was associated with a reduction in mental health problems [4]. McGee and Thompson showed that 18- to 25-year-old unemployed Americans had up to three times higher odds of depression compared to employed individuals [35] and Goldman-Mellor et al. found that 18-year-olds Not in Education Employment or Training (NEET) in Great Britain had a higher risk for both major depressive and generalized anxiety disorder than non-NEETs [36]. Only one study found mixed results: Aguilar-Palacio et al. found a significant higher risk for mental health problems for young Spanish men (aged 16 to 24) that were unemployed compared to employed young men in 2012 (a recession period). However, they failed to find similar results in 2006 (no recession period). For women, no significant differences were found between employed and unemployed individuals either in 2006 or 2012 [37].

Cohort studies – not controlling for confounders

Of the 12 cohort studies, seven provided information on the effect of unemployment at a young age on mental health without controlling for confounders. All of these studies (seven out of seven) found mainly significant effects of youth unemployment on mental health.

Cohort studies – controlling for confounders except mental health at baseline

When controlling for confounding variables (see Table III for the confounders included in the different studies) but without considering mental health at baseline, the effect of unemployment at a young age on mental health was reduced in most cases compared to not controlling for confounders. Nevertheless, the majority of the studies (six out of eight studies) still found mainly significant effects of unemployment among young people on mental health.

For instance, unemployment at age 20 to 24 in Australia was significantly associated with depression at age 24 to 32 [18]. A similar effect was observed in the United States, where unexpected unemployment between age 19 and 27 was found to have a significant effect on depression at age 29 and 37 [38]. Additionally, high exposure (16 to 183 weeks) of self-reported unemployment at age 21 to 25 in Sweden was found to have a significant effect on depression and anxiousness at age 39 or 43 in times of economic boom and recession [19]. Age-related differences were observed in a study by Strandh et al., which showed that although unemployment of more than 6 months among Swedish youths between age 18 and 21 had a significant effect on psychological problems at age 21, 30 and 42, unemployment of more than 6 months at age 21 to 30, or being unemployed both at age 18 to 21 and at age 21 to 30, had a significant effect on psychological problems at age 30 but not at age 42 [21]. Similarly, age- and sex-related differences were observed by Reine et al., who showed that being unemployed for more than 6 months between age 16 to 21 was significantly associated with poor psychological health at age 21 for both Swedish men and women, whereas more than 6 months of unemployment between the age of 25 and 30 was significantly associated with poor psychological health for men at age 30, but not for women [39]. Finally, Thern et al. found that registered unemployment at age 17 to 24 for more than 6 months significant effected mental health at age 34/35. In addition, total unemployment (including short-term unemployment) affected nervous- and stress-related disorders at times of financial prosperity, but not at times of financial crisis in Sweden, but did not affect depression and anxiety disorders. However, because inpatient discharge data were used to assess these outcomes, only patients that needed inpatient care were captured in their analysis [40].

In contrast, in a cohort of Swiss men, no significant effect was found for NEET status at a mean age of 20 on mental health or depression 15 months later [41]. Furthermore, Breslin and Mustard found no significant effect of registered unemployment in Canada at age 18 to 30 on clinical depression or distress after 2 years of follow up [42].

Cohort studies – controlling for confounders including mental health at baseline

Eight studies reported results on the effect of unemployment among young people on mental health while controlling for confounders including mental health at baseline. The inclusion of this confounder reduced the effect size in all of these eight studies and resulted in mixed results: three out of eight studies showed mainly significant effects of unemployment at young age on mental health, one study found mixed effects, whereas four studies did not find significant effects.

Three studies showed significant effects when controlling for confounders including mental health at baseline: Strandh et al. showed that Swedish 18- to 21-year-old NEETs who were actively seeking for work had more internalized mental health problems at age 21 and 43 than employed individuals [43]. Thern et al. still found a significant effect of registered unemployment for more than 6 months at age 17 to 24 on mental health at age 34/35 when controlling for mental health at baseline [40]. In addition, Virtanen et al. still found a significant effect of high exposure (16 to 183 weeks) of self-reported unemployment at age 21 to 25 on depression and anxiousness at age 39 to 43 in times of economic boom, but no longer on depression in a period of recession [19].

One study found mixed effects: Hammarström and Janlert found that more than 6 months of unemployment between age 16 and 21 was significantly associated with increased risk of psychological symptoms at age 30 for both men and women. However, no significant effect was found for more than 18 months of unemployment between age 22 and 30 on psychological symptoms at age 30 in Sweden [27].

Four studies found no significant effect: Breslin and Mustard still did not find a significant effect for registered unemployment on depression or distress after two years of follow up when controlling for mental health at baseline [42]. Fergusson et al. found no effect in a New Zealand cohort study of unemployment between ages 16 to 21 on major depressive disorder at age 21 after controlling for confounding variables including mental health at baseline [9]. Similarly, in another study by Fergusson et al. using the same cohort, the significant effect of duration of unemployment between age 16 to 25 on major depression and anxiety disorder as a pooled result for the ages 18, 21, 25 and 30 disappeared when controlling for confounders including mental health at baseline [44]. Lastly, Mossakowski et al. no longer found a significant effect of unexpected unemployment between age 19 and 27 on depression at age 29 and 37 when controlling for mental health at baseline in addition to other confounding variables [38]. A summary of the findings of the cohort studies can be found in Table IV.

Table IV.

Summary of effects of unemployment among young people on mental health expressed in Beta coefficient (β) or OR found in cohort studies (N =12).

Author (year)	Not controlled for confounding factors	Controlled for confounding factors	Controlled for confounding factors including mental health at baseline
Baggio et al. (2015)41	N/A	Depression β 0.093 (NS)	N/A
	N/A	Mental health β -0.028 (NS)	N/A
Crowe and Butterworth (2016)18	Depression: OR 2.40 (1.74–3.32)⋆	Depression: OR 1.55 (1.06–2.25)⋆	N/A
Strandh et al. (2014)21	MH (exposure age 18–21): β 0.25 (0.09–0.41)⋆⋆⋆	MH (exposure age 18–21): β 0.21 (0.05-0.37)⋆⋆⋆	N/A
	MH (exposure age 21–30): β 0.32 (0.07–0.58)⋆	MH (exposure age 21–30): β 0.34 (0.09–0.59)⋆⋆	N/A
	MH (exposure age 18–21 and 21–30): β 0.99 (0.74–1.23)⋆⋆⋆	MH (exposure age 18–21 and 21–30): β 0.97 (0.72–1.21)⋆⋆⋆	N/A
Reine et al. (2004)39	MH at age 21 (men): OR 4.97 (2.79-8.86)⋆	MH at age 21 (men): OR 4.79 (2.24-10.23)⋆	N/A
	MH at age 30 (men): OR 2.64 (1.33–5.24)⋆	MH at age 30 (men): OR 2.29 (1.08–4.85)⋆	N/A
	MH at age 21 (women): OR 4.03 (2.25–7.19)⋆	MH at age 21 (women): OR 2.71 (1.24–5.88)⋆	N/A
	MH at age 30 (women): OR 1.48 (0.62–3.53) (NS)	MH at age 30 (women): OR 1.38 (0.55–3.44) (NS)	N/A
Breslin & Mustard (2003)42	N/A	Depression: OR 0.89 (0.3-2.8) (NS)	Depression: OR 0.83 (0.3–2.6) (NS)
	N/A	Distress β -0.002 (NS)	Distress β -0.052 (NS)
Fergusson et al. (2001)9	β 0.141⋆⋆	N/A	β 0.103 (NS)
Ferguson et al. (2014)53	Depression: OR 1.23 (1.12–1.35) ⋆⋆⋆	N/A	Depression: OR 1.17 (1.00–1.37) (NS)
	Anxiety: OR 1.15 (1.03–1.28)⋆⋆	N/A	Anxiety: OR 1.10 (0.93–1.29) (NS)
Hammarström and Janlert (2002)27	N/A	N/A	MH (exposure age 16–21, men):OR 2.6 (1.4-4.7)⋆
	N/A	N/A	MH (exposure age 16–21, women):OR 1.9 (1.0–3.5)⋆
	N/A	N/A	MH (exposure age 22–30, men):OR 1.5 (0.9–2.6) (NS)
	N/A	N/A	MH (exposure age 22–30, women):OR 1.5 (0.8–2.6) (NS)
Mossakowski et al. (2011)38	N/A	β 0.70⋆⋆	β 0.39 (NS)
Strandh et al. (2015)43	N/A	N/A	MH at age 21: OR 2.48 (1.57–3.60)⋆⋆⋆
	N/A	N/A	MH at age 43:OR 1.71 (1.20–2.43)⋆⋆
Thern et al. (2017)40	MH (exposure during economic recession): OR 3.08 (1.97–4.82)⋆	MH (exposure during economic recession): OR 3.27 (2.08–5.15)⋆	MH (exposure during economic recession): OR 2.70 (1.71–4.28)⋆
	MH (exposure during economic boom): OR 4.17 (2.52–6.90)⋆	MH (exposure during economic boom): OR 4.05 (2.44–6.72)⋆	MH (exposure during economic boom): OR 3.33 (2.00–5.57)⋆
Virtanen et al. (2016)19	Anxiousness (exposure during economic boom): OR 2.65 (1.81–3.88)⋆	Anxiousness (exposure during economic boom): OR 2.65 (1.80–3.92)⋆	Anxiousness (exposure during economic boom): OR 2.19 (1.46–3.30)⋆
	Depression (exposure during economic boom): OR 2.43 (1.68–3.50)⋆	Depression (exposure during economic boom): OR 2.37 (1.63–3.45)⋆	Depression (exposure during economic boom): OR 1.85 (1.25–2.74)⋆
	Anxiousness (exposure during economic recession): 2.25 (1.45–3.47)⋆	Anxiousness (exposure during economic recession): 2.33 (1.50–3.63)⋆	Anxiousness (exposure during economic recession): 2.13 (1.33–3.39)⋆
	Depression (exposure during economic recession): OR 1.46 (0.96–2.21) (NS)	Depression (exposure during economic recession): OR 1.55 (1.01–2.37)	Depression (exposure during economic recession): OR 1.38 (0.89–2.14) (NS)

⋆

= p <0.05; ⋆⋆ = p <0.01; ⋆⋆⋆ = p <0.001.

NS: not significant; NA: information not available; MH: mental health; OR: odds ratio.

Discussion

Discussion of findings

The aim of this study was to better understand the association between unemployment among young people and mental health. The results of this study show that unemployment at a young age is associated with mental health problems. Reviewing the results of cross-sectional studies, the findings were largely consistent in finding an association between unemployment among young people and mental health: almost all studies found evidence for a significant difference in the mental health of unemployed compared to employed young people.

However, to answer the research question ‘what effect does unemployment between the age of 16 and 30 have on the mental health of individuals in high-income countries?’, the results of the cohort studies should be considered. Results of cohort studies not controlling for confounding variables consistently showed an effect of unemployment on mental health: seven out of seven studies found a significant effect. However, controlling for confounding variables but not considering mental health at baseline already resulted in a reduced effect in most cases as well as mixed results: six out of eight studies showed a significant effect. Importantly, when controlling for confounding factors including mental health at baseline, the observed effects decreased in all cases and only a minority of the studies (three out of eight) showed a significant effect.

Interpretation of results

To our knowledge, this systematic review was the first to assess the association between unemployment and mental health specifically for young people. The finding of an association between unemployment and mental health is in line with previous meta-analyses and systematic reviews on unemployment and mental health among mixed-age groups [22 –24]. However, our findings do not replicate the results of these studies that support a causal relationship between unemployment and mental health. There are several reasons that might explain the difference between the conclusions of the current study and previous systematic reviews and meta-analyses.

First of all, our study shows that outcomes differ quite substantially depending on whether confounders are considered. As previous reviews did not differentiate studies based on the inclusion of confounders, it remained unclear how much of the relationship between unemployment and mental health found in previous meta-analyses could be explained by confounding factors, such as mental health at baseline, low socioeconomic status of the parents, etc. The importance of a number of factors that could potentially confound this relationship have been described in previous studies. For example, mental health problems at an early age were found to have a significant effect on mental health further into adulthood [27, 38] and worrying about family finances was shown to lead to more mental health problems among adolescents [45].

Second, the present review only considered evidence from cohort studies specifically related to unemployment to answer the research question. In contrast, previous systematic reviews and meta-analyses on mixed age groups combined evidence from studies investigating the effect of both unemployment and reemployment and in most cases included evidence from cross-sectional studies to reach their conclusion. However, even in these reviews, a small effect of unemployment on mental health was observed when only considering cohort studies [22 –25]. In addition, Paul and Moser provide separate evidence from a natural experiment on a factory-closure study and showed the effect of unemployment on mental health decreased tremendously compared to cohort and cross-sectional studies [22].

Third, the present study focused specifically on unemployment among young people instead of a mixed-aged population as has been done in previous reviews and meta-analyses [22 –24]. The effect of unemployment on health may differ between young people and adults based on differences in financial and family responsibilities as well as differences in the role that work plays in identity development [35, 39, 46]. This is also supported by the findings of Paul and Moser, who demonstrate that that a significant effect for adults can only be found when analysing evidence for adults and youths separately [22].

Methodological considerations

The main strength of this study is the overall good quality of the included studies. In total, 13 of the 17 included studies had a global rating of ‘strong’ in the quality assessment. Another strength is that only the evidence from cohort studies was used to assess the effect of unemployment on mental health and a differentiation was made between results without controlling for confounding variables and controlling for confounding variables, with and without mental health at baseline.

A limitation of the study is the heterogeneous character of the included studies due to differences in time of measurement, unemployment characteristics and definition of mental health. This limitation of heterogeneity was reduced by including only studies that measured mental health problems as stress, low mental wellbeing, depression, or anxiety and only focusing on high-income countries to increase the comparability of the studies. A second limitation is the over-representation of Swedish studies. Five of the 12 cohort studies are based on data from the Northern Swedish Cohort, which causes an over-representation of this cohort. However, this cohort was characterized by a high response rate, a long follow-up time and all studies received a global rating of ‘strong’ in the quality assessment. Therefore, these studies were expected to hold reliable and valuable data, which should not be disregarded. Moreover, as the included studies had different baseline measures, follow-up times and/or outcome measures, no quantitative synthesis of results was done. Instead, studies were described separately in a narrative synthesis.

Implications

The results of this study show that unemployed young people are a high-risk group for mental health problems. Ensuring appropriate services and support for this group should therefore be a priority for policymakers not only from an economic perspective but also from a public health perspective. However, to produce a sufficient knowledge base to inform policymakers on this topic, more research should be done to increase the understanding of the effect of unemployment among young people on mental health and the role that confounders play in this relationship.

Future research

To increase comparability across studies, unemployment should be defined more uniformly in future studies. This would make it possible to cluster results in a meta-analysis. In addition, a greater focus on the role of confounders, including mental health at baseline, would increase the understanding of how much of the association between unemployment at a young age and mental health is causal and what role confounders play in this relationship. This evidence would strengthen the hypothesis that there is an exposure effect of unemployment on mental health.

A second area of future research should consider the quality of employment when examining the effect of unemployment among young people on mental health. Changes in the labour market in Western European countries have led to more flexible working arrangements that also influence the quality of work [47]. This is in line with a suggestion made by Dooley, who points out that to gain a full picture of the effect of unemployment on mental health, employment should be considered as a continuous variable instead of a dichotomized one by also taking into account the quality of employment [48]. One of the few studies focusing on quality of employment in the relationship between unemployment and mental health showed that underemployment is associated with internalized mental health problems [49]. In addition, low job security is shown to be associated with increased risk on mental health symptoms and mental disorders [50]. So, to develop a full picture of the relationship between employment status and mental health, more studies considering the quality of employment are needed.

A third suggestion for future research would be to investigate the role of context in the relationship between unemployment and mental health by, for example, examining the role that unemployment benefits and active labour market programmes play in this relationship.

Conclusions

This systematic review shows an existing association between unemployment among young people and mental health. The majority of the cross-sectional studies showed that unemployment was associated with increased mental health problems, depression and anxiety disorders. However, whether there is a causal relationship between unemployment at young age and mental health is less clear. Cohort studies that did not control for confounding factors showed an effect, but when controlling for confounding factors, the effect reduced in most cases and the results were mixed. When mental health at baseline was then considered among the confounding factors, the effect decreased in all cases and only a minority of the studies found an effect. To understand how much of the association between unemployment and mental health can be considered as causation, more evidence is needed. In particular, natural experiments and longitudinal studies that control for confounding variables could provide more definite evidence.

Footnotes

Acknowledgements

The authors would like to thank the Public Health Agency of Sweden for providing the conditions to make this study possible.

Authors’ Note

Vicky H. M. Bartelink is also affiliated with Centre for epidemiology and community medicine, Stockholm county council, Stockholm, Sweden.

Conflict of interest

The authors declare that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

ORCID iD

Vicky H. M. Bartelink

References

Arnett

Žukauskienė

Sugimura

The new life stage of emerging adulthood at ages 18–29 years: Implications for mental health. Lancet Psychiatry 2014; 1: 569–576.

EuroStat. Youth unemployment by sex, age and educational attainment level, http://appsso.eurostat.ec.europa.eu/nui/submitViewTableAction.do (2018)

EuroStat. Being young in Europe today: Labour market access and participation. EuroStat Statistics Explained. Luxembourg: Publications Office of the European Union, 2015: 138–164.

Bjarnason

Sigurdardottir

TJ.

Psychological distress during unemployment and beyond: Social support and material deprivation among youth in six northern European countries. Social Sci Med 2003; 56: 973–985.

Caliendo

Schmidl

Youth unemployment and active labor market policies in Europe. IZA J Labor Pol 2016; 5: 1.

Brydsten

Hammarstrom

San Sebastian

The impact of economic recession on the association between youth unemployment and functional somatic symptoms in adulthood: A difference-in-difference analysis from Sweden. BMC Public Health 2016; 16: 230.

Nygren

Gong

Hammarstrom

Is hypertension in adult age related to unemployment at a young age? Results from the Northern Swedish Cohort. Scan J Public Health 2015; 43: 52–58.

Herbig

Dragano

Angerer

Health in the long-term unemployed. Deutsches Arzteblatt Int 2013; 110: 413–419.

Fergusson

Horwood

Woodward

LJ.

Unemployment and psychosocial adjustment in young adults: Causation or selection?

Social Sci Med 2001; 53: 305–320.

10.

Barnes

Smith

TG.

Tobacco use as response to economic insecurity: Evidence from the National Longitudinal Survey of Youth. B E J Econ Analysis Policy 2009; 9.

11.

Compton

Gfroerer

Conway

, et al. Unemployment and substance outcomes in the United States 2002–2010. Drug Alcohol Dependence. 2014; 142: 350–353.

12.

Christoffersen

Poulsen

Nielsen

Attempted suicide among young people: Risk factors in a prospective register based study of Danish children born in 1966. Acta Psychiatrica Scandinavica. 2003; 108: 350–358.

13.

Eliason

Alcohol-related morbidity and mortality following involuntary job loss: Evidence from Swedish register data. J Studies Alcohol Drugs. 2014; 75: 35–46.

14.

Krahn

Chow

Youth unemployment and career scarring: Social-psychological mediating effects?

Can J Sociol; Cahiers Canadiens De Sociologie 2016; 41: 117–138.

15.

Axelsson

Ejlertsson

Self-reported health, self-esteem and social support among young unemployed people: A population-based study. Int J Social Welfare 2002; 11: 111–119.

16.

Bacikova-Sleskova

van Dijk

Geckova

, et al. The impact of unemployment on school leavers’ perception of health. Mediating effect of financial situation and social contacts? Int J Public Health 2007; 52: 180–187.

17.

Khattab

Fenton

What makes young adults happy? Employment and non-work as determinants of life satisfaction. Sociol. 2009; 43: 11–26.

18.

Crowe

Butterworth

The role of financial hardship, mastery and social support in the association between employment status and depression: Results from an Australian longitudinal cohort study. BMJ Open 2016; 6: e009834.

19.

Virtanen

Hammarstrom

Janlert

Children of boom and recession and the scars to the mental health: A comparative study on the long-term effects of youth unemployment. Int J Equity Health. 2016; 15.

20.

Power

Clarke

Kelleher

, et al. The association between economic inactivity and mental health among young people: A longitudinal study of young adults who are not in employment, education or training. Irish J Psycholog Med 2015; 32: 155–160.

21.

Strandh

Winefield

Nilsson

, et al. Unemployment and mental health scarring during the life course. Eur J Public Health 2014; 24: 440–445.

22.

Paul

Moser

Unemployment impairs mental health: Meta-analyses. J Vocational Behavior 2009; 74: 264–282.

23.

McKee-Ryan

Song

Wanberg

, et al. Psychological and physical wellbeing during unemployment: A meta-analytic study. J Applied Psychol. 2005; 90: 53.

24.

Murphy

Athanasou

JA.

The effect of unemployment on mental health. J Occupational Organizational Psychol 1999; 72: 83–99.

25.

Cohen

Statistical Power Analysis for the Behavioral Sciences. 2nd ed., Hillsdale: L. Erlbaum Associates, 1988.

26.

Field

Discovering statistics using IBM SPSS statistics. London: Sage, 2013.

27.

Hammarstrom

Janlert

Early unemployment can contribute to adult health problems: Results from a longitudinal study of school leavers. J Epidemiol Community Health 2002; 56: 624–30.

28.

World Health Organization. Factsheet mental health: Strengthening our response. Geneva, 2016.

29.

(IHME) IfHMaE. GBD Compare Data Visualization. Seattle: University of Washington, 2016.

30.

World Bank. World Bank Country and Lending Groups, https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups (2018).

31.

Project EPHP. Quality Assessment Tool For Quantitative Studies. Hamilton, ON, 1998.

32.

Thomas

Ciliska

Dobbins

, et al. A process for systematically reviewing the literature: Providing the research evidence for public health nursing interventions. Worldviews Evidence-Based Nurs. 2004; 1: 176–184.

33.

Armijo-Olivo

Stiles

Hagen

, et al. Assessment of study quality for systematic reviews: A comparison of the Cochrane Collaboration Risk of Bias Tool and the Effective Public Health Practice Project Quality Assessment Tool: methodological research. J Evaluation Clinical Practice. 2012; 18: 12–18.

34.

Moher

Liberati

Tetzlaff

, et al. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Annals Int Med 2009; 151: 264–269.

35.

McGee

Thompson

NJ.

Unemployment and depression among emerging adults in 12 states, Behavioral Risk Factor Surveillance System, 2010. Prevent Chronic Dis 2015; 12: E38.

36.

Goldman-Mellor

Caspi

Arseneault

, et al. Committed to work but vulnerable: Self-perceptions and mental health in NEET 18-year olds from a contemporary British cohort. J Child Psychol Psychiatry. 2016; 57: 196–203.

37.

Aguilar-Palacio

Carrera-Lasfuentes

Rabanaque

MJ.

Youth unemployment and economic recession in Spain: Influence on health and lifestyles in young people (16–24 years old). Int J Public Health 2015; 60: 427–435.

38.

Mossakowski

KN.

Unfulfilled expectations and symptoms of depression among young adults. Soc Sci Med. 2011; 73: 729–736.

39.

Reine

Novo

Hammarstrom

Does the association between ill health and unemployment differ between young people and adults? Results from a 14-year follow-up study with a focus on psychological health and smoking. Public Health 2004; 118: 337–345.

40.

Thern

de Munter

Hemmingsson

, et al. Long-term effects of youth unemployment on mental health: Does an economic crisis make a difference? J Epidemiol Community Health 2017; 71: 344–9.

41.

Baggio

Iglesias

Deline

, et al. Not in Education, Employment, or Training Status Among Young Swiss Men. Longitudinal associations with mental health and substance use. J Adolescent Health 2015; 56: 238–243.

42.

Breslin

Mustard

Factors influencing the impact of unemployment on mental health among young and older adults in a longitudinal, population-based survey. Scan J Work, Environ Health 2003; 29: 5–14.

43.

Strandh

Nilsson

Nordlund

, et al. Do open youth unemployment and youth programs leave the same mental health scars? Evidence from a Swedish 27-year cohort study. BMC Public Health 2015; 15: 1151.

44.

Fergusson

McLeod

Horwood

LJ.

Unemployment and psychosocial outcomes to age 30: A fixed-effects regression analysis. Aust New Zealand J Psychiatry. 2014; 48: 735–742.

45.

Kim

Hagquist

Trends in adolescent mental health during economic upturns and downturns: A multilevel analysis of Swedish data 1988–2008. J Epidemiol Community Health 2017; 72: 101–108.

46.

Rodrı́guez YGa. Learned helplessness or expectancy-value? A psychological model for describing the experiences of different categories of unemployed people. J Adolescence 1997; 20: 321–332.

47.

Puig-Barrachina

Vanroelen

Vives

, et al. Measuring employment precariousness in the European Working Conditions Survey: The social distribution in Europe. Work (Reading, Mass). 2014; 49: 143–161.

48.

Dooley

Unemployment, underemployment, and mental health: Conceptualizing employment status as a continuum. Am J Community Psychol. 2003; 32: 9–20.

49.

Monfort

Howe

Nettles

, et al. A longitudinal examination of re-employment quality on internalizing symptoms and job-search intentions. J Occupational Health Psychol. 2015; 20: 50–61.

50.

ten Have

van Dorsselaer

de Graaf

. The association between type and number of adverse working conditions and mental health during a time of economic crisis (2010–2012). Social Psychiatry Psychiatric Epidemiol 2015; 50: 899–907.

51.

Fergusson

Horwood

Woodward

LJ.

Unemployment and psychosocial adjustment in young adults: Causation or selection?

Social Sci Med. 2001; 53: 305–320.

52.

Strandh

Winefield

Nilsson

, et al. Unemployment and mental health scarring during the life course. European J Public Health 2014; 24: 440–445.

53.

Fergusson

McLeod

Horwood

LJ.

Unemployment and psychosocial outcomes to age 30: A fixed-effects regression analysis. Aust New Zealand J Psychiatry. 2014; 48: 735–742.