Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

[Link]

ORIGINAL PAPER

Caregiving as suicide‑prevention: an ecological 20‑country study

of the association between men’s family carework, unemployment,
and suicide
Ying‑Yeh Chen1,2 · ZiYi Cai3 · Qingsong Chang4 · Silvia Sara Canetto5 · Paul S. F. Yip3,6

Received: 3 August 2020 / Accepted: 7 April 2021 / Published online: 5 May 2021
© Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Purpose Suicide rates are generally higher in men than in women. Men’s higher suicide mortality is often attributed to
public-life adversities, such as unemployment. Building on the theory that men’s suicide vulnerability is also related to their
private-life behaviors, particularly men’s low engagement in family carework, this ecological study explored the association
between men’s family carework, unemployment, and suicide.
Methods Family-carework data for twenty Organization for Economic Co-operation and Development (OECD) countries
were obtained from the OECD Family Database. Sex-specific age-standardized suicide rates came from the Global Burden
of Disease dataset. The association between men’s engagement in family carework and suicide rates by sex was estimated,
with OECD’s unemployment-benefits index and United-Nations’ Human Development-Index (HDI) evaluated as controls.
The moderation of men’s carework on the unemployment-suicide relationship was also assessed.
Results Overall and sex-specific suicide rates were lower in countries where men reported more family carework. In these
countries, higher unemployment rates were not associated with higher male suicide rates. In countries where men reported
less family carework, higher unemployment was associated with higher male suicide rates, independent of country’s HDI.
Unemployment benefits were not associated with suicide rates. Men’s family carework moderated the association between
unemployment and suicide rates.
Conclusion This study’s findings that higher levels of men’s family carework were associated with lower suicide mortality,
especially among men and under high-unemployment conditions, point to the suicide-protective potential of men’s family
carework. They are consistent with evidence that where gender equality is greater, men’s and women’s well-being, health,
and longevity are greater.

Keywords Suicide · Men · Family carework · Unemployment · Human Development Index · Gender equality

4
* Silvia Sara Canetto School of Sociology and Anthropology, Xiamen University,
[Link]@[Link] Xiamen, China
5
* Paul S. F. Yip Department of Psychology, Colorado State University,
sfpyip@[Link] Fort Collins, CO, USA
6
1 Hong Kong Jockey Club Center for Suicide Research
Taipei City Psychiatric Centre, Taipei City Hospital, Taipei,
and Prevention, University of Hong Kong, Pok Fu Lam,
Taiwan
Hong Kong
2
Department of Public Health, Institute of Public Health,
National Yang-Ming University, Taipei, Taiwan
3
Department of Social Work and Social Administration,
Faculty of Social Sciences, University of Hong Kong,
Pok Fu Lam, Hong Kong

13
Vol.:(0123456789)
2186 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

Introduction United States unemployment explained a small proportion

of suicide rates variance [7].
In most parts of the world suicide rates are higher in men The reasons for the national and time-period variability
than in women [1]. National exceptions to men’s predomi- in the association between economic factors and suicide are
nance among the suicide, in recent decades, have included unclear. One perspective is that the association between a
China, Pakistan, Bangladesh, Indonesia, and Iraq [2]. Col- specific economic factor, unemployment, and suicide would
lectively, these countries include a major proportion of the be weaker where there are work-protection programs (e.g.,
global population so their suicide patterns cannot be dis- programs providing expenditures to compensate for unem-
missed as a minor case. There are also within-country excep- ployment) or active labor programs (e.g., programs aimed
tions when the data are examined by age or ethnicity. In at helping with finding employment). Studies from a diver-
other words, men’s overrepresentation among those who die sity of countries indicate that the impact, on suicide rates,
by suicide is not universal and, therefore, not inevitable [3]. of these programs can be significant, but with variations
Still, the prevailing pattern is that male suicides outnumber by population (e.g., by sex, age, ethnicity), country, region
female suicides. of a country, and type of program [14–16]. For example, a
Many theories have been proposed for this dominant 26-countries, European study found that between 1970 and
pattern. Most link men’s high suicide mortality with the 2007 a reduction in suicide rates by 0.038% required a US
stresses and the demands, on men, of employment and the $10 per person increased investment in labour market pro-
economic-provider role [4]. The typical prediction has been grams [16]. A 1990–2014 Italian study found that active-
that male suicide rates are higher when men’s employment labor programs were effective in moderating the association
and economic-provider role are under threat [5, 6], example, between unemployment and suicide only among men ages
when male employment declines (as is the case during eco- 45–54, and only in central Italy [15].
nomic downturns) or ends (for most men, in late adulthood), Multi-country evidence also suggests a suicide-protective
or when female employment raises; and the typical suicide- impact of public-domain gender equality in terms of male
prevention recommendation has been to shore up or extend suicide, and during difficult economic times. Specifically, a
men’s employment/economic provider role [7]. European-Union 20-countries 1991–2011 study found no
Studies testing the employment/economic-provider the- relation between unemployment and suicide in countries
ory of male suicide have yielded mixed findings [8]. Early (e.g., Sweden and Austria) with higher levels of gender
studies found that unemployed men generally had higher equality (as measured, for example, by the smaller size of
suicide mortality than unemployed women [9, 10]. Recent the female/male pay-gap and by higher female participation
evidence, however, challenges the assumption that unem- in political governing bodies). In fact, living in a country
ployed men are inevitably and consistently at higher suicide with more gender-egalitarian norms was especially suicide-
risk than unemployed women [e.g., 11]. Studies of the asso- protective for men under conditions of higher unemployment
ciation of female labour participation and suicide have also [17].
yielded variable findings, depending on time and context. A different perspective on male suicide and unemploy-
In an early, 53-countries study, higher female labour par- ment has been taken by Canetto (1992, 1994, 1997) [4, 18,
ticipation was associated with higher male suicide rates in 19]. Building on family and work theory and research [20],
countries with higher gender equality, in 1980 [5]. However, as well as on evidence about suicide trends by sex, globally
in a more recent study of 13 Eastern European countries in [3, 21], Canetto theorized that men’s suicide mortality is not
1990–2008, that is, post-Soviet Union, female labour partici- just a function of economic factors and of men’s public-life
pation was not related to higher male suicide rates [12]. The experiences, like unemployment. Men’s private-life behav-
largest sample (70 countries) and most recent study found iors, particularly their low engagement in family carework,
that in 2012 male suicide rates were not higher where both also matter. According to Canetto, men overinvest in eco-
female labour participation and human development were nomic-provider work and roles, and underinvest in family
high [13]. Other evidence that the association between eco- carework and roles—a pattern that leaves them vulnerable
nomic factors and male suicide varies by national context when economic-provider work and roles are threatened
comes from a study focusing on Canada, the United States, or lost. In her view, reducing men’s suicide vulnerability
and Europe during the 2008–2010 economic recession. This requires men to invest in, actually do family carework—as
study recorded marked cross-national variations in suicide a way to diversify their sources of meaning and purpose
rates—though suicide rates generally rose among men more as well as their social capital and networks. In support of
than among women during this recession. In Austria suicide her argument, Canetto pointed to how, in the United States,
rates did not increase when unemployment rose while in the older adult women have lower suicide mortality than older
adult men, despite the fact that over the lifespan women
experience more discrimination and adversities than men,

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2187

including in terms of employment, as well as, in late life, (regardless of whether that person lived in the household).
in terms of chronic illnesses, poverty, and widowhood. A Carework included the provision of personal care to a child
U.S. adversity paradox of suicide is that “White” men have or an adult, the supervision and the education of a child as
the highest suicide rates of all older adults [22]. A suicide- well as transporting children. Going with the child to the
relevant difference between U.S. women’s and men’s life- cinema, watching television with a child, and the like were
experiences is that most women do family carework as well considered a secondary activity of care work [25]. Male to
as financial-provider work, while most men primarily do female ratios of the proportion of time spent on carework,
financial-provider work. Canetto’s prediction is that male as primary and secondary activities, for parents aged 25–44,
suicide vulnerability will not be reduced just via employ- were used as an indicator of men’s involvement in family
ment-focused social-protection programs because these carework, in each country. The ratio ranged between 0.21
programs further reinforce men’s already strong identifi- (Japan) and 0.64 (Spain) for parents with one child, and 0.23
cation with work/economic-provider roles. It is by adding (Korea) to 0.62 (Norway) for parents with two children or
family-carework responsibilities to their economic-provider more (Table 1). Data on parents with one child versus two or
work-responsibilities that men can increase their psychologi- more children were analyzed separately. The Canada survey
cal resilience and social capital, and with that, boost their did not stratify by parents with one child or parents with two
suicide protection: “If a man feels responsible for another or more children. Hence for the analysis of parents with one
person’s well-being, he may reject suicide if only to avoid child, all 20 countries were included, while the analysis of
causing distress to... [that] person” [19]. The idea that fam- parents with two children or more excluded Canada. The
ily carework and responsibilities may be suicide deterrents survey was conducted in different calendar years in different
is indirectly supported by the findings of studies of reasons countries. Countries’ survey years are listed in Appendix 1.
for living. Women endorse responsibility to family as a rea- Data on unemployment were retrieved from the World
son for living more than men do (see McLaren, 2011, for Bank [26], using the same calendar year as the carework
a review) [23]. All evidence considered, therefore, there is data. Unemployment data for Austria, Estonia and the UK
ground for the hypothesis that doing family carework might were obtained from the OECD dataset [25] because the
be suicide-protective for men, particularly during economic World Bank did not collect unemployment data for these
downturns, as it is likely protective for women. This hypoth- three countries for the specific calendar years of interest.
esis has not been empirically tested. Overall and sex-specific suicide mortality data were
The current study examines questions of suicide, men’s obtained from the Global Burden of Disease dataset [27].
family carework, and unemployment at the ecological level, The International Classification of Diseases, 10th revision
based on data from 20 countries. Two theory-based [4, 18, (ICD-10) code of X60-X84 (intentional self-harm) was used
19] hypotheses were tested: (1) that the suicide rates of men to identify deaths by suicide. Data were age-standardized
and women would be lower in countries where men report based on 2000 world-population data. We used data from the
more family carework; and (2) that in countries where men same calendar year as the family carework data.
have higher engagement in family carework, the association The Human Development Index (HDI), composed by
of unemployment and suicide would be reduced, particularly the United Nations Development Programme (“UNDP”)
among men. measures a country’s performance in life expectancy, edu-
cational achievement, and standard of living [28]. HDI was
used as a control co-variate in the current analyses. We also
Methods considered unemployment benefits as a control variable.
Unemployment benefits were operationalized as the net-
Data replacement rate for 6-month unemployment in a single-
earner family with two children, an indicator available via
Data on family carework were obtained from the Time Use the OECD [29].
Survey of the “Time use for work, care and daily house-
hold chores module” of the OECD (Organization for Eco- Analytic strategies
nomic Co-operation and Development) Family Database
[24]. Detailed data-collection methodology is available in Univariate analyses were conducted to assess the associa-
the OECD website [25]. Twenty countries were included in tion between the predictors (i.e., men’s family carework
the database (complete country-list can be found in Appen- and unemployment), control covariates (i.e., HDI, unem-
dix 1). Respondents were asked to record their time-allo- ployment benefits) and suicide rates. The control covari-
cation over one or several representative days for a given ates, when relevant, were included in negative binomial
time period. Family carework was defined as the time spent regression analyses conducted to estimate the association
caring for a school-age child (or children) or another adult between men’s family carework, unemployment rates, and

13
2188 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

Table 1  Hours spent per day and proportion of time dedicated to carework, by number of children, in men and women aged 25–44
Men Women Male to female ratio of carework time
One child Two or more children One child Two or more children One child Two or more children
Hours (%) Hours (%) Hours (%) Hours (%)

Austria 1.03 (4.3) 1.90 (7.9) 2.74 (11.4) 4.22 (17.6) 0.38 0.45
Belgium 0.96 (4.0) 1.34 (5.6) 2.14 (8.9) 3.50 (14.6) 0.45 0.38
Canada# 1.70 (7.1) 3.22 (13.4) 0.53
Estonia 1.34 (5.6) 1.49 (6.2) 3.07 (12.8) 4.20 (17.5) 0.44 0.35
Finland 1.27 (5.3) 1.56 (6.5) 2.57 (10.7) 3.10 (12.9) 0.50 0.50
France 1.06 (4.4) 1.25 (5.2) 1.87 (7.8) 3.02 (12.6) 0.56 0.41
Germany 1.42 (5.9) 2.02 (8.4) 3.46 (14.4) 5.09 (21.2) 0.41 0.40
Italy 1.20 (5.0) 1.58 (6.6) 2.30 (9.6) 3.43 (14.3) 0.52 0.46
Japan 0.60 (2.5) 0.98 (4.1) 2.81 (11.7) 3.98 (16.6) 0.21 0.25
Korea 0.67 (2.8) 0.72 (3.0) 3.00 (12.5) 3.14 (13.1) 0.22 0.23
Latvia 0.84 (3.5) 0.74 (3.1) 2.74 (11.4) 2.71 (11.3) 0.31 0.27
Lithuania 0.79 (3.3) 1.80 (7.5) 2.47 (10.3) 4.30 (17.9) 0.32 0.42
Mexico 1.49 (6.2) 1.61 (6.7) 4.22 (17.6) 5.42 (22.6) 0.35 0.30
Norway 1.34 (5.6) 1.92 (8.0) 2.21 (9.2) 3.10 (12.9) 0.61 0.62
Poland 1.54 (6.4) 1.75 (7.3) 3.36 (14.0) 4.39 (18.3) 0.46 0.40
Slovenia 1.22 (5.1) 1.66 (6.9) 2.69 (11.2) 4.54 (18.9) 0.46 0.37
Spain 1.51 (6.3) 2.14 (8.9) 2.38 (9.9) 3.70 (15.4) 0.64 0.58
Sweden 1.56 (6.5) 2.16 (9.0) 3.12 (13.0) 4.13 (17.2) 0.50 0.52
United Kingdom 1.66 (6.9) 1.99 (8.3) 3.91 16.3) 5.33 (22.2) 0.42 0.37
United States 1.34 (5.6) 1.56 (6.5) 2.16 (9.0) 2.90 (12.1) 0.62 0.54
Mean (SD) (h/day) 1.23 (0.33) 1.59 (0.43) 2.82 (0.61) 3.91 (0.83) 0.45 (0.12) 0.41 (0.11)

Source: Organization for Economic Co-operation and Development (2019) OECD family database. Organization for Economic Co-operation and
Development [Link]
#
The Canada survey did not stratify by parents with one child or parents with two or more children

suicide rates, stratifying by sex and number of children not related to female suicide rates. A correlation matrix of
(i.e., one child versus two or more children). Finally, the the study’s variables is in Appendix 2.
possible moderation role of men’s family carework for the Statistical estimates of the univariate associations
relationship between unemployment rates and suicide rates between men’s family carework, unemployment rates, unem-
was assessed by adding an interaction term of men’s family ployment benefits, and HDI with suicide rates are presented
carework × unemployment rate. To evaluate whether men’s in Table 2. Higher levels of men’s family carework were
family carework moderated the association between unem- associated with significantly lower overall and sex-specific
ployment rates and suicide rates, by sex and number of chil- suicide rates. These patterns held for families with one child
dren, we organized men’s family carework into tertiles (i.e., and families with two or more children. Higher unemploy-
high, moderate, and low). ment rates were associated with higher suicide rates, par-
This study used data that were publicly accessible and ticularly in men. Unemployment benefits were not signifi-
anonymous at the origin. Therefore, ethics committee’s cantly associated with suicide rates. A higher country-HDI
approval for the study was waived. was associated with lower suicide rates. The main predictor
model, a model that included men’s family carework and
unemployment rates, shows significant associations between
Results male family carework and suicide rates; and also that the
unemployment-suicide relationship was significant for men,
Suicide rates were lower in countries where men reported but not for women. After adjusting for country’s HDI, the
more family carework. This pattern was consistent for men association between the main predictors and suicide rates
and women, and across number of children (see Fig. 1). remained generally the same. The moderation model shows
Countries with higher unemployment rates had higher that in families with one child there was a significant inter-
male suicide rates (see Fig. 2). Unemployment rates were action between men’s family carework and unemployment

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2189

Fig. 1  Associations between men’s carework (male to female ratio of carework time) and suicide rates by sex

Fig. 2  Associations between unemployment rates and suicide rates by sex

in terms of overall and male suicide rates (Table 2, last col- levels were marginal (p = 0.09 for overall suicide rates, and
umn). These findings indicate that in countries where men p = 0.07 for male suicide rates).
reported more family carework, the association between Because unemployment benefits were not a significant
unemployment rates and suicide rates was weaker, particu- suicide predictor, they were not included as a control covari-
larly in men. Interaction-terms estimates for families with ate. An analysis was conducted with HDI, unemployment
two or more children were similar, but the significance benefits, and the interaction between unemployment rate and

13
2190 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

Table 2  Men’s family carework and suicide rates: main effects and moderation effects
Univariate model (unadjusted) Main predictor model Adjusted for HDI Moderation model

β coefficient 95% CI β coefficient 95% CI β coefficient 95% CI β coefficient 95% CI

Total suicide rates

Parents with one child
Men’s carework − 2.43** (− 4.18, − 0.72) − 2.65*** (− 4.19, − 1.13) − 2.80** (− 4.68, − 0.96) − 0.36 (− 3.02, 2.28)
Unemployment 0.03 (− 0.01, 0.09) 0.04* (0.01, 0.08) 0.05* (0.00, 0.10) 0.21** (0.06, 0.36)
rate
Unemployment − 1.00 (− 2.75, 0.74) – –
benefits
Human Develop- − 3.83* (− 7.77, − 0.02) 0.65 (− 3.84, 5.06) 1.86 (− 2.16, 5.99)
ment Index
(HDI)
Men’s care- − 0.35* (− 0.66, − 0.05)
work × unem-
ployment rate
Parents with two or more children
Men’s carework − 2.31* (− 4.48, − 0.15) − 2.41* (− 4.41, − 0.43) − 2.41 (− 4.90, 0.04) − 0.40 (− 3.84, 3.01)
Unemployment 0.03 (− 0.02, 0.09) 0.04 (− 0.01, 0.08) 0.04 (− 0.02, 0.10) 0.18* (0.00, 0.36)
rate
Unemployment − 0.95 (− 2.79, 0.87) – –
benefits
Human Develop- − 3.70 (− 7.77, 0.24) 0.02 (− 5.44, 5.25) 0.86 (− 4.27, 5.84)
ment Index
(HDI)
Men’s care- − 0.32 (− 0.73, 0.08)
work × unem-
ployment rate
Male suicide rates
Parents with one child
Men’s carework − 2.62** (− 4.54, − 0.74) − 2.77*** (− 4.34, − 1.22) − 2.78** (− 4.70, − 0.91) 0.09 (− 2.66, 2.80)
Unemployment 0.05* (0.00, 0.10) 0.06** (0.02, 0.10) 0.06* (0.01, 0.11) 0.23*** (0.09, 0.37)
rate
Unemployment − 1.21 (− 2.85, 0.41) – –
benefits
Human Develop- − 5.16** (− 9.17, − 1.30) 0.04 (− 4.63, 4.49) 1.26 (− 2.72, 5.17)
ment Index
(HDI)
Men’s care- − 0.38** (− 0.68, − 0.09)
work × unem-
ployment rate
Parents with two or more children
Men’s carework − 2.50* (− 4.87, − 0.14) − 2.49* (− 4.52, − 0.48) − 2.31 (− 4.81, 0.14) − 0.02 (− 3.61, 3.52)
Unemployment 0.05* (0.00, 0.10) 0.05* (0.01, 0.10) 0.05 (− 0.01, 0.11) 0.19* (0.01, 0.37)
rate
Unemployment − 1.17 (− 2.89, 0.54) – –
benefits
Human Develop- − 5.03** (− 9.17, − 1.04) − 0.70 (− 6.31, 4.56) 0.03 (− 5.10, 4.87)
ment Index
(HDI)
Men’s care- − 0.34 (− 0.74, 0.06)
work × unem-
ployment rate
Female suicide rates
Parents with one child
Men’s carework − 2.36** (− 3.90, − 0.84) − 2.35** (− 3.94, − 0.77) − 2.95** (− 4.72, − 1.18) − 1.88 (− 4.54, 0.85)
Unemployment − 0.02 (− 0.07, 0.03) 0.00 (− 0.04, 0.04) 0.02 (− 0.03, 0.07) 0.10 (− 0.06, 0.27)
rate
Unemployment − 0.49 (− 2.87, 1.87) – –
benefits
Human Develop- − 0.20 (− 4.23, 3.83) 2.59 (− 1.41, 6.95) 3.31 (− 0.94, 8.08)
ment Index
(HDI)

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2191

Table 2  (continued)
Univariate model (unadjusted) Main predictor model Adjusted for HDI Moderation model

β coefficient 95% CI β coefficient 95% CI β coefficient 95% CI β coefficient 95% CI

Men’s care- − 0.18 (− 0.53, 0.15)

work × unem-
ployment rate
Parents with two or more children
Men’s carework − 2.31* (− 4.29, − 0.37) − 2.24* (− 4.24, − 0.28) − 3.00* (− 5.33, − 0.72) − 1.97 (− 5.22, 1.25)
Unemployment − 0.02 (− 0.07, 0.03) − 0.01 (− 0.05, 0.04) 0.01 (− 0.04, 0.06) 0.09 (− 0.09, 0.28)
rate
Unemployment − 0.42 (− 2.92, 2.06) – –
benefits
Human Develop- − 0.06 (− 4.22, 4.09) 2.66 (− 1.93, 7.50) 3.19 (− 1.57, 8.29)
ment Index
(HDI)
Men’s care- − 0.19 (− 0.62, 0.22)
work × unem-
ployment rate

Note: Data on unemployment benefits for Estonia, Latvia, Lithuania, Mexico, and Slovenia were not available. Men’s carework indicates male to
female ratio of family carework time
*p < 0.05; **p < 0.01; ***p < 0.001

Fig. 3  Moderation of men’s carework (male to female ratio of carework time) for the relationship between unemployment rates and suicide rates

13
2192 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

carework in the model. Estimates for this moderation model their spouse—after controlling for demographic, personal-
did not reach statistically significant levels due to inadequate ity, physical health, mental health, and marriage variables.
sample size (Appendix 3). Moderation effects of men’s fam- Receiving support had no effect on mortality once giving
ily carework are illustrated in Fig. 3. Its upper panel shows support was considered [32].
that for families with one child, and in countries where men This study’s findings are also congruent with a diversity
reported less family carework, the association between unem- of findings suggesting that having both family-carework and
ployment rates and men’s suicide rates was statistically sig- family-economic responsibilities is more conducive to well-
nificant (p value for slope = 0.02) whereas in countries where being, health, and longevity than a gendered division of
men reported moderate or high amounts of family carework, family labor [20]. This has been known about women for a
unemployment rates were not associated with suicide rates. long time. Women’s well-being, health, and longevity have
Men’s family carework did not affect the association between increased concomitant with increases in their participation
unemployment and suicide rates in women. The patterns in the paid labor force [34]. Employed women do better
were the same in families with two or more children (Fig. 3, than unemployed women because they have independent
lower panel). financial resources. They also enjoy broader social networks
Because work-and-family gender norms and behavior as well as greater self-esteem and confidence. Employment
evolve, we tested our model on the 12 countries with 2006 contributes to women’s greater power within and outside the
and later carework data. In this time-focused sample, higher home, and also to gains in the well-being of their children.
men’s family carework rates were still significantly associ- The benefits for men of participating in family carework
ated with lower suicide rates among men, but the modera- are less well-documented. This is partly because men have
tion effect disappeared due to inadequate data points (see yet to assume family-carework responsibilities to the same
Appendix 4, Sensitivity Analyses). degree that women have taken on family financial-responsi-
bilities, via joining the paid labor force [35]. Women still do
about two to ten times more family carework than men [36].
Discussion Until recently, family carework was considered just a cost—
a burden to be assumed by social subordinates. Disengage-
Main findings summary and interpretation ment from family carework was a symbol, and a vehicle of
men’s status and power over women. For sure, men’s low
In countries where men reported more family carework, family-carework participation contributes to their labor-
overall and sex-specific suicide rates were lower. Although market advantages (e.g., greater and more time-sustained
unemployment rates were associated with male suicide labor-force participation; higher wages; higher employment
rates, in countries where men reported more family care- quality) [36]. However, men’s low family-carework partici-
work, higher unemployment rates were not associated with pation also comes with costs. It has been hypothesized that
higher suicide rates in men. By contrast, in countries where men who do less family carework may be more impacted
men reported less family carework, higher unemployment by threats to self-esteem that arise in the workplace or from
rates were associated with elevated male suicide rates. Men’s unemployment [20].
family carework moderated the relation between unemploy- The findings of this study point to another potential cost
ment and men’s suicide rates, including when adjusting for of men’s low contribution to family carework: lesser suicide
country’s level of development. Taken together, the findings resilience, particularly when employment is threatened or
of this ecological study suggest that men’s family carework lost. As reasoned by Canetto [4, 18, 19], if the mechanisms
may protect men against suicide, particularly under difficult through which employment can be suicide-protective are
economic conditions. that employment provides a social network, as well as, in
the best of cases, power, meaning, purpose, responsibility,
Interpretation of current findings in light belonging, and mattering, then the isolation and the negative
of previous research findings feelings (e.g., of powerlessness, meaninglessness, worthless-
ness, and burdensomeness) that may follow unemployment
This study’s suicide and family-carework findings are con- could be prevented or reduced via investment in other work
sistent with observations of stronger benefits of giving sup- and in other relationships, including family carework. This
port over receiving it—in terms of longevity as well as in study’s findings challenge Durkheim’s theory that involve-
terms of quality of life (e.g., subjective health) [30–33]. For ment in public life and the financial provider role are the
example, in a U.S. study, mortality was significantly lower main drivers of suicide mortality [37]. Evidence is growing
among individuals who reported providing instrumental sup- [20, 34] that neither family carework nor family financial-
port to friends, relatives, and neighbors, and also among provider work is health-protective or longevity-promoting by
individuals who reported providing emotional support to

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2193

itself. Rather, a combination of the two appears to promote family carework. To reduce self-report bias, future studies
well-being and longevity, in men as in women. would ideally include behavioral carework data, and from the
This study’s findings are most closely related to those of point of view of multiple informants. Future research might
a study by Reeves and Stuckler [17], that more egalitarian also explore the association of non-family (e.g., volunteer)
gender practices mitigate the impact of unemployment on carework and suicide. Finally, this study relied on suicide
male suicide. Together, the findings of Reeves and Stuckler’s mortality as the outcome measure. Other suicidality measures
study and of this study challenge a trope that male suicide (e.g., nonfatal suicidal behavior) as well as other measures
is a symptom of a crisis of masculinity triggered by grow- of unwellness should be included in future studies so as to
ing gender equality [3, 38]. In fact, this study’s findings are capture the variety of idioms and signs of distress that may
consistent with growing evidence that gender equality is be used by people, depending on, for example, culture and
associated with a range of positive health outcomes in men, cohort, in response to economic adversities, like unemploy-
not only in women [39–42]. ment [3, 4, 21, 22].

Limitations
Conclusions and implications for suicide
This study’s results should be interpreted in light of its prevention
design, including its limitations. A limitation is that this is
an ecological study. Population-level observations may not Consistent with Canetto’s theory of gender and suicide [4,
reproduce at the individual level [6]. For example, we do 18, 19], and with the findings of work and family research
not know if the men who died of suicide were men who did [20], this ecological study found an association between
not engage in family carework. In addition, in this study we men’s higher engagement in family carework and men’s
did not have parental suicide data stratified by number of lower suicide rates, particularly during high-unemployment
children in the family. The value of ecological studies is that conditions. Previous studies documented lower suicide rates
they identify population-level social and economic factors among people with children (a structural proxy of family
that may be driving population-specific patterns, including carework) in high-income countries [e.g., in Denmark, 10].
suicide rates patterns [43]—with implications for universal However, no prior study had explored the possible associa-
and/ or selective prevention. Another limitation of this study tion men’s family carework and men’s suicide, and across
is its cross-sectional design. This means that many expla- several countries.
nations for this study’s findings remain viable. It may be, This study’s findings point to new directions for suicide
for example, that an abundance of psychological and social prevention. They suggest incorporating support for men’s
resources enables both men’s carework and suicide resilience. engagement in family carework in social-protection pro-
Understanding the pathways possibly linking men’s family grams aiming at reducing men’s suicide rates. This change
carework and men’s suicide, over time, would be an impor- would expand dominant frameworks of men’s suicide pre-
tant direction for future research. Third, the fact that only vention beyond employment-focused programs (e.g., active
20, mostly European and mostly Christian-majority countries labour market programs) that reinforce men’s already high-
had family carework data is a limitation—though the multi- identification with paid work, and that appear to have rela-
country design is a strength. It may be that countries that tively narrow, population-specific effectiveness (e.g., by age
collect carework data have better work and family policies group) [15]. It would also expand dominant frameworks of
or more egalitarian gender norms and practices than coun- men’s suicide prevention beyond models that view men’s
tries that do not. Therefore, studies covering a wider diversity high rates of suicide as a mental health problem to be solved
of countries are important to test and extend our findings. with mental health “treatments.” Another benefit of adding
Fourth, although two potential system-level confounders (i.e., support for family carework to social-protection programs
unemployment benefits and country’s human development is that it would affirm that family carework is a family and
level) were evaluated in this study, other potential system- social responsibility—with family-carework costs and ben-
level (e.g., country’s healthcare infrastructure) or individual- efits to be shared by men and women. In a more general
level (e.g., ethnicity, marital/relationship status) confound- sense, this study’s findings are consistent with evidence [35,
ers were not assessed. Fifth, the carework data used in this 39–42, 44, 45] that advancing gender equality may be an
study spanned a relatively broad time-period (2002–2012). important component of initiatives aimed at lessening men’s
Studies with a narrower and recent time-range of carework morbidity and mortality across a range of causes.
data are needed to assess the stability of this study’s find-
ings. Sixth, this study is based on men’s self-reports of their

13
2194 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

Appendix 1 Appendix 2: Spearman’s correlation

coefficients for the relationship
between predictors and suicide rates, by sex
Country list (year of time-use survey)

Austria (2008)
Suicide rates
Belgium (2006)
Canada (2010) Total Male Female
Estonia (2000)
Men’s ­carework# involvement (one − 0.46* − 0.45* − 0.45*
Finland (2009) child)
France (2009) Men’s ­carework# involvement (two or − 0.43 − 0.42 − 0.42
Germany (2012) more children)
Italy (2008) Unemployment rate 0.31 0.34 0.01
Japan (2011) Human Development Index − 0.28 − 0.33 − 0.10
Korea (2009) Unemployment benefits − 0.27 − 0.35 − 0.01
Latvia (2002)
Lithuania (2002) *p < 0.05
#
Mexico (2002) Male to female ratio of family carework time
Norway (2010)
Poland (2004)
Slovenia (2001)
Spain (2009)
Sweden (2001)
United Kingdom (2001)
United States (2010)

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2195

Appendix 3: Men’s family carework and suicide rates: main effects and moderation effects,
controlling for level of country’s human development index and unemployment benefits

Main effect model Moderation model

β coefficient 95% CI β coefficient 95% CI

Total suicide rates

Parents with one child
Men’s carework − 2.28* (− 4.05, − 0.51) − 1.32 (− 3.71, 1.11)
Unemployment rate 0.01 (− 0.05, 0.08) 0.14 (− 0.09, 0.38)
Unemployment benefits 0.24 (− 1.40, 1.88) 0.96 (− 1.07, 3.09)
Human Development Index (HDI) 0.34 (− 8.47, 9.07) 0.41 (− 8.21, 8.93)
Men’s carework × unemployment rate − 0.24 (− 0.68, 0.18)
Parents with two or more children
Men’s carework − 2.71 * (− 4.87, − 0.57) − 1.61 (− 4.18, 0.97)
Unemployment rate 0.01 (− 0.05, 0.07) 0.16 (− 0.06, 0.37)
Unemployment benefits 0.34 (− 1.35, 2.02) 1.12 (− 0.82, 3.08)
Human Development Index (HDI) 1.38 (− 7.87, 10.47) 2.40 (− 6.59, 11.21)
Men’s carework × unemployment rate − 0.31 (− 0.74, 0.12)
Male suicide rates
Parents with one child
Men’s care work − 2.22** (− 4.04, − 0.51) − 0.95 (− 3.30, 1.41)
Unemployment rate 0.03 (− 0.03, 0.09) 0.19 (− 0.03, 0.40)
Unemployment benefits − 0.09 (− 1.66, 1.48) 0.64 (− 1.13, 2.45)
Human Development Index (HDI) 1.06 (− 6.88, 9.01) 1.55 (− 5.98, 9.09)
Men’s carework × unemployment rate − 0.29 (− 0.69, 0.10)
Parents with two or more children
Men’s carework − 2.54* (− 4.63, − 0.48) − 1.08 (− 3.58, 1.42)
Unemployment rate 0.02 (− 0.04, 0.08) 0.21 (0.00, 0.42)
Unemployment benefits − 0.01 (− 1.67, 1.65) 0.88 (− 0.91, 2.69)
Human Development Index (HDI) 1.82 (− 6.70, 10.34) 3.63 (− 4.45, 11.70)
Men’s carework × unemployment rate − 0.38 (− 0.80, 0.04)
Female suicide rates
Parents with one child
Men’s care work − 2.65* (− 4.91, − 0.40) − 2.35 (− 5.40, 0.77)
Unemployment rate − 0.03 (− 0.11, 0.06) 0.02 (− 0.29, 0.34)
Unemployment benefits 1.24 (− 1.03, 3.56) 1.49 (− 1.29, 4.51)
Human Development Index (HDI) − 1.27 (− 13.43, 10.40) − 1.53 (− 13.84, 10.28)
Men’s carework × unemployment rate − 0.09 (− 0.69, 0.49)
Parents with two or more children
Men’s carework − 3.29* (− 5.91, − 0.71) − 2.77 (− 6.04, 0.51)
Unemployment rate − 0.03 (− 0.11, 0.04) 0.04 (− 0.24, 0.31)
Unemployment benefits 1.42 (− 0.86, 3.73) 1.76 (− 0.85, 4.40)
Human Development Index (HDI) 0.39 (− 11.95, 12.03) 0.35 (− 11.94, 11.95)
Men’s carework × unemployment rate − 0.15 (− 0.74, 0.41)

Note: Data on unemployment benefit for Estonia, Latvia, Lithuania, Mexico, and Slovenia were not available. Men’s carework indicates male to
female ratio of family carework time

13
2196 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

Appendix 4: Sensitivity analyses based on the 12 countries with more recent carework data
(after 2006)

Model 1 Model 2 Model 3

β coefficient 95% CI β coefficient 95% CI β coefficient 95% CI

Total suicide rates

Parents with one child
Men’s carework − 2.00*** (− 3.17, − 0.83) − 1.77* (− 3.31, − 0.27) − 1.69 (− 4.53, 1.21)
Unemployment rate − 0.02 (− 0.08, 0.05) 0.00 (− 0.42, 0.41)
Men’s carework × unemployment rate − 0.02 (− 0.67, 0.63)
­ ore#
Parents with two children or m
Men’s carework − 2.54*** (− 3.94, − 1.14) − 2.17** (− 3.83, − 0.59) − 2.03 (− 4.90, 0.91)
Unemployment rate − 0.03 (− 0.09, 0.03) − 0.01 (− 0.35, 0.32)
Men’s carework × unemployment rate − 0.03 (− 0.63, 0.57)
Male suicide ­rates^
Parents with one child
Men’s carework − 40.09* (− 70.35, − 9.84) − 41.90* (− 82.82, − 0.98) − 23.48 (− 117.00, 70.05)
Unemployment rate 0.11 (− 1.39, 1.60) 2.83 (− 9.52, 15.18)
Men’s carework × unemployment rate − 4.33 (− 23.81, 15.14)
Parents with two children or more
Men’s carework − 50.11* (− 87.47, − 12.74) − 48.25* (− 93.91, − 2.58) − 38.20 (− 150.92, 74.53)
Unemployment rate − 0.12 (− 1.50, 1.26) 1.08 (− 11.11, 13.28)
Men’s carework × unemployment rate − 2.14 (− 23.64, 19.37)
Female suicide rates
Parents with one child
Men’s carework − 2.68*** (− 4.12, − 1.24) − 2.17* (− 4.10, − 0.34) − 2.83 (− 5.77, 0.24)
Unemployment rate − 0.03 (− 0.12, 0.05) − 0.15 (− 0.60, 0.27)
Men’s carework × unemployment rate 0.20 (− 0.51, 0.91)
Parents with two children or more
Men’s carework − 3.42*** (− 5.24, − 1.60) − 2.67* (− 4.84, − 0.68) − 3.10 (− 6.27, 0.19)
Unemployment rate − 0.05 (− 0.13, 0.03) − 0.11 (− 0.48, 0.25)
Men’s carework × unemployment rate 0.12 (− 0.57, 0.79)

*p < 0.05; **p < 0.01; ***p < 0.001

^Linear regression

Acknowledgements YY Chen was supported by the Ministry of Sci- References

ence and Technology (MOST 108-2314-B-532-007-MY2), National
Health Research Institutes (NHRI-EX108-10818PI), and Taipei City 1. World Health Organization (2019) Global Health Observatory
Department of Health (10901-62-008). PSF Yip was supported by a (GHO) data-Suicide rates (per 100 000 population). World Health
Humanities and Social Sciences Prestigious ellowship Scheme (HSSPF, Organization. [Link]
37000320), and a General Research Fund (106200063). rates_​male_​female/​en/. Accessed 1 Feb 2021
2. World Health Organization (2014) Preventing suicide: a global
Declarations imperative. World Health Organization. [Link] ​ ww.w​ ho.i​ nt/m
​ en-
tal_​health/​suici​de-p​ reve​ntion/w
​ orld_​report_2​ 014/​en/. Accessed 1
Feb 2021
Conflict of interest The authors declare that they have no conflict of
3. Canetto SS, Cleary A (2012) Men, masculinities and suicidal
interest.
behaviour. Soc Sci Med 74(4):461–465. [Link]
socsc​imed.​2011.​11.​001
Ethical approval The study was conducted in accordance with the
4. Canetto SS (1997) Gender and suicidal behavior: theories and
Declaration of Helsinki’s standards and its later amendments. Ethics-
evidence. In: Maris RW, Silverman MM, Canetto SS (eds) Review
committee approval was waived because all data were anonymous and
of suicidology. Guilford, New York, pp 138–167
publicly available.
5. Stack S (1998) The relationship of female labor force participation
to suicide: a comparative analysis. Arch Suicide Res 4:249–261

13
Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198 2197

6. Yip PSF, Caine ED (2011) Employment status and suicide: the Organization for Economic Co-operation and Development. [Link]
complex relationships between changing unemployment rates www.o​ ecd.o​ rg/e​ ls/f​ amily/L
​ MF2_5_T
​ ime_u​ se_o​ f_w​ ork_a​ nd_c​ are.​
and death rates. J Epidemiol Community Health 65(8):733–736. pdf. Accessed 1 Feb 2021
[Link] 26. The World Bank (2019) Unemployment, total (% of total labor
7. Reeves A, McKee M, Stuckler D (2014) Economic suicides in force) (modeled ILO estimate). The World Bank. [Link]
the great recession in Europe and North America. Br J Psychiatry world​bank.​org/​indic​ator/​SL.​UEM.​TOTL.​ZS?​most_​recent_​year_​
205(3):246–247. [Link] desc=​false. Accessed 1 Feb 2021
8. Rehkopf DH, Buka SL (2006) The association between suicide 27. Institute for Health Metrics and Evaluation (2019) GBD Data.
and the socio-economic characteristics of geographical areas: a Institute for Health Metrics and Evaluation. [Link]
systematic review. Psychol Med 36(2):145–157 hdata.​org/​gbd/​data. Accessed 1 Feb 2021
9. Sundquist J, Johansson SE (1997) Indicators of socio-economic 28. United Nations Development Programme (2017) Human devel-
position and their relation to mortality in Sweden. Soc Sci Med opment data-Human Development Index. United Nations Devel-
45(12):1757–1766. [Link] oi.​o rg/​1 0.​1 016/​s 0277-​9 536(97)​ opment Programme. [Link] Accessed 1 Feb
00107-x 2021
10. Qin P, Mortensen PB (2003) The impact of parental status on the 29. Organization for Economic Co-operation and Development (2020)
risk of completed suicide. Arch Gen Psychiatry 60(8):797–802. Net replacement rate in unemployment. Organization for Eco-
[Link] nomic Co-operation and Development. [Link]
11. Kposowa AJ, Aly Ezzat D, Breault K (2019) New findings on Index.​aspx?​DataS​etCode=​NRR. Accessed 1 Feb 2021
gender: the effects of employment status on suicide. Int J Womens 30. Abolfathi Momtaz Y, Ibrahim R, Hamid TA (2014) The impact of
Health 11:569–575. [Link] giving support to others on older adults’ perceived health status.
12. Kolves K, Milner A, Varnik P (2013) Suicide rates and socioeco- Psychogeriatrics 14(1):31–37. [Link] oi.o​ rg/1​ 0.1​ 111/p​ syg.1​ 2036
nomic factors in Eastern European countries after the collapse of 31. Bar-Tur L, Ifrah K, Moore D, Katzman B (2019) Exchange of
the Soviet Union: trends between 1990 and 2008. Sociol Health emotional support between adult children and their parents and
Illn 35(6):956–970. [Link] the children’s well-being. J Child Fam Stud 28(5):1250–1262
13. Chen YY, Chen M, Lui CSM, Yip PSF (2017) Female labour 32. Brown SL, Nesse RM, Vinokur AD, Smith DM (2003) Providing
force participation and suicide rates in the world. Soc Sci Med social support may be more beneficial than receiving it: results
195:61–67. [Link] from a prospective study of mortality. Psychol Sci 14(4):320–327.
14. Kaufman JA, Livingston MD, Komro KA (2020) Unemployment [Link]
insurance program accessibility and suicide rates in the United 33. Thomas PA (2010) Is it better to give or to receive? Social support
States. Prev Med 141:106318. [Link] and the well-being of older adults. J Gerontol B Psychol Sci Soc
2020.​106318 Sci 65B(3):351–357. [Link]
15. Mattei G, Pistoresi B, De Vogli R (2019) Impact of the economic 34. United Nations (UN) Women (2015) Transforming economies,
crises on suicide in Italy: the moderating role of active labor mar- realizing rights: progress of the world’s women 2015-2016 sum-
ket programs. Soc Psychiatry Psychiatr Epidemiol 54(2):201–208. mary. United Nations (UN) Women. [Link] rogre​ ss.u​ nwome​ n.o​ rg/​
[Link] en/​2015/​pdf/​UNW_​progr​essre​port.​pdf. Accessed 1 Feb 2021
16. Stuckler D, Basu S, Suhrcke M, Coutts A, McKee M (2009) 35. Barker G (2014) A radical agenda for men’s caregiving. IDS Bull
The public health effect of economic crises and alternative 45:85–90
policy responses in Europe: an empirical analysis. Lancet 36. Ferrant G, Pesando LM, Nowacka K (2014) Unpaid care work:
374(9686):315–323. [Link] the missing link in the analysis of gender gaps in labour outcomes.
61124-7 OECD Development Centre. [Link]
17. Reeves A, Stuckler D (2016) Suicidality, economic shocks, and opment-​gender/​Unpaid_​care_​work.​pdf. Accessed 1 Feb 2021
egalitarian gender norms. Eur Sociol Rev 32(1):39–53. [Link] oi.​ 37. Durkheim É (1897) Suicide: a study in sociology. Free Press, New
org/​10.​1093/​esr/​jcv084 York
18. Canetto SS (1992) Gender and suicide in the elderly. Suicide Life 38. Jordan A, Chandler A (2019) Crisis, what crisis? A feminist
Threat Behav 22(1):80–97 analysis of discourse on masculinities and suicide. J Gend Stud
19. Canetto SS (1994) Gender issues in the treatment of suicidal indi- 28(4):462–474
viduals. Death Stud 18(5):513–527 39. Holter ØG (2014) “What’s in it for men?”: old question, new data.
20. Barnett RC, Hyde JS (2001) Women, men, work, and family: an Men Masc 17(5):515–548
expansionist theory. Am Psychol 56(10):781–796. [Link] oi.o​ rg/​ 40. Kavanagh SA, Shelley JM, Stevenson C (2017) Does gender ineq-
10.​1037//​0003-​066x.​56.​10.​781 uity increase men’s mortality risk in the United States? A multi-
21. Canetto SS (2008) Women and suicidal behavior: a cultural analy- level analysis of data from the National Longitudinal Mortality
sis. Am J Orthopsychiatry 78(2):259–266 Study. SSM Popul Health 3:358–365. [Link]
22. Canetto SS (2017) Suicide: why are older men so vulnerable? Men ssmph.​2017.​03.​003
Masc 20:49–70 41. Stanistreet D, Bambra C, Scott-Samuel A (2005) Is patriarchy the
23. McLaren S (2011) Age, gender, and reasons for living among Aus- source of men’s higher mortality? J Epidemiol Community Health
tralian adults. Suicide Life Threat Behav 41(6):650–660. [Link] 59(10):873–876. [Link]
doi.​org/​10.​1111/j.​1943-​278X.​2011.​00061.x 42. Van de Velde S, Huijts T, Bracke P, Bambra C (2013) Macro-level
24. Organization for Economic Co-operation and Development (2019) gender equality and depression in men and women in Europe.
OECD Family Database. Organization for Economic Co-operation Sociol Health Illn 35(5):682–698. [Link]
and Development. [Link] 9566.​2012.​01521.x
Accessed 1 Feb 2021 43. Alothman D, Fogarty A (2020) Global differences in geography,
25. Organization for Economic Co-operation and Development (2019) religion and other societal factors are associated with sex dif-
LMF2.5: time use for work, care and other day-to-day activities. ferences in mortality from suicide: an ecological study of 182

13
2198 Social Psychiatry and Psychiatric Epidemiology (2021) 56:2185–2198

countries. J Affect Disord 260:67–72. [Link] 45. Kavanagh SA, Shelley JM, Stevenson C (2018) Is gender ineq-
jad.​2019.​08.​093 uity a risk factor for men reporting poorer self-rated health in
44. Heymann J, Levy JK, Bose B, Rios-Salas V, Mekonen Y, Swami- the United States? PLoS ONE 13(7):e0200332. [Link]
nathan H, Omidakhsh N, Gadoth A, Huh K, Greene ME, Darm- 10.​1371/​journ​al.​pone.​02003​32
stadt GL, Gender Equality N, Health Steering C (2019) Improving
health with programmatic, legal, and policy approaches to reduce
gender inequality and change restrictive gender norms. Lancet
393(10190):2522–2534. [Link]
30656-7

13
Social Psychiatry & Psychiatric Epidemiology is a copyright of Springer, 2021. All Rights
Reserved.