The Emotional Toll of Long-Term Unemployment: Examining the Interaction Effects of Gender and Marital Status

THE GREAT RECESSION AND LONG-TERM UNEMPLOYMENT

The Great Recession is the longest and most devastating economic downturn since the Great Depression (Elsby, Hobijin, and Sahin 2010; Sum et al. 2009). From May 2007 to October 2009 the labor force lost over 7.5 million jobs and the unemployment rate climbed from 4.8 percent to 10.1 percent (Grusky, Western, and Wimer 2011; Katz 2010). Compared to previous recessions, during and after the Great Recession job loss was much higher, reemployment rates were lower, spells of unemployment were longer, the rate of underemployment was higher, and the increase in the number of discouraged workers was substantial (Farber 2011).

The level of economic dislocation and financial devastation brought on by the Great Recession was accompanied by an enormous toll on emotional well-being. Among the most salient indicators of this negative emotional toll are the effects on families, leading to rising divorce rates (Morgan, Cumberworth, and Wimer 2011), increased domestic violence (Schneider, Harknett, and McLanahan 2014), reduced fertility rates (Schneider and Hastings 2014), and diminished child development (McLanahan, Tach, and Schneider 2013).

One of the distinguishing characteristics of the Great Recession was the rise of long-term unemployment, defined by the U.S. Bureau of Labor Statistics as unemployment lasting twenty-seven weeks or longer. May 2015 Bureau of Labor Statistics data showed that 29 percent of the unemployed were long-term unemployed, compared to 18 percent in 2007.¹ In March 2010, the number of long-term unemployed individuals reached over 6.5 million, representing 44.1 percent of the unemployed population, and the long-term unemployment rate increased to 4.3 percent, much higher than the prior postwar era peak of 2.6 percent in 1983 (Katz 2010).

The picture is even more distressing if we additionally consider the estimated number of long-term unemployed workers who became discouraged and have dropped out of the labor force since the Great Recession. In March 2015 the Economic Policy Institute estimated that the U.S. has 3 million “missing workers,” defined as potential workers who are neither working nor looking for work due to weak labor market conditions (Gould 2015). Qualitative data suggest that a sizable proportion of missing workers withdrew from the labor market after a prolonged and unsuccessful search for work (Sharone 2013). The ongoing crisis of long-term unemployment can also be seen by examining patterns in unemployment duration. Whereas the mean unemployment duration in past recessions was twenty weeks, in the Great Recession mean duration increased to thirty-five weeks (Farber 2011).

Long-term unemployment affects men and women across varied segments of American society. It is widespread across different education levels, gender and race groups, and occupations and industries (Gould 2015). Recent studies show that long-term unemployment is not limited to individuals who are inflexible and therefore closed to opportunities in other occupations or industries, or to individuals who lack advanced education (Sharone et. al 2015). Although for college-educated individuals the rate of long-term unemployment is lower than for individuals without a college degree, the college-educated long-term unemployment rate is twice as high as it was before the Great Recession. Among the few predictors of long-term unemployment is age. More than half of all unemployed workers over the age of fifty-five are long-term unemployed across all levels of education (Evangelist and Christman 2013).

The severe and negative emotional toll of long-term unemployment is well established, and has been shown to be as harmful to well-being as the loss of income itself (Winkelmann and Winkelmann 1998; Sharone 2013). A line of qualitative studies dating back to the Great Depression (for example, Bakke 1933) consistently show that prolonged unemployment in the United States frequently leads to diminished self-esteem and increased self-blame, which is particularly intense among white-collar workers (Newman 1999; Smith 2001; Wanberg et al. 2012; Sharone 2013; Chen 2015). For example, Katherine Newman’s (1999) study of unemployed managers found that most managers perceived their unemployment as a personal failing. Sharone (2013) likewise describes the emotional toll on long-term unemployed white-collar workers, who come to feel that they are “flawed” or “defective.”

Meta-analytic reviews of quantitative studies also consistently show that unemployed individuals suffer from significantly lower levels of mental health compared to employed individuals (McKee-Ryan et al. 2005). Longitudinal studies show that when individuals become unemployed their psychological health deteriorates, and upon reemployment their well-being recovers. In addition, meta-analytic findings show that as unemployment duration increases, so does the negative emotional toll and the risk for mental and physical health issues (Paul and Moser 2009). For example, according to a Gallup poll (Crabtree 2014), the depression rate among long-term unemployed workers is 18 percent, compared to 12.4 and 5.6 percent for all unemployed and employed workers, respectively. Underlying the negative emotional toll of unemployment is not only stress induced by the loss of income and savings and the exhaustion of unemployment benefits but also the loss of self-esteem and self-confidence due to the experience of repeated rejections by employers. The longer individuals remain unemployed the more they experience deprivation (Brief et al. 1995), exhaustion of coping resources (Kinicki, Prussia, and McKee-Ryan 2000), and even personality changes (Boyce et al. 2015). Alan B. Krueger and Andreas Mueller (2011) found that unemployed workers’ unhappiness increases the longer they stay unemployed, and such unhappiness is particularly salient when they are engaged in job-searching activities.

The association between unemployment and negative physical health outcomes is also established by a long line of research and meta-analyses (for example, Sullivan and von Wachter 2009; McKee-Ryan et al. 2005; Paul and Moser 2009; Brand 2015). A number of studies link the emotional toll of unemployment just described—especially the experience of self-blame and the internalization of stigma among unemployed workers—with negative health outcomes (Rantakeisu, Starrin, and Hagquist 1999; Eales 1989; Creed and Bartrum 2007). Over the past decade a rapidly growing literature on the connection between stigma and health shows that the internalization of stigma is linked, among other mechanisms, to increased diastolic blood pressure (Harrell, Hall, and Taliaferro 2003; Guyll, Matthews, and Bromberger 2001) and increased cortisol output (Townsend et al. 2011; Dettenborn et al. 2010). These physical responses to internalized stigma are described in this literature as connected to experiences that threaten people’s identity and status in social hierarchies, and specifically to experiences involving conditions of uncertainty about outcomes, and in which a successful outcome involves a process of “social evaluation” of the individual by others (Keating 2009, 75; Major and O’Brien 2005). The core experience of unemployment and searching for work is precisely one of repeatedly subjecting oneself to social evaluation under conditions of great uncertainty, which upon repeated employer rejections frequently leads unemployed workers to perceive themselves as “flawed” (Sharone 2013; Newman 1999). A further mechanism linking unemployment and negative physical health is social isolation (Keating 2009; Umberson, Crosnoe, and Reczek 2010; Kawachi and Berkman 2001).

GENDER, MARITAL STATUS, AND THE EMOTIONAL TOLL OF UNEMPLOYMENT

An important but under-studied question is how the negative emotional toll of long-term unemployment varies by gender and marital status. The qualitative literature on this issue yields mixed results. Older studies dating back to the 1930s suggest that unemployment is particularly undermining of married men, due to the effects of the traditional male breadwinner role. For example, Mirra Komarovsky’s (1940, 74) classic study of long-term unemployment in the Great Depression, using in-depth interviews, finds that “man experiences a deep frustration because in his own eyes he fails to fulfill what is the central duty of his life, the very touchstone of his manhood—the role of family provider.” According to Komarovsky, this emotional toll was more acute for men whose exclusive identity was as providers than for men who also identified as fathers and husbands. A similar finding about the particular vulnerability of married men was reported by Newman’s (1999) study of unemployed managers in the 1980s. She explains: “Unemployment strikes at the heart of the masculine ideal…. [Having] failed at the task that most clearly defines his role, he suffers a loss of identity as a man” (139). Like Komarovsky (1940), Newman (1999) also found that not all men were equally vulnerable to this emotional toll; specifically, gay men who were not married were much less likely to report the anxiety and anguish expressed by married men.

A more recent qualitative study examining the emotional toll of long-term unemployment challenges the applicability of Komarovsky and Newman’s studies in the current era and claims that with changing gender roles it is now married women, more than married men, who are vulnerable to self-blame and the emotional toll of identity loss. According to Carrie Lane (2011), men in the twenty-first century “have an alternative standard of masculinity” where relying on a partner’s income is evidence of progressive manhood” (121). According to Lane, the unemployed men she interviewed took it as a badge of honor that they were “secure enough in [their] manhood to comfortably rely on [their] wife’s income” (117). For women, on the other hand, their increasingly career-based identities mean that relying on their husbands’ income is devastating to their self-worth, evoking uncomfortable feelings of being needy and dependent. Thus, in comparing the experiences of unemployed married men and women, Lane (2011) concludes that “women’s feelings of dwindling self-worth mirror those of male managers in [Newman’s] study far more so than did the comments of their male peers” (126). Lane’s claims are interesting, but more research is needed to determine whether a historical reversal has indeed occurred. One limitation of Lane’s study is that none of the men who “took pride” in letting their wives support them were part of a dual-earner couple with children. This is significant because other contemporary studies show that the traditional male breadwinner role expectations surface most clearly when married couples have children (Gerson 2011).

Turning to survey studies that compare the unemployment experiences of men and women, we also find mixed results there. Some studies find that men experience more distress associated with unemployment than women, whereas others find no difference (Broman et al. 1995; Leana and Feldman 1991). Meta-analytic findings suggest that men are more negatively affected by unemployment than women (Paul and Moser 2009). Yet several other prior studies show that it takes more time for women to find a job after a period of unemployment, and that long-term-unemployed women’s reemployment opportunities come with a greater loss of prior salary than is the case for long-term-unemployed men (Snyder and Nowak 1984).

A different line of studies explores how marital status mediates the negative emotional toll of unemployment. The dominant perspective in the psychology literature is that marriage provides critical social support and acts as a shield in difficult times such as unemployment. Prior studies that examine the effect of being married on unemployment typically focus on the social support aspect of the marriage relationship. For example, cross-sectional studies consistently find that married unemployed individuals are in a better psychological state than single unemployed individuals (Cooke and Rousseau 1984; Leana and Feldman 1991). These findings are consistent with sociological studies suggesting that marriage benefits both men and women and enhances mental health (Simon 2002). In their meta-analytic study, Frances McKee-Ryan and her colleagues (2005) found that married unemployed individuals are more satisfied with their lives than single unemployed ones. Similarly, using British panel data, Andrew E. Clark and Andrew J. Oswald (1994) found that married unemployed individuals experience less mental health distress than single unemployed individuals.

Alongside these studies showing the supportive effects of marriages, the literature also reveals that some unemployed individuals experience negative social support from their spouses, including findings of social undermining, which refers to negative evaluation of the person in terms of his or her attributes or actions (Vinokur and Van Ryn 1993). One of the reasons for such negative social support in marriages is the economic strain experienced in times of unemployment. Economic hardship changes the quality of the relationship and creates conflict in married couples (Conger, Rueter, and Elder 1999). For example, Saul D. Hoffman and Greg J. Duncan (1995) found that a husband’s low income significantly increases the risk of marital dissolution. Similarly, Peter Jensen and Nina Smith (1990) found that if the husband is unemployed, divorce is more likely to occur, but this is not the case when the wife is unemployed. However, more recent studies find that either a husband’s or wife’s unemployment leads to increased likelihood of marital dissolution (Jalovaara 2003; Hansen 2005). Another complicating factor in the current era of stagnating wages is that increased economic interdependence between married men and women prevents both husbands and wives from contemplating separation (McManus and DiPrete 2001).

Although the studies described analyze the effects of gender and marital status on the emotional toll of unemployment, prior quantitative studies have not considered how marital status interacts with gender in mediating this emotional toll. The interaction of gender and marital status is likely important, given the well-documented significant effects of marriage on well-being and given the persistence of gender differences in the experiences of men and women in contemporary society. In light of the existing literature on the effects of marriage, we expect that for both long-term-unemployed men and women, marriage offers some protective benefits and buffers against the emotional toll of unemployment. Marriage is a ready-made support system. In addition to emotional support, marriages also play a social and economic insurance role. However, based on prior qualitative studies such as Komarovsky (1940) and Newman (1999), discussed earlier, for long-term-unemployed men, we hypothesize that marriage is a double-edged sword. On the one hand it provides varied forms of support, but on the other hand—to the extent that stereotypical gendered expectations persist, which focus the male role in the family on breadwinning—long-term unemployment is likely to create tensions. Specifically, we expect that the inability of long-term unemployed men to fulfill the breadwinner role will generate marital distress that will offset the supportive benefits of marriages for men. Given that this tension is premised on the persistence of stereotypical male gender roles, we expect that this particular interaction of gender and marital status will affect men more than women. With respect to women, given the well-documented research on the labor market obstacles facing long-term-unemployed women, we expect that the emotional toll of long-term unemployment will be as difficult for women as for men, and perhaps more so, but that in the case of women, marriage will follow the general pattern of having a salubrious effect on their well-being.

QUALITATIVE DATA AND ANALYSIS

To begin our exploration of how the emotional toll of long-term unemployment varies by gender, and whether and how it is mediated by marital status, we turn to our qualitative data. The qualitative data discussed in this paper are derived from fifty in-depth interviews conducted with long-term-unemployed job seekers in 2013 and 2014. To recruit interview subjects we contacted job seekers through Boston-area career centers, networking groups, and libraries, and invited job seekers to sign up for the opportunity to participate in research in exchange for receiving free support to be provided by volunteer career coaches and counselors. Interested job seekers were asked to complete a short survey so that we could see whether they met the following criteria:

1. Unemployed six months or longer

2. Between the ages of forty and sixty-five

3. Working in white-collar occupations and college-educated

4. Looking for work in the Boston area

Over 800 job seekers signed up for the opportunity to participate, but only 125 met the criteria for participation. Our sample of 125 unemployed job seekers was 55 percent male and 45 percent female, with 70 percent of the sample being married and with a mean age of fifty-four. Our recruiting focused on unemployed workers over the age of forty because, as previously discussed, older workers are more likely than younger workers to get trapped in long-term unemployment. We expect that due to self-selection our sample consists of unemployed job seekers with higher-than-average levels of motivation to continue with their job search and to receive job search support, but we do not expect this particular selection bias to affect our findings regarding how gender and marital status mediate the emotional toll of long-term unemployment. Our focus on white-collar workers who are college educated may bias our qualitative findings, but such bias likely cuts against our hypothesis regarding how gender and marital status mediate the emotional toll of long-term unemployment. Prior research (for example, Lane 2011) suggests that male breadwinner expectations are less salient for married college-educated men than for married men without a college education. So our qualitative findings regarding college-educated white-collar workers may understate the extent of marital tensions stemming from male breadwinner expectations.

In-depth interviews with a randomly selected subset of fifty long-term unemployed job seekers were conducted prior to any of the job seekers’ receiving support. These semistructured, in-depth interviews were conducted either in person or by telephone and lasted one to two hours. We asked participants questions regarding their experience looking for work and the effect of unemployment on their well-being and personal relationships.

The in-depth interviews revealed the intense negative emotions that long-term-unemployed job seekers experience. Consistent with prior research, both male and female job seekers discussed the emotional toll of financial stress and diminishing self-esteem. For example, Deborah, a fifty-two-year-old marketing executive, discussed how job loss has entailed a “profound loss of self-esteem and confidence.” She explained: “My job was a huge part of my identity. Huge part of how people saw me. It’s hard not to feel like a throw-away.” Steven, a fifty-six-year-old computer engineer, succinctly described similar feelings: “I’m embarrassed. I’m humiliated. I feel like a loser.”

A vast literature, as previously described, finds that marriages are generally supportive of individuals’ emotional well-being. Consistent with this literature, our unmarried interviewees described painful feelings of intense isolation and the absence of support more frequently than the married unemployed workers. For example, Abby, a single woman, told us, “I have no one supporting me. I’m on my own.” Even though Abby has friends and extended family, she explains that, “It’s not like I’m talking to them every day. I really feel the weight of this is on myself. It’s really emotionally devastating to feel that sort of isolation.” Gene, a widower, described unemployment as being the “darkest period in my life,” and isolation as the most difficult aspect of this period, since his social life had been fully centered around his former work. Being home alone, “I do not talk to anyone on a daily basis.” He describes the temporary reprieve from isolation when he goes to a networking support group: “These meetings are helpful sometimes for the camaraderie for a couple hours.” But, he adds, “I find myself as the meeting is closing … I sit there and go, ‘This isn’t gonna last much longer. Maybe another twenty minutes. Then I’m going to get in my car and go home. It’s going to be dark.”

Both unmarried men and women described how as a result of their unemployment their friendships have become more fragile, and thus friends are less available as a source of support. Frank described how his friends have withdrawn: “People that I thought I knew, some have just dropped off the radar and have literally told me not to contact them. It’s like they are saying, ‘There is something wrong with you. You’re not working and we don’t want to talk to you.’” Tammy’s friends have not cut off contact, but they are “all married, so a lot of times they’re tied up with other obligations,” and even when they do get together, Tammy feels like “people don’t really want to know all the gory details about how depressing it is. They will say, ‘Oh, that’s awful, it’s depressing, let’s talk about something else.’ ” In addition to the discomforts and stigmas that lead friends to withdraw from unemployed workers, other factors lead unemployed workers to withdraw from friends. In part, such withdrawal is due to financial constraints, since being with friends usually involves spending money. Rob put it this way: “I don’t have the money to go out and do what my friends are doing.” But withdrawing from friends is also the result of social unease rooted in the loss of status. Jack put it this way: “You feel uncomfortable about where you are in your life versus where your friends are. Some of these folks have jobs, ranging from vice presidents to directors, doctors. So you tend to maybe not socialize as much as you would like to or as you did in the past.”

Some single unemployed workers also mentioned that although they would like to find a romantic partner, dating was simply not feasible during unemployment. Denise, with a laugh, rhetorically asked, “I’m already facing rejection right and left on the job front. Who needs more rejection?” But then she added, with a more serious note, “What’s the first question anybody asks you when they meet you: ‘What do you do for a living?’ Well, I don’t do anything because I’m one of these people who had the misfortune to be laid off. I can’t even see any gentleman taking me seriously.” David likewise reported wishing to be in a relationship but then added, “A part of me feels too unworthy to be…. I feel ‘less than.’”

By contrast to these unmarried job seekers’ descriptions of isolation and lack of social support, approximately half of the married job seekers we interviewed described varied forms of positive supports from their spouses. For example, Ryan shared: “I would not have been able to make it through this period without my wife. My wife could not be more supportive.” Ryan explained that in the moments when he felt most discouraged his wife would remain “really positive and hopeful” and “she’d say, ‘Oh, this is going to be your week, I can see it.’” Jen described her supportive relationship with her husband of thirty years. Despite the fact that at the time of the interview Jen and her husband were about to lose their home to foreclosure, she found much solace in the marriage, explaining: “We have been able to face the terror together and in so doing, we have built a bond that is amazingly durable. Nobody should have to go through what we’ve been through, but [the marriage] is a tremendous balm.”

While unmarried unemployed workers were more likely to describe painful isolation than married workers, marriage does not always alleviate the isolation felt by unemployed workers. In fact, about half our married interviewees also reported experiencing the emotional toll of unemployment in isolation. Steven, who said he felt “embarrassed” and “humiliated,” added that despite the fact that he is married, he feels “very much alone,” and reflected on how “it’s very hard to talk to my wife about it because she’s never been through anything like it.” Albert shared a similar experience: “My wife has no clue what it’s like.” The experience of emotional isolation within marriages is not limited to men. Linda, a fifty-eight-year-old woman who has worked for over twenty years in software sales, put it this way: “My husband doesn’t understand what I’m going through.” Because Linda’s husband’s salary is sufficient to financially support the both of them, he has difficulties relating to Linda’s negative emotions of identity loss and her growing sense of financial vulnerability.

Although our qualitative data suggest that both men and women may experience emotional isolation within marriages, when we look more closely at the narratives of married unemployed workers, a strikingly gendered pattern emerges: about a third of the men, but no women, describe marital tensions due to disagreements over job search intensity and what kind of job the unemployed spouse should seek. Tensions over job search intensity frequently took the form of men reporting that their wives did not think they were exerting sufficient effort to find work. For example, Larry is a software engineering manager specializing in speech recognition. Larry and his wife are under significant financial pressure because they have two school-age children and a large mortgage. Larry describes the strain this has put on his marriage: “[My wife] says, ‘You’re not doing enough.’ What the hell? I can only do so much. She doesn’t understand it. She hasn’t been through that herself. There is a lot of pressure if you’re in a family.”

A different but related kind of tension was discussed by Richard, a fifty-year-old public relations professional. Richard and his wife also have young children and feel intense financial stress due to Richard’s unemployment. Richard explained, “I’ve never really been in a position where I wasn’t sure whether I could pay the mortgage or buy my food or keep my car repaired.” His family is dependent on his income since his wife “doesn’t make much money. She’s mostly volunteering.” Unlike Larry’s case, the source of the marital tension, as Richard described it, is not that he is not doing enough on his job search but that he is doing too much job searching and not enough to contribute his share to the “second shift,” child care (Hochschild 1989). Richard put it this way: “My wife feels neglected because I’m wrapped up in job search related things and not spending enough time just doing other things which are a part of daily life. She might think I’m not doing anything but looking for a job is more than forty hours. You always have to be ‘on.’ That’s a stress.”

The marital tensions described by Richard and Larry are partly rooted in the difficulties spouses have of understanding the experience of unemployment. The feeling that spouses “don’t understand” is frequently reported by both men and women. The distinctly gendered pattern is the lack of understanding coupled with the suggestion that the job seeker is not doing enough on his search, or is not doing enough at home, a complaint that was only reported by male job seekers. Given the abundant research showing that married women do a disproportionate amount of house and care work in families (for example, Gerson 2011) it is perhaps not surprising that women with unemployed husbands may expect more support on the home front. At the same time, for men like Richard who struggle to do the “work” of job searching (Sharone 2013) under the stress of unemployment, this spousal expectation may signal that his wife “might think I’m not doing anything.” Interestingly, some unemployed women we interviewed did report the kind of tension described by Richard but not in the context of their marriages. Instead these tensions arose with mothers, sisters, or close friends who were disappointed at not receiving support from unemployed women who, they had presumed, had time on their hands. For example, Nadine, an unemployed unmarried woman, described changes in her relationship with her mother and sister: “They both expect me to pick up a lot more of the pieces. My mom will now ask would you mind stopping at the post office and mailing this package? The subtext is that since ‘you’re not doing anything why don’t you help me?’ They don’t understand the pressures of looking for work.”

Another kind of marital tension that was generally only reported by men focuses on the kind of job the unemployed spouse should seek. A good example is Warren, a sixty-two-year-old environmental scientist, with a seven-year-old child. Despite the fact that Warren’s wife works full-time, his family is under considerable financial stress in the absence of his income. Warren reported that his wife has pressured him to seek any available job, whether or not in his field, including a low-paying retail position. Warren explained: “It’s a huge stress on our marriage. My relationship with my wife is really fraught with difficulty because I didn’t realize how much she really wanted a breadwinner in the mix…. My wife has been pressuring me to get a job in retail sales, selling camping gear or something.”

In his long career as an environmental scientist Warren has taught at Ivy League universities and conducted research funded by multi-million-dollar grants. Although Warren has expanded the breadth of his search beyond teaching and research to include university administration, he has maintained a focus on positions that would allow him to continue making a contribution in his field of expertise. Yet the unexpected toll on his marriage from his hesitation to consider jobs such as in retail sales weighs heavily on Warren as he confides, “I don’t want to let my loved ones down. That’s a huge thing.”

In Warren’s case marital tensions arose from his spouse feeling that his search was too narrow; other men report marital tensions arising from spouses thinking their husbands’ searches are too broad. James, a fifty-three-year-old former corporate manager, has expanded his search to include finance-related positions in nonprofit companies. James believes that his corporate finance experience can be a valuable asset to a nonprofit company, and that down the road such a position will provide more security and meaning than his former work, even if this would involve an initial pay cut. Yet, James is experiencing tensions in his marriage because his wife does not support his inclusion of nonprofit companies in his search:

I feel I should take a lower-paying job and she’s like, “Oh, my god, that’s so low.” She doesn’t want me to take it. She’s thinking we’re going nowhere with me taking a lower-paying job. I feel like I have a longer-range view of things than her…. She’s just looking for how much they are going to pay me right now. And I’m looking at how much I’m going to end up getting after five to ten years.

The kinds of marital tensions described by Warren and James that were common among long-term unemployed men were strikingly absent from the accounts of marital tensions described by women. In both of these cases the marital tensions arising from unemployment are implicitly linked to the extent to which the unemployed male is perceived to be fulfilling the breadwinner role. For Tom, a fifty-five-year-old unemployed sales agent, the link of breadwinning and the state of his marriage was made explicit. Tom noted that the strain in his marriage has been somewhat mitigated due to savings that he had accumulated when he had a job, which means “I’ve still been able to provide.” But, after a pause, he added, “Absolutely, if I wasn’t able to provide, I probably would not be married today.”

To summarize, as we review the qualitative data, we find that consistent with the existing literature on the positive effect of marriage on wellbeing, about half of our married unemployed workers benefitted from forms of positive support that were less frequently available to unmarried unemployed workers. Yet our qualitative data also reveal two kinds of marital tensions that exacerbate the emotional toll of long-term unemployment. First, both men and women report that marital tensions may arise because of spouses’ inability to understand and relate to the difficult experiences of unemployment. Second, unemployed married men but not women describe marital tensions arising from spouses’ feeling that the unemployed spouse is not doing enough on his search or is not looking for the right kind of job. It is important to note that the second kind of marital tension, which was only reported by men, was in almost every case accompanied by descriptions of the family’s severe economic stress. This is perhaps not surprising, given that tensions surrounding issues like the appropriate level of job search intensity and appropriate job targets are more likely to arise under conditions of economic duress. Although the qualitative data are derived from a relatively small sample, this striking pattern does support our hypothesis that while both men and women generally benefit from being married, in the context of unemployment, lingering male breadwinner expectations for men mean that the benefits of marriage are counterweighed by marital tensions that intensify the emotional toll of long-term unemployment. In the remainder of this paper we explore this hypothesis, using a larger dataset and quantitative analysis.

QUANTITATIVE DATA AND ANALYSIS

To further examine the effects and interactions of gender and marital status in mediating the emotional toll of long-term unemployment, and specifically to explore the hypothesis generated by our qualitative data and the literature discussed in the previous section, we exploit a set of data collected right after the end of Great Recession from a sample of unemployed workers in New Jersey. The Survey of Unemployed Workers in New Jersey was conducted by the Princeton University Survey Research Center from October 2009 to April 2010. The sample consists of 6,025 unemployed individuals who were recipients of unemployment insurance in New Jersey.² The unemployed individuals in this study were surveyed for twelve consecutive weeks, with a subset of respondents who were long-term unemployed at the time of the study surveyed for an additional twelve weeks for a total of twenty-four weeks. In total, 39,201 surveys were completed. Table 1 provides information on the sample characteristics.

The survey had two parts: an initial survey, which was administered in the first week and collected information on demographics and income, and a weekly survey, administered in the first week and in each subsequent week, that gathered a wide range of information about respondents’ ongoing job search activities, time use, reservation wages, job offers, and emotional states (for a detailed description of the Survey of Unemployed Workers in New Jersey, see Krueger and Mueller 2011). Although at the time of the survey New Jersey’s overall unemployment rate was similar to the national unemployment rate, it is important to note that New Jersey’s long-term unemployment rate, with 40 percent of its unemployed workers being long-term unemployed, was among the highest in the United States.

The New Jersey dataset is well suited for addressing our theoretical and empirical questions regarding the effects of gender and marital status in shaping the experience of unemployment because it includes information on the emotional state of unemployed respondents as measured by life satisfaction, as well as time spent in a negative mood on a weekly basis. Life satisfaction was measured with the question “Taking all things together, how satisfied are you with your life as a whole these days?” Respondents were asked to pick a level of life satisfaction from a four-point scale ranging from very satisfied to not at all satisfied. The survey also measured respondents’ mood with the question “Now we would like to know how you feel and what mood you are in when you are at home. When you are at home, what percentage of the time are you: ‘in a bad mood’, ‘a little low or irritable mood’, ‘in a mildly pleasant mood’, ‘in a very good mood?’” Respondents were asked to indicate the percentage of time that they experienced each mood category.

DESCRIPTIVE RESULTS

We begin our analysis by examining descriptive statistics about the reported level of life satisfaction among the unemployed respondents. Table 2 breaks down these responses by gender and marital status. To provide some context for interpreting the responses of the unemployed individuals in the New Jersey survey we also present the responses of employed individuals to the same life satisfaction question from the Princeton Affect and Time Use Survey (PATS). PATS is a national telephone survey that was conducted in the spring of 2006 by the Gallup Organization. Although the whole PATS sample consists of nearly four thousand respondents, for purposes of our comparison with the New Jersey data we limit our analysis to the responses of employed individuals in the PATS data. As seen in table 2, the differences in the life satisfaction of employed and unemployed individuals are dramatic. Only 5.5 percent of unemployed single females are very satisfied with their lives, compared to 37 percent for employed single females. Overall, married females are more satisfied than single females, regardless of employment status, but again the effect of unemployment is enormous, with the percentage of unemployed married women who are very satisfied with their lives at 8.7 percent, compared to 55 percent for employed married females. Table 2 shows a similar pattern for men, with immense differences in life satisfaction between employed and unemployed men, and as is the case with women, married men report higher levels of life satisfaction than single men. While the comparison between the PATS and New Jersey data is limited by the fact that PATS data were collected three years earlier and are national in scope (as opposed to focusing only on New Jersey), the dramatic differences in the answers to the identical questions about life satisfaction provide a useful baseline for contextualizing the emotional state of unemployed individuals in the New Jersey survey.

For additional descriptive statistics on how the emotional toll of unemployment varies by gender and marital status, we draw on the New Jersey survey data to compare the means of the responses given to the life satisfaction question and the percentage of time respondents reported spending in a bad mood for each group (single female, married female, single male, and married male). In table 3 ANOVA (analysis of variance) statistics show that there is a significant difference between groups. Both unemployed men and unemployed women who are married report more life satisfaction and less time spent in a bad mood than single unemployed individuals.

Our descriptive results are not surprising. We see, first, that consistent with the literature on the emotional toll of unemployment, employed individuals are indeed far more satisfied with their lives than unemployed individuals, and, second, consistent with the literature on the salubrious effects of marriage, married unemployed and employed individuals are more satisfied with their lives than single employed and unemployed individuals.

CROSS-SECTIONAL ANALYSES

We next analyze whether the interaction of gender and marital status significantly predicts individuals’ life satisfaction and mood. Using data from the initial survey of unemployed New Jersey workers, as seen in table 4, we first regressed life satisfaction on gender, marital status, and a host of control variables such as education, number of children, household income, race, age, and unemployment duration. In this model, levels of life satisfaction are higher for unemployed women than for unemployed men, and higher for married than single individuals of both genders. When we added the interaction of gender and marital status we see that the interaction term is significantly and negatively related to life satisfaction. This result shows that single males experience the lowest levels of life satisfaction, followed by single females. Married women reported higher levels of satisfaction than married men. The interaction term is also significantly and negatively related to time spent in a good mood. Again, single males, as a group, spend the least amount of time in a good mood, followed by single females. Married women report spending the most time in a good mood, followed by married men. The cross-sectional analyses of the data support our descriptive findings.

PANEL DATA ANALYSIS

One of the distinguishing characteristics of the Great Recession is the crisis of long-term unemployment that came in its wake. For this reason we are particularly interested in exploring how the experience of unemployment changes over time as unemployment duration increases. Because the New Jersey survey of unemployed individuals provides data on the emotional state of respondents on a weekly basis, we are able to run individual fixed effects models to see how the emotional toll of unemployment unfolds over time for married and single men and women. Using panel data also helps reduce the likelihood that our findings are driven by omitted variables, which have the potential to be in relationship with the independent variable. By using fixed effects models we can eliminate unobserved heterogeneity across individuals, which might introduce spurious results.

In tables 5 and 6 we present fixed effects models for single females, married females, single males, and married males with the dependent variables of life satisfaction and time spent in a bad mood, respectively. Table 5 shows that as unemployment duration increases, life satisfaction increases for all groups except single males, but the effect sizes are extremely small and therefore not very meaningful.³ Much more striking are the findings with regard to time spent in a bad mood. As seen in the table 6, as the duration of unemployment increases, time spent in a bad mood significantly increases for all groups.

In short, our findings consistently show that for married and single men and women time spent in a bad mood dramatically increases with longer periods of unemployment, and because we use fixed effects models to look at the same individuals over time we can eliminate the possibility that these findings are driven by observed and unobserved time-invariant differences among individuals. Looking more closely at these findings to explore whether the extent of the increase in time spent in a bad mood significantly differs across groups, we use our panel data to run random effects models, which include the interaction of unemployment duration with married and single men and women. As seen in table 7, the interaction of unemployment duration with gender and marital status reveals a significant difference between single men and married men. The increase in time spent in a bad mood as unemployment continues is significantly higher for single men than for married men. However, it is important to note that although the increase in time spent in a bad mood as unemployment continues is likewise higher for single females than for married females, this difference is not significant.

Taken as a whole, the findings described suggest that marriage mediates and diminishes the negative emotional toll of unemployment for both men and women. In the final step of our analysis we attempt to dig deeper to understand what it is about marriage that is helpful. Our analysis here is guided by our qualitative findings, which suggest that marriages may be less supportive when married couples feel acute financial stress. Moreover, the pattern in the qualitative findings suggest that marital tensions as a result of financial stress are more likely to arise in couples with an unemployed man as compared to an unemployed woman.

To explore whether the salubrious effects of marriage vary by gender under conditions of financial stress, we separated our analysis for married and single men and women. For this analysis, following Daniel Kahneman and his colleagues (2006), we created a new dependent variable called “negative mood” by combining time spent in a bad mood and time spent in a low/irritable mood. We first examine the relationship among marital status, economic strain, and negative mood for men. As can be seen in table 8, model 1, when we regress marital status on negative mood, and control for unemployment duration, we find that, as expected, single men spend significantly more time in negative moods than married men. However, in model 2, when we add household income to the regression we find that after controlling for household income marital status became insignificant. In other words, once we control for the extra household income that typically comes from being married, there is no significant difference between the time spent in a negative mood for single men and married men. To see the effects of other variables in model 3 we included all other control variables, except household income, and still found the significant difference between single men and married men with regard to time in a negative mood. In model 4, when we added household income along with the other control variables, the marital status variable once again became insignificant. One interpretation of this striking finding is that for unemployed men, the benefits of marriage derive more from added income than from other forms of intangible or emotional supports.

We next ran the same models for women, and as can be seen in table 9, the results are strikingly different. In model 1, when we regress marital status on negative mood, and control for unemployment duration, we find that single women, just like single men, spend significantly more time in a bad mood than their married counterparts. The difference between men and women arises when we control for household income in model 2. Unlike men, women benefitted from being married even after controlling for household income, as well as after controlling for all other variables (models 3 and 4). The differences in the effects of marriage for men and women suggested by tables 8 and 9 are broadly consistent with the pattern observed in our qualitative data, which suggested more marital tensions arising for couples under financial stress with an unemployed male.

Finally, since our findings suggest that a lingering male breadwinner role increases the emotional toll of unemployment for married men, and since according to some existing studies the salience of the male breadwinner role varies by education levels—and specifically has diminished for married college-educated men (for example, Lane 2011)—to explore variations by levels of education we looked at whether men and women’s education level moderates the relationship between unemployment duration and negative mood. Our regression analyses using panel data in tables 8 and 9 show that overall, unemployed college-educated workers (which for purposes of those tables we define as workers with at least some college education) spend significantly less time in a negative mood than unemployed workers who are not college-educated. Prior studies comparing the experiences of college-educated white-collar and blue-collar unemployed workers find that white-collar workers are more likely to experience negative emotions from blaming themselves for their unemployment (Newman 1999; Sharone 2013; but see Chen 2015 finds self-blame among blue-collar workers). However, our findings suggest that when looking more broadly at negative moods, which may be triggered by a variety of sources beyond self-blame, they are more frequently experienced by workers who are not college-educated. The benefit of higher education levels holds for both men and women, but differences emerge between men and women when we control for income and other demographic variables. Among men, after controlling for income and demographic variables, college-educated unemployed workers spent significantly less time in a negative mood than unemployed male workers who did not go to college. By contrast, among women, after controlling for income and demographic variables, college-educated unemployed workers do not significantly differ in time spent in a negative mood from unemployed female workers who did not attend college.

We further examined whether education level moderates the effect of unemployment duration on negative mood for men and women separately. For this analysis, we regressed negative mood on the interaction terms of unemployment duration with education level dummies. As seen in table 10, there is a significant difference for men between the reference category of some high school and the higher education categories. As the education level increases, this difference gets bigger. However, we do not see a similar pattern for women. Figure 1 graphically shows the difference by gender in the time spent in a negative mood as it relates to level of education. The graph shows that for men, but not for women, time spent in negative mood decreases as education level increases. These findings lend some support to claims in the literature that the salience of the male breadwinner role varies by education levels, and specifically, that the male breadwinner expectations, and the emotional toll that accompanies such expectations during times of unemployment, are less salient for married college-educated men than for men without a college education.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Percentage of Time Spent in Negative Mood

DISCUSSION AND CONCLUSION

Although the Great Recession officially ended in June 2009, its disastrous effects are still with us. Strikingly, the rate of long-term unemployment remains at levels unseen in the postwar era, wreaking havoc on the finances and well-being of millions of American families. Voluminous research has shown that long-term unemployment is associated with a variety of social ills including, among others, job search discouragement leading workers to drop out of the workforce, and deteriorating mental and physical health. As discussed in the introduction to this paper prior research has linked both job search discouragement (Sharone 2013) and the deterioration of health (for example, Rantakeisu, Starrin, and Hagquist 1999; Creed and Bartrum 2007) to the severe and negative emotional toll of long-term unemployment. Drawing on both qualitative interview data as well as survey data this paper breaks new ground by examining how the negative emotional toll of long-term unemployment is shaped by the interaction of gender and marital status.

A well-established literature shows the salubrious effects of marriage for well-being. Yet, given the ongoing debate in the qualitative literature about the degree to which marriage may in some cases exacerbate the emotional toll of unemployment for married men (for example, Newman 1999) or for married women (Lane 2011), it is surprising that to our knowledge no prior survey study has focused on the interaction of gender and marital status in mediating the emotional toll of long-term unemployment.

Our in-depth interviews with long-term unemployed job seekers in the Boston area reveal a suggestive pattern. As can be expected, given the existing literature on the positive effects of marriage, many interviewees discuss the benefits of marriage. Yet approximately half the interviewees also brought up marital tensions that exacerbated the emotional toll of long-term unemployment. Although both unemployed men and women reported some tensions in their marriages arising from the fact that their spouses had difficulties understanding their prolonged unemployment, it was only men who described a second kind of tension involving spouses who suggested the unemployed male was not doing enough on their search, not doing enough at home, or not looking for the right kind of job. This pattern in the qualitative data, along with prior studies, led us to hypothesize that for unemployed men the generally supportive effects of being married would be counterweighed by marital tensions that tend to intensify the emotional toll of long-term unemployment.

Turning to our analysis of survey data of unemployed job seekers in New Jersey we confirm the immense emotional toll of unemployment, finding large differences in life satisfaction between employed and unemployed individuals, as well as the extent to which negative moods increase over time with prolonged unemployment. How does the interaction of gender and marital status shape this emotional toll? Overall our data suggest that marriages are helpful to the well-being of both unemployed men and women. Yet, an interesting finding emerges when we attempt to understand what it is about being married that is particularly helpful to well-being. Marriages may be helpful in various ways, including as a source of emotional comfort and support and as a form of economic insurance with spousal contributions to household income. It is striking that when we look at the emotional toll of unemployment for men (as measured by time spent in a negative mood) our analysis revealed there were no significant benefits to being married once we controlled for household income. The survey data we analyzed suggest that the main emotional benefit of marriage for unemployed men derives from increased household income. Further research is needed to examine this finding more closely. For unemployed women, by contrast, the benefits of being married for their emotional well-being remained significant even after controlling for household income. These findings are broadly consistent with the pattern in our qualitative data showing heightened marital tensions among married couples with an unemployed male.

This study has important implications for policymakers and practitioners working to respond to the negative consequences of the ongoing crisis of long-term unemployment. In thinking about responses to the negative emotional toll generated by unemployment—which to an important extent underlies issues such as job search discouragement and deteriorating mental and physical health—traditionally available sources of emotional support during times of crisis, most notably spouses, may not be well positioned to provide such support in the specific context of financial stress and long-term unemployment. This points to the importance of policies and practices that bolster public support institutions for long-term-unemployed job seekers and to the need to reverse the trend of defunding public sources of such support (McKenna, McHugh, and Wentworth 2012).