Working conditions matter for our wellbeing–we spend about half of our waking life at work, and one of the critical attributes of our jobs is the flexibility it provides, which does affect greatly the other half of our waking life. Flexible working schedules or employee-centered flextime offers greater freedom and autonomy to conduct and navigate through our daily lives. Thus, we hypothesize, flextime will considerably improve one’s happiness.

Autonomy is not only a desire but arguably one of the basic human needs (Ryan and Deci 2000), and per livability theory (Veenhoven 2014), unfulfilled needs will make us unhappy. For instance, physicians complain that a lack of autonomy makes them unhappy (Lickerman 2012). A case example is a student, who worked at a gas station and experienced the polar opposite to having time autonomy (total flextime), and even worse than inflexible fixed work time–unpredictable and irregular working hours–and it made him very unhappy. Also, flexibility and autonomy should arguably promote intrinsic motivation among employees. Instilling intrinsic motivation and goals predict greater happiness (Schmuck et al. 2000; Roberts 2011).

Happiness is defined as “overall judgment of life that draws on two sources of information: cognitive comparison with standards of the good life (contentment) and affective information from how one feels most of the time (hedonic level of affect)” (Veenhoven 2008, p. 2). Happiness is reasonably precise, reliable, and valid measure, at least within-country or culture (Myers 2000; Oswald and Wu 2009; Diener et al. 2013). We follow usual practice in social indicators research and use terms “happiness” and “subjective wellbeing (swb)” interchangeably. Finally, to be clear, we focus here on general or overall happiness, not just a domain-specific happiness, such as job satisfaction.

There are only few studies regarding the relationship between work time flexibility and happiness. Bryson and MacKerron (2016) use smartphone data to study the context of work in the UK. Moen et al. (2016) study flexible schedules as workplace intervention.Footnote 1Golden et al. (2013) took an approach similar to ours, but we extend the previous work in several ways. First, we use more recent data for 2010 and 2014. Second, we add a key measure–the employee’s perceived input into their work schedule. Third, the prior research pertained mainly to the differences between hourly paid and salaried workers. Fourth, the estimation method includes more control variables, such as for workers’ region of residence. Finally, the present study situates the issue less in the literature of work-life and more in the philosophical conceptions of subjective wellbeing and work in a market society. An important limitation of earlier investigations is that they do not explain why flexibility should be associated with greater happiness, or why fixed work schedules should lead to unhappiness?

Wage Slavery and Commodification

“You are hired slaves instead of block slaves. You have to dread the idea of being unemployed and of being compelled to support your masters” (p. 283 Goldman et al. 2003).

Critics argue that under a system of capitalism, workers may be considered to be like “wage slaves,” at least in some important ways.Footnote 2 (Goldman et al. 2003; Stefan 2010) It is, to use Marcuse (2015) language, ’voluntary servitude’–it is voluntary because one can pick her master, it is servitude, because one has to have a master (unless one is a master or capitalist herself).

Esping-Andersen thinks of labor as of a commodity, and hence a notion of “commodification,” and its reverse “decommodification”– “labor is decommodified to the degree to which individuals or families can uphold a socially acceptable standard of living independent of market production”(1990, p 37).Footnote 3 Esping-Andersen goes on to argue that “the market becomes to the worker a prison within which it is imperative to behave as a commodity in order to survive” (1990, p. 36). Lane (2000) contends that markets are indifferent to the fate of individuals and that markets make people unhappy. Radcliff (2001) follows the thought: “I argue that the principal political determinant of subjective well-being is the extent to which a program of “emancipation” from the market is ’institutionalized’ within a state.”

It has been shown multiple times at the societal level that decommodification is associated with greater happiness (Lane 2000; Radcliff 2001; Pacek and Radcliff 2008a, b; Radcliff 2013; Okulicz-Kozaryn et al. 2014). Herein, we see flextime and setting one’s own work schedule as one step in the direction of emancipating one’s time from the vagaries of market, becoming more autonomous and free, thus, becoming less of a wage slave.

In addition, the quality of jobs more generally have been associated with both subjective and objective measures of wellbeing among those employed (Budd and Spencer 2015). This includes the role of working time as one of the important objective conditions of a job or work that contribute to a worker’s subjective wellbeing indicators, such as job or life satisfaction (Findlay et al. 2013). Employees’ level of subjective wellbeing, in turn, can feed back to work and the workplace productivity–thus, job and general life can be and has been improved by quality of work programs (Oswald et al. 2015).

Data and Method

We use the US General Social Survey (GSS) dataset containing two attached modules, the Quality of Worklife (QWL) and International Social Survey Program’s (ISSP) Work Orientations (WO). We pool data from 1998, 2002, 2006, 2010, and 2014. The GSS is a nationally representative sample collected from face-to-face interviews. We retain only respondents working full-time or part-time. The GSS contains a standard happiness question, which reads “Taken all together, how would you say things are these days–would you say that you are very happy, pretty happy, or not too happy?” and answers are coded as 1=”not too happy,” 2=”pretty happy,” and 3=”very happy.”Footnote 4. All variables are defined in Table 1. Distributions of all variables are shown in the Appendix in Fig. 1. Table 1 lists two measures of flexibility that come from the QWL, plus one measure from the WO survey (who set working hours). The typical controls used in the empirical literature regarding respondent happiness are then listed (Okulicz-Kozaryn 2016; Berry and Okulicz-Kozaryn 2011). One additional variable is included, the number of hours worked last week. It is important to distinguish between schedule flexibility and and the length of work hours. Perhaps, schedule flexibility is relatively more meaningful for the wellbeing of certain workers, such as those who also work long hours.

Table 1 Variable definitions
Full size table
Fig. 1
figure 1

Variables’ distribution

Full size image

We also control for the important role for income–we use household income and not personal income because there are many more missing observations on personal income, and also one’s happiness is clearly affected by household income, at least indirectly, not only by personal income. Also, household income may matter more for the relationship between flexibility and one’s happiness–flexibility may contribute more to happiness if household income is low–one can save a lot of time and money with flextime: avoid traffic congestion, take advantage of off-peak pricing, manage care of children or elderly better, and coordinate work with other household members and responsibilities better than if schedules were fixed.

We add in a control for one’s self-rated level of health. There is some disagreement about the direction of causality, i.e., whether health predicts happiness or happiness predicts health (Diener 2015). The most recent evidence suggests that the health causing happiness is predominant (Liu et al. 2016), and we follow it here. We also postpone introduction of health variable to last step in model elaboration.

In addition to these variables, we also include two sets of dummy variables. The occupation dummies are based on ISCO classification of 1-digit occupations: professional, administrative/managerial, clerical, sales, service, agriculture, production, transport, craft, and technical. Occupation dummies are important to control for because there are differences across occupations in working conditions that could affect happiness, and there are differences in flexibility across occupations. We seek to pick up the direct influence of flexible work scheduling, controlling for the other specific aspects of occupations. We also include twelve regional dummies (census regions) to control for potential place or cultural differences in work or wellbeing: New England, Middle Atlantic, E. Nor. Central, W. Nor. Central, South Atlantic, E. Sou. Central, W. Sou. Central, Mountain, and Pacific.

We use OLS estimation, which Ferrer-i-Carbonell and Frijters (2004) showed will yield substantially the same results with those from discrete models, and indeed, OLS became the norm in the literature measuring associations with happiness (Blanchflower and Oswald 2011).

Results

Results for each of the three measures of flexibility are presented in a separate table, and each table contains four models. The first model is bivariate. The second model sequentially adds family income, reflecting a clearly important characteristic of jobs or households that influences a worker’s happiness. The third column adds socio-demographic variables known to predict happiness, and the occupation and region dummies. The last, fourth column, adds health and number of hours worked last week, another key characteristic of one’s job–these two variables are added at the very end because they have many missing observations.

Table 2 shows results for who set working hours (i.e. schedules). The base case is the middle category ’i decide w/limts.’ It turns out that such limited flexibility is no more significantly associated with happiness than having no flexibility or discretion at all (’employer decides’). Full flexibility (’free to decide’), on the other hand, is associated with considerable happiness in column a1. Although elaboration of the model in subsequent columns attenuated somewhat the effect of flexibility, its effect persists despite all control variables included. In fact, only schedule flexibility, married, and health variables remain significant in the full model. Also, note that the size effect is as much as half that of being married (.13 v .27), and about as big as one step on 4-step health scale (.15), for instance, having control over one’s work schedule contributes as much to happiness as having one’s health go from “good” to “excellent.” Thus, the effect of having discretion into one’s work schedule is salient and meaningful.

Table 2 OLS of happiness on who set working hours
Full size table

The other two flexibility variables are available for multiple years in the QWL: 2002, 2006, 2010, and 2014. Table 3 shows results for can change schedule (start and end times of work). The base case is the lowest category ‘never’. As in Table 2, where there was no difference between the two lowest categories ‘employer decides’ and ‘i decide w/limits’, ‘rarely’ is no different from ‘never.’ Having flexibility only on rare occasion yields no difference in terms of happiness. An ability to change one’s daily schedule ‘often’, on the other hand, is associated with markedly greater happiness. The positive impact of ‘often’ remains robust, with all controls included, although its size effect is a bit muted.

Table 3 OLS of happiness on can change schedule
Full size table

Finally, Table 4 shows results for not hard to take time off. Again, the base is the lowest category ‘very hard’, and again, there is no difference between second lowest category ‘somewhat hard’ and the base. The most flexible category ‘not at all hard’ is not only very significant statistically, but also substantially.

Table 4 OLS of happiness on not hard to take time off
Full size table

The Appendix contains beta coefficients that confirm that schedule flexibility has a strong positive association with happiness, indeed, about as strong as the effect of income, and about one fourth of the size effect of health. The effect of having considerable frequency or ease of schedule flexibility is large and thus unambiguously positive, given that controlling for occupation would capture most other contributing working conditions—arguably larger than what most people would expect vis-à-vis other contributors to happiness.

Discussion

Almost 100 years ago, Keynes ([1930] 1963) envisioned the future for grandchildren of his generation who thanks to continued economic growth would finally enjoy the fruits of painful laboring for centuries. Keynes envisioned more leisure and enjoyment. This has not (yet) transpired in most countries. It is debated whether the average length of working hours is in general declining, or just for certain subsets of workers–in particular, those who are not salaried (Golden and Figart 2000)–but actually we do now devote more hours to labor than before industrial revolution (Schor 2008). Moreover, average real wage rates have stagnated over the past half acentury despite agrowing rate of labor productivity (Bivens and Mishel 2015). Societal happiness does not depend on economic growth (Easterlin et al. 2010), but rather depends on growth in wage rates (Fischer 2008). Another explanation of Easterlin paradox may be “wage slavery.” As Marcuse put it (2015):

“Happiness,” said Freud, “is no cultural value.” Happiness must be subordinated to the discipline of work as full-time occupation, to the discipline of monogamic reproduction.

One key working condition, having discretion or more control over one’s work schedule, such as with daily flextime, arguably may serve to lessen the degree of exploitation of labor resulting from longer hours for no greater real wage level. Indeed, the ability to control the timing of work, with full flextime, improves not only individuals’ subjective wellbeing, but moves a society towards a more humanistic civilization for which philosophers, social theorists, and intellectuals have been advocating for decades (Fromm 1944, 1962, 1964, [1941] 1994; Marcuse 2015; Maslow 2013; Harvey 2014).

Similar although earlier and more limited GSS and Quality of Worklife datasets were used to study the relationship between happiness and the other dimensions of working hours, such as their length (Golden and Okulicz-Kozaryn 2015), involuntary nature of extra working hours (Golden and Wiens-Tuers 2006), and a focus on outcomes other than happiness, such as work-family conflict (Golden et al. 2011). Moreover, the present paper controls for the influence of number of work hours and focuses on the isolated role of flexible work schedules, including a question from a second data source reflecting workers’ decision input into their work schedules. By focusing on flexibility and happiness, it is thus a contribution that is differentiated from the vast empirical literature on hours mismatches or long hours and health (e.g., Costa et al. 2006; Dembe et al. 2008; Beckers et al. 2008; Kleiner and Pavalko 2010; Bell et al. 2012; Başlevent and Kirmanoğlu 2014) including one using the 2002 GSS data (Grosch et al. 2006), association of hours and happiness (Rätzel 2009) or other aspects of wellbeing (Wooden et al. 2009; Wunder and Heineck 2013) and the association of flexibility and work-life balance (e.g., Lyness et al. 2012; Golden et al. 2016).

The usual caveat is that, without experimental data, causality is difficult if not impossible to establish, but real experiments are almost never possible and quasi-experiments are often inadequate to ensure causality as well. We would argue that one’s work schedule is often quite exogenously determined–few people have the luxury of picking among many jobs or their conditions, or their precise daily work schedule with their jobs. Rather, jobs and their schedule characteristics are mostly given, and presented as a take-it-or-leave it choice for applicants and incumbent workers. Thus, we can safely assume that the direction of causality runs from schedule flexibility to happiness, although we may not entirely rule out that happier workers self-select (in the longer run) into jobs featuring more schedule flexibility (this would be testable with panel data, which controls for the individuals’ pre-existing level of happiness). It is unlikely, however, even in the long run, that most happy people would end up in flexible jobs and unhappy people in inflexible jobs, particularly as some kind of discretionary choice. If anything, there is more risk of unobserved characteristics that may affect both jobs and happiness, such as personality attributes. That is arguably a key potential limitation–certain personalities (e.g., extroverts) may be more likely than others (e.g., introverts) to end up in occupations with flexible scheduling opportunities. Personality traits and other potential confounders are likely to be relatively stable over time, and hence, use of panel data with observations on pre-existing personality traits should help to alleviate this problem. However, as of now, there is no long running panel for the US containing happiness, schedule flexibility, and personality items.Footnote 5

That one working condition, having a great deal of work schedule flexibility matters as much as income or as much as quarter of the effect of one’s health is arguably larger than in the common wisdom. This is thus a new area ripe for additional happiness research–to point to surprising or nonintuitive findings so that irrational human beings (Ariely 2009) can make better informed choices, choices that will make them happy.

In terms of public policy, our results support contemporary modifications of the basic US Fair Labor Standards Act workweek rules, as well as, workplace and organizational flexibility practices generally. In particular, employers can improve employees happiness with a more advanced human resource management of providing system more frequent discretion of when employees engage in work activity. In addition, public policy makers could institute an individual worker “right to request” a change in the timing (and number) of their work hours and time off, protected from retaliation from making such requests, and be granted that request unless there is a clear business disruption–the result would likely be happier workers, firms no worse off, and perhaps better productivity or performance.