Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
By Sébastien LaRochelle-Côté and Claude Dionne
Full article in PDF format
One of the most significant social transformations of the past few decades has been the increase in the total time spent at the workplace by couples, essentially driven by the substantial rise in the labour market participation of women (Marshall 2009). While this increase in labour market participation has been advantageous in many ways (e.g. rising economic output, more income to meet family needs), parents may feel they have less and less time available for their children or for themselves, and may find it increasingly challenging to reconcile family and work responsibilities—especially if they consistently work long hours year after year.
This paper looks at the work patterns of families over a five-year period. The longitudinal focus is necessary because other studies have shown that individual work patterns may vary extensively over time (Bluestone and Rose 1997). It is also advantageous because relationships between work time and indicators of well-being are likely to be more robust when studied over a longer period (see Data source and definitions). Furthermore, longer-term patterns of labour market participation are likely to be more representative of what families experience in terms of time spent at work and elsewhere (Heisz and LaRochelle-Côté 2006).
The paper also documents differences in work patterns between families with children and families without children and discusses the potential effects of long work hours on the well-being of families with children. Families with children may face a particular set of challenges related to work–life balance when working long hours. Families with long hours are those with two adults working full time, with at least one working a particularly high number of hours.
The study of work patterns over several years requires a careful approach as the work patterns of individuals and families may vary substantially over time. To deal with this, a relatively simple method (Bluestone and Rose 1997, and Heisz and LaRochelle-Côté 2006) can be used (Chart A).
The first category—those never working—consisted of individuals who did not participate in the labour market in any of the five years (12% of adults in sample). The second category covered workers with at least one year below 1,500 hours and none above the 2,300-hour threshold (42% of adults). These workers were considered to be working ‘low' hours since they averaged 1,000 hours per year over the five years. The third category contained individuals consistently working 1,500 to 2,300 hours (22%). This is the ‘standard' category since the average 2,000 hours per year corresponds roughly to one full year at 40 hours per week. The fourth category was those with ‘long' hours—at least one year above the 2,300-hour threshold and no year below 1,500 hours (16%). These individuals worked 2,500 hours per year on average, surpassing the standard group by 25%. Finally, in the ‘high-low' category were individuals with particularly variable work hours—less than 1,500 hours in at least one year, more than 2,300 in at least one other—but with an average very similar to the standard category (1,800 hours compared with 2,000).
Work patterns are not necessarily problematic as they are often the product of individual choices. However, those that involve longer hours may become more challenging when they are associated with adverse effects on well-being. Stress, in particular, is an important effect that is widely used as a prime indicator of well-being in the literature, as it is associated with adverse effects on psychological and physiological health (Wilkins and Beaudet 1998). Stress is also a natural consequence of ‘role overload'—having too much to do and too little time to do it (Higgins and Duxbury 2002).
The importance of stress has led a number of commentators to investigate the association between stress levels and work hours (Higgins and Duxbury 2002, Hébert and Grey 2006, and Heisz and LaRochelle-Côté 2006). As a result, stress can reasonably be used as a good proxy for work patterns more likely to be associated with adverse effects on well-being.3
Individuals working long hours consistently reported significantly higher levels of stress (Chart B). For instance, 16.9% of individuals with long hours reported higher stress levels in at least three of the five years, compared with 10.9% of the population as a whole and 9.7% among those with consistently standard schedules. Nearly half of all individuals with long hours were stressed in at least one year, compared with 38.5% of the population as a whole. This suggests that individuals with long schedules are more likely than others to feel the adverse effects of work time. It also suggests that long hours are less likely to be welfare-maximizing choices for individual workers.4
Describing long-term work patterns of individuals is relatively straightforward, but describing family work patterns is more complicated since every family has two adults who may have variable work schedules over time. To simplify this, the high-low and standard categories were combined. The merger of these two categories is perhaps debatable as high-low workers might face different labour market challenges (and they also report slightly higher stress levels than standard individuals), but it is reasonable since they work as many hours as standard workers on average and are closer to standard workers than individuals with long hours are in terms of stress levels. The work patterns of the two adults in the family were then used to create 10 family work patterns, ranging from the least labour intensive (both adults not working) to the most (both with long hours) in terms of average annual family work hours over five years.
Families were clearly concentrated in certain patterns (Table 1). More specifically, almost 43% of families had one adult with low hours and another with a standard or long hours. Families having one adult with low hours and one with standard hours put in approximately 3,100 hours per year on the job, while those having one adult with low hours and one with long hours did approximately 3,500 hours.
The category with both adults working a consistently standard schedule was only 14%, which suggests a lot of variation in family work patterns and underscores the need to examine patterns over a longer run. Consistently standard families spent an average 3,900 hours per year at work, which is the equivalent of two full-year schedules at 40 hours per week.
Work-intensive categories—one adult with long hours and the other with at least a standard schedule—also accounted for 14% of families (only 3% had both adults with consistently long hours). These families averaged at least 4,400 hours per year on the job.
At the other end of the spectrum, 9% of families had one adult not working at all over the five years but the other with at least a standard schedule. Those with the working partner putting in long hours did nearly 2,500 hours on average; those with a standard-schedule partner, 1,900. The three least labour-intensive categories together accounted for approximately 20% of families with two adults.
Lack of time raises a different set of well-being issues for families with children. For instance, studies have shown that children enjoying more available parental hours fare better at school (Curtis and Phipps 2000). Other studies also correlate children's health with hours worked by parents (Anderson et al. 2003). Significant differences in work patterns can be seen between families with children and families without children, even after adjusting for age differences (Table 2).5 More particularly, after adjusting for age differences, families with children were less likely to have both parents working a consistently standard schedule (14%) than families without children (21%). Families with children were also much more likely to have one parent with low hours and the other with at least a standard schedule—51% compared with 41% of age-adjusted families without children. Parents with children were also less likely to fall into the two most work-intensive categories. These results suggest that the presence of children is correlated with differences in work patterns. The greater share of families with children having at least one parent with low hours (mostly mothers) also suggests that many families with children are organized so that at least one parent (mostly mothers) spends less time at a paid job.6
Families with very long work hours likely face extra challenges in balancing personal and work responsibilities, with the hours spent by both adults on the job leaving little time for family or personal duties. Who are these families? Clearly, those with both parents consistently putting in long work hours qualify, with 5,000 hours annually (100 hours per week) over five years. Both individuals are more likely to report higher levels of stress and suffer other adverse effects of long work hours. Arguably, families having at least one parent with fewer work hours should not be part of this definition as this parent has, at least in theory, more time available to compensate for the increased workload of the other parent. Similarly, families with two adults consistently working standard hours should also be excluded because individuals with standard hours tend not to exhibit higher levels of stress, and, despite the relatively high level, these hours are less variable year over year (Heisz and LaRochelle-Côté 2006), facilitating the dual management of work and family responsibilities.
According to the literature on work time, it appears reasonable to include families having at least one parent with long hours and the other with a consistently standard schedule—particularly families with children—in the long hours group, for several reasons. First, these families spend a considerable number of hours on the job (4,400 per year on average), which reduces the time available for parental duties and family activities (Curtis and Phipps 2000). Second, most families with two full-time, full-year paid jobs face a challenge with work–life balance as conflicting demands and role overload increase (Burton and Phipps 2007), with these likely to be particularly sensitive among families with children. Third, a parent with long hours may also affect the well-being of the other parent since these spouses, mainly women, see increased parental work (and stress) in response to work stress experienced by their partner (MacDonald et al. 2005 and Bolger et al. 1989). Finally, families with both parents working at least 45 hours per week (approximately 4,500 per year) can be described as very short of time (Burton and Phipps 2007), which reinforces the argument that these families face a particular challenge in maximizing their welfare due to time constraints.
For this study, ‘families with long hours' includes those with two adults working long hours as well as those with one adult working long hours and the other a consistently standard schedule. Based on this definition, 14% of families with children had particularly long hours (compared with 20% for age-adjusted families without children).
For simplicity, the remaining categories were also regrouped to create four categories of family work patterns. These categories accounted for the major differences shown in work patterns between families with children and without children. In addition to families with long hours, the categories were families with both adults consistently working standard hours; families with one parent working low hours and another with at least a standard schedule; and all other family work arrangements involving less than standard hours (Table 3).
Families with and without children showed substantial differences in work patterns. For instance, 14% of families with children worked long hours compared with 20% of those without. Furthermore, while 21% of families without children consistently worked standard hours, only 14% of families with children did so. Finally, 51% of all families with children were in the one low, ‘one at least' standard mould, compared with 41% of families without children—suggesting that the model whereby one parent has more time available for purposes other than work is common among families with children.7
If long hours do have a particular impact on the welfare of families with children, then there may be a negative association between long hours and the presence of children. While the average number of children under 18 was virtually identical by family work pattern (Table 4), differences were apparent in the proportion of families with young children (under age 6). More preschool children were in families with less intensive work patterns (15% to 17%) than in families with long hours (9%) or consistently standard hours (11%).
Since the presence of children may be related to other family or personal characteristics, a series of regressions were conducted to test the robustness of the association between the presence of children (including young children) and long family hours. Both the presence of children and young children were negatively associated with long hours when demographic characteristics were taken into account. The presence of children, but not young children, was negatively associated with consistently standard hours (see Children and family work patterns). Such results raise the possibility that families with children are less likely to choose situations that would expose them to long work hours and time-crunch issues. It also suggests that families with young children are particularly averse to long hours.
It is often argued that long hours are associated with detrimental effects on well-being, particularly for families with children. The association between well-being and hours by looking at the relationship between long family hours and various statistical indicators, and also by examining whether these indicators tend to be more significant when the focus is restricted to families with children.
A good starting point is the link between family hours and family earnings. The issue of time and money is a crucial one for families in general, and for families with children in particular. For instance, higher-income parents might be able to substitute money for their own time—at least partially—by hiring nannies or housekeepers (Burton and Phipps 2007). In other words, if families with long hours can generate more earnings from their longer work hours, then the welfare consequences of an elevated workload may be smaller.
Among families with children, those working long hours made significantly less money on average than consistently standard families, despite working 600 (or 15%) more hours—$86,500 per year on average, compared with $97,700 (Table 6). The difference was even larger at the 25th percentile, where families with long hours were worse off by $17,200. At the 75th percentile, however, earnings levels became similar.8
Such differences in earnings levels were not seen among families without children, even if similar differences were found in average hours across family work patterns. At first glance, the lower earnings of parents with the most hours compared with those working consistently standard hours appears counterintuitive. Some parents may have had to work long hours in order to maintain a minimum standard of living—they could not afford to reduce their hours. Such findings suggest that long-hour families with children do not necessarily have additional resources to better cope with work–life balance issues.
Other indicators can also be used to investigate the relationship between long hours and well-being. Job and occupation characteristics, in particular, can be related to differences in work time and have the potential to reveal information about family well-being (Heisz and LaRochelle-Côté 2006 and 2007). Differences between families with (and without) children across family work patterns could therefore reveal more about the preferences of families with children, and, by extension, their state of well-being.9 Since job information was available only for when individuals were employed, only the first three work-pattern categories were examined: families with long hours, consistently standard families, and families with one low, one at least standard parent (Table 7).
Job-quality indicators are used by many analysts to classify jobs as good or bad. Good jobs tend to have better pension and union coverage, and are more likely to be found in large firms. More particularly, good jobs also tend to be associated with stable, full-time hours, and bad jobs with more ‘unstable' work arrangements (Gunderson and Riddell 2000). In general, families with and without children were not significantly different in terms of job-quality indicators. However, fathers in families working long hours tended to be more unionized than their counterparts without children. Since unionized jobs tend to be more secure and associated with more predictable shifts, this may indicate that, given the long work hours, families with children are looking for more security and stability. It also suggests that parents may try to reduce the adverse effects of long work hours on their families.
Differences were also examined by occupation and industry (Table 8). Mothers in families working long hours were more likely than other women to work in the public sector. Since husbands typically spend the most time on the job in such families, mothers may be compensating for their husband's long hours by working in industries generally known for more stable schedules to ensure that one parent has hours that help them fulfill their parental duties. Furthermore, parents in families with long hours were also much more likely than non-parents to be self-employed. Among those with children, 31% of fathers and 24% mothers were self-employed, compared with just 22% and 10% of non-parents. Since the self-employed typically have more control over their schedules than paid employees, this may not be a surprise as parents with long hours may need more flexibility to deal with parental duties.10
Mothers in consistently standard families were much more likely than other women to be managers. This is not too surprising since consistently standard work still involves a large number of hours, which means these mothers may be more likely to need (or choose) to put in the hours for professional reasons.11
The results suggest that parents working long hours may respond to the presence of children by making different choices to reduce the welfare impact of long hours on the family. To test that hypothesis, an empirical strategy was needed to examine whether long work hours had different welfare implications on parents. Although Survey of Labour and Income Dynamics (SLID) does not provide much information on the state of family well-being, it does enquire about the general level of perceived stress. This measure is not perfect since stress can be caused by many factors not necessarily related to work hours. Furthermore, the direction of the causality is not always clear as work hours can cause stress, but stress can also affect work hours. The best that can be done is to develop a family measure of stress by using information on individual stress levels, and by assuming that a measure of family stress is a good proxy for family well-being.12 One measure used was the proportion of families in which both parents reported at least one episode of stress over the period (Table 9).13 As expected, families with long hours had significantly higher levels of stress (28%) than consistently standard couples (17%), and more than families with one low hours and another with at least standard hours (22%), although the latter difference was not significant.
However, a different picture emerged when family stress levels were examined separately for families with and without children. While families with long hours reported relatively high levels of stress even in the absence of children, consistently standard families with children were much more likely than those without children to report higher levels of stress (22% compared with 13%), suggesting that consistently standard families with children—who also spend a large number of hours in the labour market—also face well-being issues of their own.
Because stress levels can also be associated with other demographic and job characteristics, the robustness of the association between family stress and family work arrangements was tested with regressions that included a dummy variable to account for the presence of children and used families with consistently standard hours as a reference group. Once again, families with long hours were much more likely to be stressed than consistently standard families (Table 10). Families in the one low, one at least standard group were also more likely to be stressed than consistently standard families, albeit by a less significant margin.
After adding a dummy variable to account for children's interactions with family work patterns, both coefficients associated with work patterns remained positive and significant—especially in the case of families with long hours, indicating that these families experienced more stress than consistently standard families. However, the coefficient associated with the dummy variable for presence of children in consistently standard families was positive, indicating that those with children tended to report significantly higher levels of stress than those without children. Furthermore, the child interaction coefficients associated with one low, one at least standard families and with long hour families were not significant, which means that the presence of children did not seem to be associated with higher stress levels in these families. All coefficients stayed the same when demographic characteristics were taken into account, but the significance of the coefficient associated with long hours became lower when job characteristics were considered, which suggests that at least some of the stress experienced by families working long hours could be due to job factors.
The results imply that families working long hours typically experienced higher stress levels regardless of the circumstances. Consistently standard families reacted to the presence of children as they tended to report lower levels of stress than families with long hours in the absence of children, and similar levels as other family types in the presence of children.
That said, such findings require a word of caution. Parents working long hours may face well-being issues that are not necessarily captured by their stress levels. When working long hours, stress may also be different in the presence of children than in the absence of children. Clearly, additional work is required to better understand the well-being implications of work patterns on families with children. Ideally, a larger set of family well-being indicators should be applied to a reasonably large sample of families.
Over the past few decades, women increased their labour market participation substantially. While this is advantageous in a number of ways (higher family income, more equality between men and women), it also brings challenges as families might find it more difficult to reconcile work and family responsibilities—especially if both parents consistently work long hours year after year. This paper looked at the work patterns of families over five years. Families were grouped into four family work patterns: with long hours; with two adults consistently working on a standard basis; with one parent working short hours and the other at least a consistently standard schedule; and other patterns (with fewer family hours). Families with long hours had at least one adult with particularly long hours (at least once above 2,300 hours without ever going below 1,500 hours) and another with a consistently standard schedule (always between 1,500 and 2,300 hours or the equivalent). The rationale for this definition was that individuals in these families showed an increased tendency to have higher stress levels, and were therefore likely to face more work–life balance challenges.
Significant work-pattern differences were found between families with children and those without children. For instance, 14% of families with children under 18 were in the long-hours group, compared with 20% of families without children. Furthermore, families with children were much more likely to fall in the one low, one at least standard mould (51% versus 41% for families without children) and less likely to have two parents with consistently standard schedules (14% versus 21%). Long hours were also negatively associated with the presence of young children in the family.
Families with children might have different work patterns because of the well-being implications of working long hours. This paper examined the characteristics of families working long hours, and whether such characteristics differed from families without children. Families with parents working long hours were financially worse off than consistently standard parents even though they worked 15% more hours—a difference not seen among families without children. Families with children were more likely to work in unionized jobs (fathers), more likely to work in the public sector (mothers) and more likely to be self-employed (both), thereby increasing the possibility that their long hours were not always by choice, and, when facing the prospect of long hours, they organized themselves to reduce the negative impact.
This hypothesis was tested with a measure of family stress—defined as both adults reporting at least one episode of stress over the five-year period. While families with long hours were more stressed than other types of families, the presence of children did not appear to have much impact on their stress levels. Rather, the presence of children seemed to affect the stress of consistently standard families. This is not necessarily surprising. The marginal stress effect of children was probably lower among long-hour families since they already had high stress levels.
The longitudinal Survey of Labour and Income Dynamics (SLID) is conducted every year to collect information about income and labour market activity. Respondents are asked about hours usually worked at all jobs, which are then aggregated into annual paid hours. Paid hours include paid holidays, paid sick or maternity leave, and usual paid overtime. For example, an individual reporting 2,000 hours per year is typically working a 40-hour week, 52 weeks per year.
Since information on work hours was gathered for six years for all individuals age 16 and over, it was possible to create categories of long-term work patterns as suggested in Bluestone and Rose 1997. The work patterns of couples were then regrouped into family work patterns.
Three longitudinal panels (1996 to 2001, 1999 to 2004, and 2002 to 2007) were combined to create a sample of two-adult families with sufficient labour and demographic information for both in at least five of the six years.1 Families with missing information for two or more years were dropped from the sample and the weights of the remaining sample were adjusted to compensate.2 Because of the requirement for families to be in sample for all years, those that experienced a change in marital status (divorce, separation or death) also had to be excluded, but these amounted to a relatively small portion. Of the 8,800 families remaining in sample, approximately 4,800 had at least one child under age 18 in all six years (excluding children born over the period). As work patterns might have different implications for families with children, they are shown separately. Standard errors were generated using bootstrap weights.
To ensure that the association between work patterns and the presence of children was not due to other personal or family characteristics, a regression was designed to control for demographic characteristics that might affect work time patterns—a multinomial logit to determine the probability of being in one of the four family work patterns. The objective was to see if the relationship between the presence of children and certain family work patterns remained when all demographic characteristics were taken into account (Table 5).
The presence of children was negatively correlated with the probability of being in consistently standard families or in families with long work hours. However, after adding a dummy variable indicating the presence of young children, both child variables were negatively associated with the probability of being in consistently standard- or long-hour families—but the presence of young children was negatively correlated only with long hours. These results confirm that families may have a preference for fewer hours on the job when children—particularly young ones—are present, even after demographic and family characteristics are taken into account.
Anderson, Patricia M., Kristin F. Butcher and Phillip B. Levine. 2003. “Maternal employment and overweight children.” Journal of Health Economics. Vol. 22, no. 3. May. p. 477-504.
Bluestone, Barry and Stephen Rose. 1997. “Overworked and underemployed: Unraveling an economic enigma.” The American Prospect. Vol. 8, issue 31. March. 11 p.
Bolger, Niall, Anita DeLongis, Ronald C. Kessler and Elaine Wethington. 1989. “The contagion of stress across multiple roles.” Journal of Marriage and the Family. Vol. 51, no. 1. February. p. 175-183. (accessed August 5, 2009).
Burton, Peter and Shelley Phipps. 2007.“Families, time and money in Canada, Germany, Sweden, the United Kingdom and the United States.” Review of Income and Wealth. Series 53, no. 3. September. p. 460-483. (accessed August 5, 2009).
Curtis, Lori and Shelley Phipps. 2000. Economic Resources and Children's Health and Success at School: An Analysis Using the NLSCY. W-01-1-4E. Catalogue no. MP32-28/01-1-4E-IN. Working Paper Series. Applied Research Branch. Human Resources Development Canada. Ottawa. 37 p. (accessed August 5, 2009).
Gunderson, Morley and W.Craig Riddell. 2000. “The changing nature of work: Implications for public policy.” Adapting Public Policy to a Labour Market in Transition. W. Craig Riddell and France St-Hilaire (eds.). Chapter 1. Institute for Research on Public Policy (IRPP). Montreal. p. 1-37.
Hébert, Benoît-Paul and Alex Grey. 2006. “Time-related stress: Incidence and risk factors.” Horizons. Vol. 8, no. 3. April. Policy Research Initiative. p.14-18. (accessed August 5, 2009).
Heisz, Andrew and Sébastien Larochelle-Côté. 2007. Understanding Regional Differences in Work Hours. Statistics Canada Catalogue no. 11F0019MIE – no. 293. Analytical Studies Branch Research Paper Series. Ottawa. 48 p. (accessed August 5, 2009).
Heisz, Andrew and Sébastien Larochelle-Côté. 2006. Work Hours Instability in Canada. Statistics Canada Catalogue no. 11F0019MIE – No. 278. Analytical Studies Branch Research Paper Series. Ottawa. 38 p. (accessed August 5, 2009).
Higgins, Chris and Linda Duxbury. 2002. The 2001 National Work–Life Conflict Study: Report One. Report prepared for the Healthy Communities Division, Health Canada. Ottawa. 99 p. (accessed August 5, 2009).
MacDonald, Martha, Shelley Phipps and Linda Lethbridge. 2005. “Taking its toll: The influence of paid and unpaid work on women's well-being.” Feminist Economics. Vol. 11, no.1. March. p. 63-94.
Sébastien LaRochelle-Côté is with the Labour and Household Surveys Analysis Division. He can be reached at 613-951-0803. Claude Dionne is with the Income Statistics Division. He can be reached at 613-951-5043. All authors can be reached at firstname.lastname@example.org