Article body

This paper examines whether flexible work schedules in Canada are created by employers for business reasons or to assist their workers achieve work-life balance. We focus on long workweek, flextime, compressed workweek, variable workweek length and/or variable workweek schedule. Statistics Canada’s 2003 Workplace and Employee Survey data linking employee microdata to workplace (i.e., employer) microdata are used in the analysis. Results show that more than half of the workers covered in this data have at least one of the five specified types of flexible work schedules. Employment status, unionized work, occupation, and sector are factors consistently associated with flexible work schedules. Personal characteristics such as marital status, dependent children, and childcare use are not significantly associated with flexible work schedules, and females are less likely to have a flexible work schedule than are males. Overall, results suggest that flexible work schedules are created for business reasons rather than individual worker interests.This paper examines whether flexible work schedules in Canada are created by employers for business reasons or to assist their workers achieve work-life balance. We first explore the prevalence of different types of flexible work schedules in Canada, and then examine work, personal, human capital, workplace, and sector characteristics that might be associated with flexible work schedules. For flexible work schedules, we focus on having a long workweek, flextime, a compressed workweek, a variable workweek length and/or a variable workweek schedule.

The term “flexibility” is used frequently in all business circles, though it has a plethora of meaning and contexts (Blyton and Morris, 1991; Piore, 1986; Zeytinoglu, 1999). There are several reasons for this inconsistency in meaning: first, the range of subjects that it covers is substantial; second, the term means vastly different things to different constituencies, such as employers, unions, employees, governments, and intergovernmental organizations; and third, its use is often ideological, reflecting our views and perceptions of the “value” of flexibility (Atkinson, 1987).

In the last three decades, two streams of literature have emerged on flexibility. One stream of literature discusses flexibility as demand-driven; that is, a strategic initiative of employers to enhance the business requirements of the firm (Beechy and Perkins, 1987; Golden, 2005; Jenson, Hagen and Reddy, 1988; Zeytinoglu, 1999). The other stream of literature discusses flexibility as supply-driven, where employees have the ability to influence decisions about the nature of their work schedules and where employees, especially women, demand flexible work schedules for work-life balance (Bailyn, Drago and Kochan, 2001; Kropf, 1999; Marshall, 2001; Zeytinoglu, 1993). Thus, we ask are flexible work schedules created for business reasons or to assist workers achieve work-life balance?

The topic of this study is important for policy initiatives. Influential intergovernmental organizations, such as the Organisation for Economic Co-operation and Development (OECD, 2006) and the European Union (Goetschy, 2006) are promoting flexible work schedules as a significant employment strategy to ease the rigidity in labour markets and to assist in creating work-family balance for women. The International Labour Organisation is also supporting flexible work schedules that are family-friendly and promote gender equality (Messenger, 2006). In Canada, attracting and retaining female workers is a strategy promoted by the federal government, and a flexible work schedule is seen as a means to achieve that goal (Human Resources and Social Development Canada, 2006). The results of this study can assist policy makers by providing evidence on whether flexible work schedules are created for business reasons or for individual worker interests.

Theory and Empirical Literature

From a theoretical perspective, flexible work schedules can be discussed within the framework of industrial relations systems theory (Dunlop, 1958; Meltz, 1993) and its extension, the business strategy model (Kochan, Katz and McKersie, 1986). Both global and domestic work environments influence companies’ business strategies, resulting in the offering of a variety of flexible employment contracts (Cappelli et al., 1997). Employers use a core of relatively permanent full-time workers, that is, standard workers, and use a number of flexible, non-standard workers in the periphery. Non-standard employment is seen by companies as a business strategy to achieve flexibility in hiring/firing decisions and for cost effectiveness, while continuity in business operations is accomplished by using standard workers (Houseman, 2001). Our study builds on this flexibility concept and extends it by focusing on the scheduling aspect of flexibility.

Using Zeytinoglu’s (1999) typology of employment contracts, we categorize all types of work based on two dimensions: regularity (or continuity) of employment contract and hours of work. Employment status is categorized as standard (regular full-time) and non-standard (regular part-time, temporary full-time and/or temporary part-time). We argue that the four types of employment status include flextime, compressed workweek, variable workweek length and/or variable workweek schedule, but the long workweek can only occur within the regular or temporary full-time employment status. Research suggests that temporary employment status is more commonly associated with flexible work schedules (Bosch, 2006).

In addition to employment status, the empirical literature suggests a number of work, personal, human capital, workplace, and sector characteristics that are associated with flexible work schedules. For work characteristics, we focus on whether the job is unionized or not. Many workers in Canada are not unionized, and workers in non-standard work have lower unionization rates than workers in standard employment status. There is ample research showing the benefits of unionization for workers (Verma, 2005) including negotiating for standard work schedules and limits to flexible work (Golden, 2005; Zeytinoglu, 1993).

In terms of personal characteristics associated with flexible work schedules, gender, marital status, dependent children and age are discussed in the literature. Female workers are inevitably linked to the flexible work schedules debate. There is a plethora of research on work pressures and interest in flexible work schedules for work-family balance for female workers (Duxbury and Higgins, 2001; Parasuraman and Simmers, 2001). However, empirical research shows that women, including mothers with young children, are working longer hours than in the past, although still less than men (Usalcas, 2008), and have less flexible work schedules than men (Comfort, Johnson and Wallace, 2003; Golden, 2005; McCrate, 2005). In terms of age, recent analysis shows that younger workers are more likely than middle-aged and older workers to have a flexible work schedule (Cooke, 2007).

For human capital characteristics, research shows that education affects work hours and those with post-secondary education tend to have more standard work schedules (Cooke, 2007; Usalcas, 2008). The other human capital variables of occupation and wage level are, in a way, a proxy for power and control on the job. Research suggests that the autonomy and power of the individual to control their work schedules can dictate whether one has a flexible work schedule or not (MacDermid, 2005). Those in positions highly demanded by the employer will have the power to negotiate flexible work schedules for work-family balance (Kossek, Lautsch and Eaton, 2005). Research shows that those in managerial and professional occupations tend to have better access to flexible work schedules than those in blue-collar and service jobs (Comfort, Johnson and Wallace, 2003; Golden, 2005; McCrate, 2005).

In terms of workplace characteristics, the size of the workplace is shown to be associated with flexible work schedules in a number of studies but results are inconclusive. Some studies suggest flexible work schedules exist in larger workplaces (Chaykowski and Slotsve, 2003), while others show smaller workplaces are more likely to introduce a number of flexible work schedules (Houseman, 2001).

Sector is another characteristic shown to be associated with flexible work schedules. Generally speaking, flexible work schedules exist more in the service sector whereas the manufacturing sector has more structured work schedules, long workweeks, and less room for flexibility due to business reasons (Usalcas, 2008). In a recent study, Levesque (2007) showed that the health care and education sectors have more flexible work schedules than other sectors.

As Lewis, Gambles and Rapoport (2007) argue, the discussion surrounding flexible work schedules is essentially two distinct yet discernable and overlapping discourses. One focuses on the individual and personal control of time and work, and the other focuses on workplace flexibility as decided by the employer according to business needs and goals. The effects of flexible work schedules on workers range from potentially very favourable to very unfavourable, depending on the nature of those schedules and the preferences of the affected workers (see Baltes et al., 1999; Boulin, Lallement and Michon, 2006; Cooke, 2005; Lee and McCann, 2006). For example, for a worker, flextime can be favourable for personal reasons such as having dependent children and/or needing to come to work early or leave early. A compressed workweek is most often created for business reasons but it can also be favourable for workers if it is their choice to work longer hours in a workday in order to have more uninterrupted personal time. Long workweeks are primarily created for business interests and are unfavourable for workers interested in more time for personal reasons; however, they can be favourable if workers are interested in earning more money. Variable workweek length and schedule can be unfavourable for those who seek schedule regularity, but can be favourable for those who prefer to decide when and how long to work, when the option is offered. Research shows that flexible work schedules can be implemented by employers primarily to address operational needs, employee preferences, or a combination of the two (Boulin, Lallement and Michon, 2006; Cooke, 2005).

Based on the theory and literature reviewed, we argue that if we find work, workplace, and sector variables to be associated with flexible work schedules, it suggests that employers are creating flexible work schedules for business reasons rather than for individual worker interests. On the other hand, if we find personal characteristics, such as gender, marital status and/or dependent children to be associated with flexible work schedules, it suggests that employers create flexible work schedules for individual worker interest. If we find significance of personal characteristics along with human capital variables, these suggest that employees have increased control over their work schedules and employers are accommodating their interests.

Methodology

Data and Sample

The study uses Statistics Canada’s linked Workplace and Employee Survey (WES) 2003 microdata, which include 20,834 employees from among 6,565 workplaces, based on a response rate of 82.7% and 83.1% respectively. On a weighted basis, this equates to 12.1 million respondents. The unit of analysis in this paper is the individual worker. The data, which contain linked responses from employers and their employees, are well suited for this study due to the large number of work, personal, human capital, and workplace variables. The WES surveys all business locations operating in Canada except employers in Yukon, Nunavut and Northwest Territories, and employers operating in crop production, animal production, fishing, hunting and trapping, private households, religious organizations and public administration. The weighted microdata and the bootstrapping technique allow us to generalize from the results to the Canadian labour market, subject to exclusions listed (for more on sampling and sample design, see Statistics Canada, 2006).

Variables

The dependent variables are long workweek, flextime, compressed workweek, variable workweek length, variable workweek schedule, and those having a standard work schedule (and each coded as 1 = yes, 0 = otherwise). Having a workweek averaging at least 48 hours per week is a long workweek. Working a certain number of core hours but being able to vary start and stop times while working the equivalent of a full workweek is flextime. Working longer hours each day to reduce the number of days in a workweek is a compressed workweek. Variable workweek length consists of having a workweek length that varies weekly, that is, normally not working the same number of paid hours per week. Having a work schedule in which the hours and/or days of the week vary is considered to be a variable workweek schedule. A standard work schedule is not having any of the above characteristics, and thus working between 30–48 hours per week with a fixed schedule of work hours and workdays.

The independent variables are work characteristics, including employment status and unionized; personal characteristics of gender, marital status, presence of dependent children, use of childcare, and age; and human capital characteristics including full-time work experience, education, occupation, and wage level. The workplace variable is size (in number of employees). Sector refers to the sector in which the workplace operates.

The work characteristics of employment status consists of four categories: regular full-time (employment contract with no contractual or anticipated termination date and working 30 hours or more per week), regular part-time (similar employment contract but working less than 30 hours per week), temporary full-time (seasonal, casual, on-call, or term employment contract with specified termination date and working 30 hours or more per week), and temporary part-time (similar employment contract but working less than 30 hours per week), each coded as 1 = yes, 0 = otherwise with regular full-time as the reference category. Unionized (being a member of a union or covered by a collective agreement) is coded as 1 = yes, 0 = no.

All personal characteristics are coded into dummy variables: gender (1 = female, 0 = male), marital status (married or in a common law relationship = 1, single, separated, divorced, or widowed = 0), presence of dependent child(ren) (1 = yes, 0 = no), use of childcare (have children in the care of someone other than the respondent or another legal guardian [excluding regular school hours] 1 = yes, 0 = no), and age (three categories of < 30, 30–50 [reference group], 50+), each coded as 1 = yes, 0 = otherwise.

The human capital characteristic of full-time work experience is measured in number of years (and a squared version is also included in the multivariate analysis). Other human capital characteristics are education (categorized as at most high school education [reference group], some post-secondary education, and university degree), each coded as 1 = yes, 0 = otherwise; occupation (consisting of managers [reference group], professionals, lower white-collar such as marketing/sales or clerical/administrative, and blue-collar such as technical/trades or production workers degree), each coded as 1 = yes, 0 = no; and wage levels (< $10 [reference group], $10–$19.99, $20–$39.99, and $40+), each coded as 1 = yes, 0 = no.

The only workplace characteristic we use is workplace size (referring to the number of employees, although a log form is used in the multivariate analysis). Sector is coded as dummy variables consisting of primary and manufacturing, construction and transportation (reference group), education and health, and other service sector.

Analysis

The analysis consists of several sets of descriptive statistics followed by correlations between variables (not presented here due to space limitations but available upon request) and multivariate (logistic regression) analyses. For each regression model, the odds ratio and significance level of each variable are provided. In addition, two goodness of fit measures (i.e., the pseudo R2 and Wald chi-square) are presented for each iteration. All of the analyses have been generated using weighted microdata accessed at the Statistics Canada Research Data Centre (RDC) of McMaster University. Due to the complex survey design of the WES dataset, bootstrapping is used in the multivariate analyses. Bootstrapping refers to a process of repeatedly drawing random samples, with replacement, from the data at hand (Hamilton, 2003). Bootstrapping involves “resampling” the data with replacement many, many times in order to generate an empirical estimate of the entire sampling distribution of a statistic. Any data that involve complex surveying methodology requires special analytic consideration of bootstrapping (Mooney and Duval, 1993). All presented regression results have been bootstrapped using Statistics Canada’s recommended set of weights via the Stata function developed and discussed by Chowhan and Buckley (2005). All descriptive statistics are weighted as recommended.

On a technical note, although some of the independent and/or control variables in our regression models are positively correlated, multicollinearity is not an issue in our analyses. For instance, while education and occupation are inter-correlated, the absolute magnitude of bivariate correlations between pairs of our independent and/or control variables tend to be below .3, with very few exceptions. Moreover, the odds ratios for occupation were not tangibly different when generating alternative regression models that excluded education, wage level, or employment status variables, for example. The one exception is worker age and years of full-time work experience. These are, not surprisingly, strongly correlated. However, there are conceptual reasons for including each of these two (sets of) variables in the regressions. Moreover, the resulting odds ratios for these variables are not counterintuitive, which would have been a sign of multicollinearity.

Sample Characteristics

Approximately 5% of workers have a long workweek, 36% have flextime, 7% a compressed workweek, 13% a variable workweek length, and 16% a variable workweek schedule. Only two in five (43%) Canadians have a standard work schedule. Other studies also show that flexible work schedules are extensive in Canada, with less than one third of Canadian workers employed in full-time, continuous employment with a stable work schedule (Lipsett and Reesor, 1998). Using a more restrictive scope, when considering only short or long workweeks, non-permanent employment, and/or home-based work, Cooke (2005) showed that in 1999, 42% of Canadian workers had a flexible work arrangement. The Comfort, Johnson and Wallace (2003) study found that 40% of Canadian workers use flextime, and Hall (1999) and Shields (2000) showed that a sizable proportion of Canadians have work hours that exceed “normal” full-time hours. Descriptive statistics for these and all other variables are presented in Table 1.

Results

Descriptive Statistics of Flexible Work Schedules

Figure 1 shows each type of flexible work schedule among workers sorted by employment status. Among workers with regular full-time employment status, 6% have a long workweek, 34% have flextime, and 7% have a compressed workweek. About 8% and 12% have a variable workweek length or a variable workweek schedule, respectively, while more than half of workers with regular full-time employment status (52%) have a standard work schedule.

Compared to those in regular full-time employment, those with part-time employment status are much more likely to have flextime (42%), a variable workweek length (26%), or a variable workweek schedule (32%), but are less likely to have a compressed workweek (5%). Although having a compressed workweek is sometimes perceived to be possible for full-time workers only, the definition in our dataset does not preclude a part-time worker from having this arrangement. As an example, a worker with a 24-hour workweek might arrange with their employer to work three eight-hour shifts rather than four six-hour shifts. Since this variable is self-reported, the worker is the one indicating whether or not they have a compressed workweek. Nonetheless, it is unsurprising that compressed workweek is less prevalent among part-time workers because part-time schedules inherently involve fewer workdays per week in most cases. Thus, it is difficult to further shorten the workweek given a particular number of weekly hours. We also remind readers that workers in part-time hours—whether regular or temporary—cannot have a long workweek, nor a standard work schedule. On the whole, among workers with regular employment status, those with part-time hours are more likely to have other flexible work scheduling components in their job, relative to those with full-time hours.

Table 1

Descriptive Statistics of All Variables Used in the Study

 

Proportion / Mean

Std. Dev.

Dependent Variables

 

 

Long workweek

5.1

-

Flextime

36.2

-

Compressed workweek

6.6

-

Variable workweek length

13.0

-

Variable workweek schedule

16.2

-

Standard work schedule

43.0

-

Independent Variables

 

 

Regular full-time

80.5

-

Regular part-time

11.0

-

Temporary full-time

4.1

-

Temporary part-time

4.4

-

Unionized

25.7

-

Gender (i.e., female)

53.1

-

Marital status:

 

 

   Married/common-law

68.3

-

   Single/separated/divorced/widowed

31.7

-

Dependent children

46.6

-

Child care

8.6

-

Age:

 

 

   < 30

23.3

-

   30–50

56.7

-

   50 +

20.0

-

Full-time work experience

17.2

11.2

Education:

 

 

   At most high school education

29.7

-

   Some post-secondary education

51.3

-

   Has university degree

19.0

-

Occupation:

 

 

   Manager

12.8

-

   Professional

16.1

-

   Lower white-collar

22.9

-

   Blue-collar

48.1

-

Wage levels:

 

 

   < $10

14.1

-

   $10–$19.99

44.2

-

   $20–$39.99

35.8

-

   ≥ $40

6.0

-

Workplace size (in number of employees)

414.2

1085.4

Sector:

 

 

   Primary and manufacturing

18.2

-

   Construction and transportation

14.6

-

   Education and health

21.0

-

   Other service sector

46.1

-

Weighted sample size: 12,119,794

 

 

Sample: All workers.

-> See the list of tables

Workers in temporary full-time employment are much more likely than workers in regular full-time employment to have a long workweek (13%), flextime (46%), a variable workweek length (33%), or a variable workweek schedule (26%), and are somewhat more likely to have a compressed workweek (13%). Conversely, less than one third (32%) of workers in temporary full-time employment have a standard work schedule. Thus, a distinct pattern emerges in which those with temporary full-time employment status are much more likely to have one or more type of flexible work schedule than those with regular full-time employment status.

Finally, we consider those with a temporary part-time employment status. These workers are more likely to have flexible work schedules, relative to their regular counterparts, except that having a compressed workweek is less prevalent (3%). More than half of workers in temporary part-time employment have a variable workweek length (53%), over half have a variable workweek schedule (52%), and close to half have flextime (44%).

Figure 1

Proportions of Flexible Work Schedules Among Workers Sorted by Employment Status

Proportions of Flexible Work Schedules Among Workers Sorted by Employment Status

Note: By definition, a part-time worker cannot have a long workweek or standard work schedule. Thus, the proportion is zero for both of the flexiblework schedules for the two sub-samples of part-time workers.

-> See the list of figures

Regression Results

Six sets of regression results are presented in Table 2, one for each of the five types of flexible work schedules and one for a standard work schedule. Since several models are presented, the discussion of results for each is necessarily brief. As such, focus is limited to those with strongly statistically significant variables (at the p < .05 or p < .01 levels).

Table 2

Factors Associated with Flexible Work Schedules (Multivariate Regression Results)

Dependent Variables

Long workweek

Flextime

Compressed workweek

 

Odds

 

Odds

 

Odds

 

 

Ratio

Sig.

Ratio

Sig.

Ratio

Sig.

Independent Variables

 

 

 

 

 

 

Regular full-time

ref.

 

ref.

 

ref.

 

Regular part-time

N/A

 

1.425

***

0.739

 

Temporary full-time

2.681

***

1.488

***

1.384

 

Temporary part-time

N/A

 

1.532

**

0.516

*

Unionized

0.534

***

0.704

***

1.126

 

Gender (i.e., female)

0.253

***

0.748

***

1.032

 

Marital status:

 

 

 

 

 

 

   Married/common-law

ref.

 

ref.

 

ref.

 

Single/separated/divorced/widowed

1.202

 

0.983

 

0.955

 

Dependent children

1.064

 

0.985

 

1.109

 

Child care

0.786

 

1.038

 

0.898

 

Age:

 

 

 

 

 

 

   < 30

0.523

**

1.074

 

1.281

 

   30–50

ref.

 

ref.

 

ref.

 

   50 +

1.049

 

0.931

 

1.148

 

Full-time work experience

1.006

 

0.984

 

1.003

 

Full-time work exp. squared

1.000

 

1.000

 

1.000

 

Education:

 

 

 

 

 

 

   At most high school

ref.

 

ref.

 

ref.

 

   Some post-secondary

0.718

*

0.902

 

0.970

 

   Has university degree

1.165

 

0.974

 

0.882

 

Occupation:

 

 

 

 

 

 

   Manager

2.279

***

1.466

***

0.342

***

   Professional

1.335

 

1.358

**

0.668

***

   Lower white-collar

0.225

***

1.012

 

0.386

***

   Blue-collar

ref.

 

ref.

 

ref.

 

Wage levels:

 

 

 

 

 

 

   < $10

ref.

 

ref.

 

ref.

 

   $10–$19.99

0.849

 

0.825

 

0.677

 

   $20–$39.99

0.740

 

1.111

 

0.861

 

   ≥ $40

0.781

 

1.511

*

0.505

*

Workplace size (logform)

0.951

 

0.962

*

1.032

 

Sector:

 

 

 

 

 

 

   Primary and manufacturing

ref.

 

ref.

 

ref.

 

   Cons. and transportation

1.441

*

1.272

**

0.521

***

   Education and health

0.737

 

1.366

*

0.627

**

   Other service sector

0.946

 

1.771

***

0.729

*

   Number of Observations

20,362

 

20,362

 

20,362

 

      Wald Chi-Square

289.740

 

230.830

 

153.020

 

      Prob > Wald

0.000

 

0.000

 

0.000

 

      Pseudo R-Square

0.145

 

0.040

 

0.040

 

Significance levels: * p < .10, ** p < .05, *** p < .01

Sample: All workers.

N/A is not applicable because this schedule is defined to exclude part-time hours.

-> See the list of tables

Table 2 continues

Factors Associated with Flexible Work Schedules (Multivariate Regression Results)

Dependent Variables

Variable

workweek length

Variable

workweek schedule

Standard

work schedule

 

 

Odds

 

Odds

 

Odds

 

 

Ratio

Sig.

Ratio

Sig.

Ratio

Sig.

Independent Variables

 

 

 

 

 

 

Regular full-time

ref.

 

ref.

 

ref.

 

Regular part-time

3.529

***

2.678

***

N/A

 

Temporary full-time

4.786

***

2.300

***

0.647

***

Temporary part-time

13.062

***

6.259

***

N/A

 

Unionized

1.284

*

1.929

***

0.980

 

Gender (i.e., female)

0.932

 

1.002

 

1.145

**

Marital status:

 

 

 

 

 

 

   Married/common-law

ref.

 

ref.

 

ref.

 

Single/separated/divorced/widowed

1.079

 

0.982

 

0.949

 

Dependent children

1.134

 

0.933

 

0.855

*

Child care

1.136

 

1.050

 

1.317

**

Age:

 

 

 

 

 

 

   < 30

1.274

 

1.063

 

1.091

 

   30–50

ref.

 

ref.

 

ref.

 

   50 +

1.422

*

1.279

*

0.910

 

Full-time work experience

0.989

 

0.990

 

1.075

***

Full-time work exp. squared

1.000

 

1.000

 

0.999

***

Education:

 

 

 

 

 

 

   At most high school

ref.

 

ref.

 

ref.

 

   Some post-secondary

0.826

 

0.883

 

1.127

*

   Has university degree

0.737

**

0.627

**

1.154

 

Occupation:

 

 

 

 

 

 

   Manager

1.441

*

0.791

 

0.863

 

   Professional

0.944

 

0.761

*

1.022

 

   Lower white-collar

0.791

 

0.902

 

1.121

 

   Blue-collar

ref.

 

ref.

 

ref.

 

Wage levels:

 

 

 

 

 

 

   < $10

ref.

 

ref.

 

ref.

 

   $10–$19.99

0.695

***

0.717

**

2.057

***

   $20–$39.99

0.871

 

0.772

 

1.630

***

   ≥ $40

1.626

 

0.565

*

1.026

 

Workplace size (logform)

0.970

 

1.078

***

1.038

*

Sector:

 

 

 

 

 

 

   Primary and manufacturing

ref.

 

ref.

 

ref.

 

   Cons. and transportation

1.532

**

0.936

 

0.784

**

Education and health

1.397

*

1.778

***

0.475

***

   Other service sector

1.315

*

1.616

***

0.557

***

   Number of Observations

20,362

 

20,362

 

20,362

 

      Wald Chi-Square

428.400

 

455.170

 

211.230

 

      Prob > Wald

0.000

 

0.000

 

0.000

 

      Pseudo R-Square

0.124

 

0.108

 

0.044

 

Significance levels: * p < .10, ** p < .05, *** p < .01

Sample: All workers.

N/A is not applicable because this schedule is defined to exclude part-time hours.

-> See the list of tables

Long workweek. The variables with a strong statistical relationship with a long workweek are employment status, unionized work, gender, age, and occupation. Controlling for other factors, compared to those in regular full-time employment, workers in temporary full-time employment are more than twice as likely to have a long workweek. Unionized and younger workers are about half as likely to have a long workweek, relative to non-union and middle-aged workers, respectively. Moreover, females and those in lower white-collar occupations are 75% and 78% less likely to have a long workweek relative to males and those in blue-collar occupations, respectively. Those in managerial occupations, as compared to blue-collar workers, are more than twice as likely to have a long workweek.

Flextime. The variables strongly and significantly related to flextime are employment status, unionized work, gender, occupation, and sector. Workers with regular part-time, temporary full-time or temporary part-time employment status are at least 42% more likely than regular full-time workers to have a flextime schedule when controlling for other factors. There is also a higher likelihood that managers and professionals have a flextime schedule relative to blue-collar workers. Also, workers in the construction and transportation sector and in other service sectors are at least one quarter more likely to have a flextime schedule relative to those in the primary and manufacturing sector. That said, unionized workers and female workers are 30% and 25% less likely to have a flextime schedule compared to their non-union and male counterparts, respectively.

Compressed workweek. Occupation and sector are the only two factors significantly associated with a compressed workweek. In particular, professionals are about one third less likely to have a compressed workweek relative to blue-collar workers, while managers and lower white-collar workers are about two thirds less likely to have this schedule. Relative to those in the primary and manufacturing sector, workers in the construction and transportation, and education and health sectors are at least one third less likely to have a compressed workweek.

Variable workweek length. Employment status, education, wage levels, and sector are significantly associated with the likelihood of having a variable workweek length. Those with regular part-time, temporary full-time or temporary part-time employment status are between 2.5 to 12 times more likely to have a variable workweek length relative to regular full-time workers. Construction and transportation sector workers are also about 50% more likely to have this type of flexible work schedule compared to primary and manufacturing sector workers. Additionally, those with a university degree are 26% less likely to have a variable workweek length schedule compared to workers with no more than a high school education, while those in the $10–$19.99 hourly wage level are 30% less likely to have a variable workweek length schedule relative to those earning less than $10 per hour.

Variable workweek schedule. Employment status, unionized work, education, wage levels, workplace size, and sector are the factors significantly associated with a variable work schedule. Those with regular part-time, temporary full-time, or temporary part-time employment status are between 1.3 to 5.2 times more likely to have a variable workweek schedule relative to those in regular full-time employment. Unionized workers are almost twice as likely as their non-union counterparts to have a variable workweek schedule. Workers with a university degree are about one third less likely to have a variable workweek schedule relative to those with at most high school education. Those in the $10–$19.99 wage level are 28% less likely to have a variable workweek schedule relative to those earning less than $10 per hour. Workers in larger workplaces are more likely to have a variable workweek length compared to those in smaller workplaces. Also, those in education and health or other service sectors are at least 62% more likely to have a variable workweek schedule than those in the primary and manufacturing sector.

Standard work schedule. Finally, we examined factors associated with having a standard work schedule since this is the base work schedule to which all other (i.e., flexible) work schedules are implicitly compared. Temporary full-time workers are significantly less likely to have a standard work schedule compared to regular full-time workers by a factor of one third. We remind the readers that those with either regular part-time or temporary part-time employment status cannot have a standard work schedule by definition. Controlling for other factors, females are about 15% more likely than males to have a standard work schedule, while those using formal childcare are 32% more likely to have a standard work schedule. Full-time work experience is positively related to having a standard work schedule. Hourly wage levels show that those earning between $10 and $39.99 per hour are much more likely to have a standard work schedule relative to those earning less than $10 per hour. Finally, those in the construction and transportation sector, education and health care sector, and other service sector are about 20–50% less likely than those in the primary and manufacturing sector to have a standard work schedule.

Conclusions and Discussion

Results show that flexible work schedules are prevalent in Canada with 57% of working Canadians having some type of flexible work schedule. The most common type is flextime, followed by a variable workweek schedule or length, a compressed workweek, and a long workweek. These findings are in-line with other studies showing that a sizeable proportion of Canadian workers have some type of flexible working arrangement (e.g., Comfort, Johnson and Wallace, 2003; Hall, 1999; Shields, 2000).

An important contribution of this study to the literature is that it examines a variety of work, personal, human capital, workplace, and sector characteristics as possible factors associated with different types of flexible work schedules. Results show that employment status is a factor consistently associated with flexible work schedules. Compared to regular full-time workers, those in regular part-time employment are more likely to have a flextime schedule, a variable workweek length and/or a variable workweek schedule. Those with temporary full-time employment status are more likely to have a long workweek, flextime, a variable workweek length, and variable workweek schedule, and less likely to have a standard work schedule. Workers with temporary part-time employment status in particular are the most likely to have flextime, a variable workweek length and/or a variable workweek schedule. These results confirm and extend previous research that showed a positive relationship between flexible work schedules and non-standard employment status (Bosch, 2006).

Results also show that unionized work is another factor associated with flexible work schedules. Unionized workers are less likely to work in long workweeks or have flextime. These results are in-line with previous research showing that unions negotiate improvement in working conditions (Verma, 2005) including limits to flexible work schedules (Golden, 2005). Though, the results show that unionized workers are more likely to work in a variable workweek length and/or a schedule.

The results also show that the sector the workplace is in has a significant association with flexible work schedules. Compared to those in the primary and manufacturing sector, workers in the construction and transportation sector, education and health, and other service sectors are more likely to have flextime, a variable workweek length and/or a variable workweek schedule, and less likely to have a compressed workweek. Workers in the construction and transportation sector are more likely to have a long workweek than those in the primary and manufacturing sector.

In terms of personal characteristics, marital status, having dependent children or using childcare are not significantly associated with having flexible work schedules. Moreover, controlling for other factors, the gender variable shows that, contrary to the commonly held views that women work in flexible work schedules for work-family balance interests, it is men rather than women who tend to have flexible work schedules such as flextime. Controlling for other factors, our results show that women are more likely to have a standard work schedule, less likely to work long workweeks, and that gender is not significantly associated with a variable workweek length or schedule. These findings are not unique to Canadian workers covered in the WES dataset. Similar results exist for US workers as well (Golden, 2005; McCrate, 2005).

Of the human capital factors, occupation is significant. Specifically, managers and professionals are more likely to have a flextime schedule and less likely to have a compressed workweek. Others using US data (Golden, 2005; McCrate, 2005) and an earlier year of WES data in Canada (Comfort, Johnson and Wallace, 2003) show similar results for flextime. As suggested in the literature, those in positions of autonomy and control have the power to negotiate flexible work schedules that fit their needs (Kossek, Lautsch and Eaton, 2005; MacDermid, 2005). Managers and professionals tend to have more control over their work hours and schedules, and our results show that these employees are the ones with favourable flexible work schedules such as flextime. They are also likely to avoid flexible work schedules that are potentially less desirable for most workers such as a compressed workweek. Flextime is discussed in the literature as a favourable flexible work schedule for workers and compressed workweek is favoured mostly by employers (Baltes et al., 1999; Boulin, Lallement and Michon, 2006; Lee and McCann, 2006; Lewis, Gambles and Rapoport, 2007). However, business norms and structures still dictate that managers work long workweeks, as shown in our results.

Taken together, the significance of some factors and lack of associations found with other factors suggest that flexible work schedules are primarily created for business reasons as the industrial relations systems theory (Dunlop, 1958; Meltz, 1993) and the business strategy model (Kochan, Katz and McKersie, 1986) indicate, rather than assisting females, married or common-law-relationship workers, and workers with dependent children to achieve work-life balance interests. Similar to the discussions in the literature that flexibility is demand-driven, that is, a strategic initiative of employers to enhance the business requirements of the firm (Beechy and Perkins, 1987; Golden, 2005; Jenson, Hagen and Reddy, 1988; Zeytinoglu, 1999), findings of our study suggest that the choice for creating flexible work schedules is the employers’ and they make the choice for business reasons rather than for individual worker interests.

These findings can have implications for policy development if employers create flexible work schedules for business reasons, and not in an effort to assist workers, particularly women, with work-family balance concerns. Thus, flexible work schedules that are being promoted by influential intergovernmental organizations such as the OECD (2006) and the EU (Goetschy, 2006) are unlikely to achieve the dual goal of providing flexibility to employers and, at the same time, work-family-life balance for workers, as our results show. We recommend that governments develop separate policies responding to workers’ work-family-life balance interests.

It is important to note several limitations of our study, mostly stemming from the limitations of the data. First, although the WES represents the Canadian labour market, there are some specified exceptions in the data coverage. Particularly, the absence of workers in the public administration; that is, federal and provincial governments and municipalities, means the flexible work schedule experiences in this large sector are not reflected in our results. Second, the WES seems to under-represent individuals with non-permanent employment status because only employees receiving T4 slips from the establishment are eligible for sampling under the WES methodology. Thus, agency temporary workers are only included if the agency itself is included as an employer. Moreover, it appears that some casual and/or on-call workers classify themselves to be “regular” employees, even though they are more accurately categorized as having a type of non-permanent employment. Thus, the proportion of temporary workers in our study is somewhat lower than other estimates of the Canadian labour market. Third, without specific questions on who chooses to work in a flexible work schedule and who decides who will work in what type of flexible work schedule, we had to use a number of proxy variables. Further studies are recommended to be more specific in their focus on how flexible work schedule decisions are made in workplaces.

Our study is one of the first to examine the flexible work schedule issue both comprehensively and with a broad representative coverage and as such, we recommend further research on the topic. In particular, we recommend further attention be paid to other emerging types of work schedules beyond those included in our typology. We also suggest that future research pay more attention to employers as they design working conditions in general and work schedules in particular. For example, it remains an open question as to whether Canadian employers utilize flexible work schedules (and employment statuses) primarily out of financial or operational necessity or primarily due to strategic considerations. Additionally, we recommend the impact of flexible work schedules on employees to be examined in more detail.