Corps de l’article

Work and family demands on women are well reported (Korabik, Lero and Whitehead, 2008). Feminization of part-time work and the lower level of benefits provided to part-time workers as compared to full timers are also reported in earlier studies (Zeytinoglu and Cooke, 2005). It is well known that many women voluntarily choose to work part-time because of family responsibilities (Lero and Lewis, 2008; Marshall, 2001; Tomlinson, 2007) and there are some others, both female and male, in involuntary part-time jobs (Marshall, 2001). With the population aging in most industrialized countries, governments and intergovernmental organizations are encouraging women’s attachment to the labour force, particularly through part-time employment (Human Resources and Social Development Canada, 2006; OECD, 2006; Goetschy, 2006). There seems to be the impression in the popular press that employers are providing extensive support to retain their staff (see, for example, Galt, 2007), though some academic research suggests that this probably occurs only for managers and professionals (Lewis, Gambles and Rapoport, 2007), and even for them such a balance might not exist (Burchielli, Bartram and Thanacoody, 2008).

These issues led us to ask following questions: What are employers doing to support workers? Do employers offer family support programs to assist their employees? Are the gender of the employee and their employment status (full-time, voluntary or involuntary part-time) associated with employer offered support programs? To answer these questions we take advantage of unique Canadian data, namely Statistics Canada’s 2003 Workplace and Employee Survey (WES). We examine the extent of employer offered family support programs in Canada and whether there are associations between these programs, gender and employment status.

In the literature, employer offered family support programs refer to “deliberate organizational changes – in policies, practices, or the target culture – to reduce work-family conflict and/or support employees’ lives outside of work” (Kelly et al., 2008: 310). These are, for example, family leaves, flexible work arrangements including part-time work, dependent care supports and organizational and supervisory support (Higgins, Duxbury and Lyons, 2008; Kelly et al., 2008). Employer offered family support programs can be initiated and introduced in the workplace only if the employer decides to provide these benefits to employees, though workers’ and/or union interests can also be a catalyst for their development. In this study, dependent care supports are the focus of interest. Employer offered family support programs are defined as childcare and eldercare support. In the data (the WES), employer offered childcare support refers to help through an on-site childcare centre, assistance with external suppliers, or assistance with informal arrangements. Employer offer for eldercare support is defined as employer offering help with eldercare services.

Childcare and eldercare needs are significant factors contributing to work-family conflict in Canada, the US and Europe (Higgins, Duxbury and Lyons, 2008; Lero and Lewis, 2008). Employer offered support programs can increase job satisfaction and commitment, reduce turnover and absenteeism, and improve employee morale and productivity (Kelly et al., 2008). Researchers question how much employers are doing to support work-family balance (Burchielli, Bartram and Tanacoody, 2008). The issues of childcare and eldercare are under-analyzed and under-discussed in the context of industrial relations. Thus, this study is important because it offers new insights on an under-researched area with a dataset that allows us to generalize from the results to the Canadian labour market, subject to some exceptions (explained below). The results of this study will contribute to our understanding of the extent of employer support for childcare and eldercare programs and provide evidence for public policy initiatives on work-life balance.

Theoretical Foundations of the Study and Empirical Research

Work-life balance is an integration of essentially two discernable, overlapping but distinct discourses. One focuses on the individual and personal control of time and the other on the workplace flexibility work-life balance discourse (Lewis, Gambles and Rapoport, 2007). The former, with its focus on the individual and the family level, applies to “relatively affluent professional and white-collar workers – both men and women – especially in the knowledge economy, who have difficulty finding time for personal life” (Lewis, Gambles and Rapoport, 2007: 361). It focuses on choice and personal responsibility for balance in life. Most management and human resources research implicitly or explicitly focuses on this relatively privileged group. The latter discourse focuses primarily on employer need for flexibility and employer choice for balancing the workplace and employee needs. This study is on the employer decision to provide family support programs to employees as system level rewards. We treat organizations as social actors (Whetten, Felin and King, 2009) and employers as the implementers of organizations’ goals and actions.

For the theoretical foundation of this paper, we incorporate two complementary frameworks: the strategic choice model of organizations (Galbraith, 1977) and the industrial relations systems theory (Dunlop, 1958).

According to the strategic choice model (Galbraith, 1977), organizations function through a continuous choice process to achieve coherence among choices of strategy, organizing mode, and integration of individual and organizational interests. Individuals are attracted to the organization to execute the tasks expected of them. To achieve their goals, organizations design and administer a reward system. Rewards affect organizational performance through job satisfaction, motivation, turnover intention and productivity. System rewards are a form of extrinsic rewards that are provided to all employees based on their membership in the system. These rewards include fringe benefits such as family support programs.

Applying this framework to our study, we propose that employers offer family support programs as a system reward with the anticipation that such programs will attract high quality workers to their organization, decrease work-family conflict, increase job satisfaction and commitment, and decrease turnover and absenteeism. These, in turn, will contribute to organizational productivity, thus reducing costs associated with turnover and absenteeism. Such programs will also improve the organization’s public image as “caring employers” leading to increased share value (Kelly et al., 2008).

We also draw on the industrial relations systems theory (Dunlop, 1958), which focuses on employers as one of the three key actors of the system interacting with the other two actors (employees and their representatives, and governments) to develop collective and individual rules in the workplace. The actors interact and develop rules and procedures guided by their goals, values (ideology), power and history (Meltz, 1993). Employers’ strategic goals are significant in the types of policies and programs introduced in the workplace (Kochan, Katz and McKersie, 1986).

Since the industrial relations systems theory is a broad framework, we apply only relevant sections to our study. Employers are influenced by external environmental factors, including social values and norms in the society. Employers’ decisions are also influenced by their own goals, values and power, and the history of work relationships in the organization. In applying the theory to our study, we argue that employers make strategic decisions to offer family support programs with the anticipation that such programs will attract the best workers to their organization, decrease work-family conflict, and contribute to organizational productivity.

The empirical research on work-family interface is extensive in social science disciplines. The literature demonstrates that childcare and eldercare are disproportionately the responsibility of working women (Marshall, 2006; Pyper, 2006). Recent data from Statistics Canada shows that 73 percent of all women with children less than 16 years of age are employed (Almey, 2007), and although women are working more hours outside the home and men are spending more hours in domestic responsibilities, still, the gendered division of housework continues to exist in dual-earner households with women having more work at home than men (Marshall, 2006).

Women and men make individual and gender-specific accommodations for work-family balance. Research shows that work-family-life balance issues are invariably intertwined with discussions on gender and part-time work (Tijdens, 2002; Owens, 2006). Throughout the 1990s, almost one in five in Canada worked part-time; most part-timers were women, and 73 percent of part-timers were voluntary part-time, i.e. working part-time for personal choice or to accommodate personal circumstances (Marshall, 2001). Research confirms that benefits coverage is lower in part-time employment as compared to full-time employment (Zeytinoglu and Cooke, 2005), though the researchers did not examine voluntariness of part-time or employer offered family support programs in their study.

The research shows the need for family-friendly workplace practices such as flextime, childcare and eldercare support, to ease work-family tensions and enhance workplace performance (Baltes et al., 1999; Duxbury and Higgins, 2001). However, empirical research shows that the availability of employer support is limited and there seems to be a gender difference in the support provided, with men more likely to obtain support. In terms of flexible hours, research from Canada (Zeytinoglu, Cooke and Mann, 2009) and the US (Golden, 2005; McCrate, 2005) shows that, contrary to commonly held beliefs, women do not have more flexible hours than men.

Focusing on part-time and family-friendly practices in Canadian workplaces and using WES 1999 data, Comfort, Johnson and Wallace (2003) show that access to family-friendly work arrangements is extremely low, and access is unrelated to employees’ personal or family characteristics but linked to establishment characteristics of size and sector. In a more recent study, Ferrer and Gagné (2006), using WES 1999-2002 data, group childcare, eldercare and other types of family support services into a single category, and show that in Canadian workplaces, these benefits seemed to be biased towards workers who are in no dire need of them, and the benefits depend strongly on workplace size, industry and occupation factors. Using a separate data set, Higgins, Duxbury and Lyons (2008) find that unpaid leave of absence and psychological/health counselling are the two most common family friendly benefits provided to Canadian workers (84% and 83%, respectively). Among their respondents, only 8% have on-site day care benefits, 7% have childcare referral service, and 6% have eldercare benefits. In the US, a national sample of medium and large private-sector employers finds that on-site childcare provision is rare (9%), but 39% of these employers have information and referral services for eldercare (Galinsky et al., 2008).

In terms of eldercare, in Canada, informal care by family members far exceeds formal care (Chappell, 2009; Keefe, 2009), and the proportion of Canadians caring for elderly persons is increasing. In 2002, more than 1.7 million Canadians age 45 and older combined paid employment and care to older adults on a regular basis (Pyper, 2006), and most caregivers worked full-time (Keefe, 2009). Gender issues in eldercare are not frequently acknowledged at the policy and corporate levels. Yet, most caregivers are women (Daly, 2009; Chappell, 2009; Keefe, 2009; Zeytinoglu and Denton, 2009). Canadian experiences are similar to the US, Europe and other industrialized countries’ experiences (Lero and Lewis, 2008).

As the literature review shows, while rich in knowledge generated, research to date is limited. Some are descriptive analysis and researchers recommend multivariate analysis on the topic (Comfort, Johnson and Wallace, 2003). Others do not consider various dimensions of part-time work (Ferrer and Gagné, 2006; Lero and Lewis, 2008). This study addresses some of those limitations with its focus on both descriptive and multivariate analysis, and the voluntariness of part-time work as associated with employer offered family support programs while controlling for other factors.

Integrating the strategic choice model of organizations, the industrial relations systems framework, and the empirical research on the topic, we focus only on the employer offering family support programs. In other words, we are examining the determinants of offering such programs, not the individual or organizational outcomes such as job satisfaction or productivity. In the conceptual framework of this study, employer offered childcare and eldercare programs are dependent variables and gender and employment status are independent variables. We control for a number of individual, workplace and sector factors that, independently and collectively, can influence an employer’s choice to offer childcare and eldercare programs.

A high percentage of employers offering childcare and eldercare support programs will indicate the importance given to the topic by employers, and it will also show the support employers give to their workers in achieving work-life balance. We presume that in providing family support programs as a benefit to workers, employers will not differentiate among workers based on individual characteristics and thus, gender is not expected to be associated with employer offered family support programs. We also presume that it is the employer who decides whether to offer family friendly programs and the coverage according to employment status. We expect that full-time workers, as the core group, will more likely be offered family support programs, and compared to full-timers we expect voluntary and involuntary part-time workers will less likely be offered family support programmes.

Given the limits of theoretical approaches, the ambiguity in conceptualizing and understanding the work-life balance phenomenon (Fleetwood, 2007), an oversimplification of the phenomenon in the literature (Lewis, Gambles and Rapoport, 2007) and the implicit gendered (Phipps, Burton and Osberg, 2001) and classed nature of the work-life balance literature (Collins, 2007), our expectations should be taken as exploratory.

Methodology

Data

Statistics Canada’s 2003 Workplace and Employee Survey (WES) is used in the study. The advantage of the WES is that it links employer and employee responses, and data contains a large number of variables not usually available in other data sets. It provides accurate and unbiased estimates of the effect of particular workplace and worker characteristics, and gives an indication about how the attributes of the employees and workplace activities jointly affect worker outcomes. 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, i.e. federal and provincial government and municipalities. The broader public sectors of education and health care are included. The sample is drawn from the Business Register. The 2003 WES has data on 20,834 employees from 6,565 workplaces. Using weighted micro-data and the bootstrapping technique allows us to essentially replicate the entire Canadian labour market with only the exceptions listed above. (For more on sampling, sample design, and response rates, see Statistics Canada, 2006). Thus, our results are generalizable within Canada.

Variables

To examine the extent of employer offered family support programs, the following questions are used: “Does your employer offer personal support or family services such as childcare, employee assistance, eldercare, fitness, and recreation services or other types of services?” Those who respond yes to this question are then asked: “Does your employer offer help for childcare either through an on site centre or assistance with external suppliers or informal arrangements?” For eldercare the question is: “Does your employer offer help with eldercare services?” The dependent variable in the multivariate analysis is the childcare question transformed into a dichotomous dependent variable of “offered childcare support” (1 = yes, 0 = no). A similar variable was created for eldercare but, as we explain below, was not used. These questions illustrate the advantages and disadvantages of the WES. The advantage is that it contains many variables and constructs not included in other datasets. However, the large breadth of the dataset is at the expense of depth of some individual variables such as childcare or eldercare, where the extent of “help” is not quantified.

The independent variables are gender and employment status. Gender is coded as female = 1 and male = 0. Employment status is from the employee survey and is categorized into full-time, voluntary part-time and involuntary part-time work.[1] Full-time refers to those working at least 30 hours per week. The voluntary part-time variable is derived from the questions pertaining to hours of work and preference to work the same number of hours for the same pay. Those working less than 30 hours per week and who prefer to work the same number of hours for the same pay are coded as voluntary part-time. Those working less than 30 hours per week and who do not prefer to work the same number of hours for the same pay are coded as involuntary part-time. In the multivariate analysis, full-time work is included as the reference variable. We also created an interaction variable of gender and employment status, and used this in the multivariate analysis, but it was not significant and therefore not discussed.

Many other individual, workplace and sector characteristics factors are included as control variables in the multivariate analysis. Individual characteristics include marital status, dependent children, occupation, age, wage, education, and being covered by a collective agreement, and they are from the employee survey. Marital status is categorized according to whether they are married or in a common-law relationship (= 1) or other marital status, i.e. never married, separated, divorced, or widowed (= 0). Dependent children is coded as a dummy variable (1 = yes, 0 = no). Occupation is listed in four groups as lower white collar, manager, professional or blue collar (dummy variables with the last as the reference group). Age is separated into groups representing younger, middle, and older workers (with middle age as the reference). There are four dummy variables for wage: very low (< $10/hour (the reference group)), low ($10 to < $20/hour), medium ($20 to < $40/hour), or high (≥ $40/hour). Education is coded into three categories as high school graduate or lower (the reference group), some post-secondary education, or university degree. Whether workers are covered by a collective agreement is a dummy variable (1 = yes, 0 = no). Size, coded in the number of employees, and operationalized in the multivariate analysis as the log of size, is the workplace characteristic, and the question is from the workplace survey. Sector variables include four dummy variables of manufacturing (the reference group), service, finance and insurance, and education and health care.

Analysis

Descriptive statistics are calculated and logit and logistic regression analysis are conducted. We present odds ratios, regression coefficients, bootstrapped standard errors, and significance levels of each variable. The analyses are generated using weighted micro data accessed at Statistics Canada’s McMaster University Research Data Centre (RDC). Bootstrapping is used in the multivariate analysis due to the complex survey design of the WES. We used Statistics Canada’s mean bootstrap weights using the Stata file developed by Chowhan and Buckley (2005). Their file calculates variance estimates using bootstrap weights allowing researchers to calculate reliable variance.

Limitations of the Study

There are a few limitations of the study worth noting. First, ideally employer responses should be used for examining employer offered family support programs but these questions are not included in the employer survey, thus we use the questions from the employee survey. It is possible that employees might not have good knowledge of what is offered and the responses might not be accurate. Second, the question on eldercare support is particularly vague since the employer offering help with eldercare is not defined in the survey. However, this question follows the childcare question and we assume that recalling the items listed in the childcare question, respondents might have referred to them in their responses. Also, as mentioned earlier, the extent of “help” provided by employers is not available for either variable. Third, there were no specific questions on whether workers wanted to work part-time voluntarily and therefore they were derived from two questions. This, however, creates some potential ambiguity in the definition of involuntary part-time work. Lastly, data does not include federal and provincial civil servants and municipal workers who might have higher coverage of employer offered family support programs. Despite these limitations, the data provides a rare opportunity to investigate Canadian workers’ experiences on this topic.

Sample Characteristics

Descriptive statistics of all variables are shown in Table 1. Among all workers (as shown in Table 1, Column 1), the majority of the sample is working full-time (85 percent) and the rest are in part-time work. Voluntary part-time workers are about 10 percent of the sample and involuntary part-time workers are about 6 percent of the sample. Slightly more than half of all workers are female, more than two-thirds are married or in a common-law relationship, and almost half have dependent children. In terms of occupation, almost half are in blue collar jobs, followed by those in lower white collar jobs, and professional and managerial occupational categories. Turning to worker age, more than one-fifth are younger, another one-fifth are older, and the remainder are in the middle age category. One in seven workers has very low wages, almost half are low-waged, one-third of workers are medium-waged, while the remaining 6 percent have high wages. Slightly less than one-third of all workers have high school education or lower, about one-fifth have university degrees, while the remaining half of workers have some post-secondary education. The mean level of full-time work experience is 17 years. One quarter of workers are covered by a collective agreement. The mean organization size is 414 workers, and in terms of sector, one-third are in manufacturing, 7 percent are in finance and insurance, 21 percent are in the education or health care sectors, and the remaining 39 percent are in other service industries.

Table 1

Characteristics of All Workers and Those with Employer Offered Family Support Programs (descriptive statistics (i.e. mean values) of variables among five worker sub-groups)

Characteristics of All Workers and Those with Employer Offered Family Support Programs (descriptive statistics (i.e. mean values) of variables among five worker sub-groups)

-> Voir la liste des tableaux

Results

The Extent of Employer Offered Family Support Programs

As shown in Figure 1, regarding whether the employer offers personal support or family services, about a third report receiving some form of family support. Subsequently, when these individuals are asked whether their employer offers childcare support, one in five respond positively and among all respondents, these workers represent only 7 percent. For eldercare support, of those offered some type of personal support or family services, 12 percent are offered eldercare support, and among all respondents this equates to 4 percent. These small percentages of employers offering these programs are an indication of the low priority employers give to the work-life balance concerns of their employees, at least with respect to these specific forms of support.

Figure 1

Availability of Employer Offered Family Support Programs (Excat numbers and percentage for the weight sample)

Availability of Employer Offered Family Support Programs (Excat numbers and percentage for the weight sample)

-> Voir la liste des figures

Descriptive Statistics of Workers Offered Family Support Programs

Characteristics of those who are offered childcare support and those who are offered eldercare support are shown in Table 1. Column 2 presents the descriptive statistics among those who are offered childcare support. Relative to all workers, those who are offered childcare support are more likely to be male and full-time, and less likely to be voluntary or involuntary part-time. For those who are offered eldercare support, the relevant statistics are presented in Table 1, Column 3. The proportion offered eldercare is similar to the proportion among all workers when considering gender, or full-time and voluntary part-time employment status but relative to all workers, those to whom eldercare support is offered are less likely to be involuntary part-time workers. Due to space limitations, further characteristics of those offered childcare or eldercare are not discussed here but are presented in Table 1.

Bivariate and Multivariate Regression Results

To examine whether employer offered family support programs are associated with gender and full-time, voluntary part-time and involuntary part-time work, we conduct bivariate and multivariate logit and logistic regressions. In the bivariate analyses, we provide results for each variable regressed onto being offered childcare support (Table 2). In the multivariate analysis, the model includes all of the employment status variables and control variables simultaneously. Since our interest is on substantive, not statistical significance, we focus on odds ratios that differ tangibly from 1.0 in the following discussion.

Table 2

Logistic Regression Analysis: The Probability of Being Offered Childcare Support

Logistic Regression Analysis: The Probability of Being Offered Childcare Support

*** p < .01, ** p < .05, * p < .1

-> Voir la liste des tableaux

For the employer offered childcare support, we first present bivariate logits. As shown in Table 2, females are more than one third less likely than males to be offered childcare support. Those in full-time employment are 80 percent more likely than part-time workers to be offered childcare support. While involuntary part-time workers are less likely to be offered childcare support compared to all others, the gap is not large relative to voluntary part-time workers.

In the multivariate model, we presumed that in providing family support programs as a benefit to workers, employers will not differentiate among workers based on individual characteristics. Contrary to our expectations, gender is significantly associated with employer offered childcare support, with females one third less likely to be offered childcare support compared to males, controlling for other factors. The results show that, as expected, workers in both voluntary and involuntary part-time employment are less likely to be offered childcare support than are full-time workers, but the gap is statistically significant only for those in involuntary part-time work. High wages, large organization size and employment in finance and insurance or education and health care sectors are significant factors associated with the probability of employer offered childcare support.

We do not provide an analysis for eldercare support because the proportion accessing employer offered eldercare support is too small and in some cases, negligible. In addition, small cell counts make the analysis unreliable. Nonetheless, the proportions provided in Figure 1 and Table 1 clearly show the very limited extent of employer offered eldercare support programs in Canadian workplaces.

Conclusions and Discussion

Results show that for most Canadian workers, there is a low level of employer support for childcare and eldercare. This low level of support indicates the lack of importance employers attach to this work-life balance issue. We base our conclusions on the low likelihood of Canadian employees being offered childcare and/or eldercare support by their employer. In addition to this low overall prevalence, we found that females are significantly less likely to be offered childcare support when controlling for other factors. The same pattern appears to exist regarding eldercare support as well, based on the extent of analysis that we were able to undertake given small sample sizes. In terms of employment status, full-time workers are more likely to be offered childcare or eldercare support. Those voluntary and involuntary part-time workers are less likely to be offered childcare or eldercare support, although this finding is statistically significant only for involuntary part-time workers.

Of course, being offered child-care and eldercare support is not the same as needing them. That is, it is reasonable to assume that some workers offered these support programs will not have a need for them. Conversely, others needing one or both supports will not have access to them from their employer. Certainly, the second situation is a far more serious mismatch than the first.

We expect that due to their extra responsibilities at home, on average, female workers would be more likely to need child- and/or eldercare support programs. Although not shown, preliminary analysis for this paper supported our presumption. The need for these support programs among part-time versus full-time workers is more ambiguous. It could be argued that full-time workers need support programs more because their longer workweek means that they have less time and energy to devote to their responsibilities outside of work. On the other hand, it is plausible that part-time workers, whether voluntary or involuntary, are as likely to need these family support programs for work-life balance since they might have this work schedule due to the extent of their other life and family commitments. In any case, we argue that the finding that females have lower access to these family support programs is particularly troublesome.[2]

It seems that the experiences of the last four decades of middle-class women’s entry into the paid labour market and converging gender roles in families did not trigger comparable changes in the workplace. Family structures have changed; dual-earner couples have increased; women’s earnings have become a significant part of the family income often raising the family to the middle-class status; and males have taken more household, childrearing and eldercare responsibilities. All workers need some type of support with childcare or eldercare at some period in their lives, and that support is crucial if the prime-age working population is expected to fully engage in paid work. Policies that facilitate work-life balance are needed and such policies will be critical in the coming years when, in addition to childcare, the population needing care grows relative to the population of worker/caregivers.

Although the characteristics of workers who need childcare versus eldercare support programs could be different, we have chosen to analyze both in this paper since there are obvious similarities. Very few Canadian workers are offered either of these types of support programs, and for childcare support in particular, having access to this support from one’s employer is apparently affected by job, industry, and workplace factors as opposed to workers’ needs or preferences. We presume that similar access issues exist for eldercare supports as well, although small sample sizes prevented us from conducting multivariate analyses. It can also be argued that, in today’s global business environment, Canadian employers cannot be expected to provide benefits such as family support programs, even if desired by their employees.

In our view, it is time that we consider childcare and eldercare as societal responsibilities rather than individual, or even corporate, responsibilities. We argue that if individuals are going to receive assistance for childcare and eldercare, that assistance is more likely going to come from the government as opposed to employers, and we recommend publicly funded universal family support programs. As Doherty (2007) says, this does not mean that there has to be a single program in which all children and adults needing care must participate. It means there should be services available to all Canadians, should they need them and want to use them, in a way similar to the universal health care coverage in Canada.

Children are the future of society and the elderly have contributed to building our society and thus both groups deserve support from the system if, and when, they need it. There is some support for eldercare in Canada, though it is not fulfilling the demand nor does childcare support in the rest of Canada come close to universal childcare support in the province of Quebec. There is decades old discussion of the need for universal childcare (Rebick, 2005) and extensive literature showing the benefits of universal government subsidized childcare for children’s developmental outcomes and work-life balance for parents (Doherty, 2007). Unfortunately, the political will to implement public policy recommendations is still lacking in Canada. For the universal government subsidized childcare and eldercare support policies to be implemented, workers themselves have to see that the issue is not an individual work-life balance problem but a societal concern that needs to be addressed by the federal and provincial governments in Canada. Given the industrialized countries’ and intergovernmental organizations’ interest on the part-time work and work-life balance topic, the results presented here can provide guidance to other countries contemplating solutions to work-life balance concerns.