Work-related travel and job autonomy
Department of Sociology
405 30 Göteborg
Telephone: 46 31 786 47 70
Preliminary draft. Please do not quote.
Paper to be presented at the WES Conference, Aberdeen, Scotland, 12-14 September, 2007
‖It is difficult for those who manage workers
on the move to keep track where they are
and observe what they are doing.‖
(Felstead, Jewson and Walters 2005: 166)
Within contemporary research on work and the labour market, there is a common claim that flexibility is one of the most significant characteristics of working life today; even if there is no general agreement on how to define the concept, or the consequences for workers and employment relations (Furåker, Håkansson and Carlsson, 2007).
This report focuses on a special aspect of flexibility in working life, namely flexible work situations. Flexible work situations can mean a lot of things. In the literature, there is often a connection to different forms of employment conditions or employment contracts, such as appointment on a temporary basis, or flexitime systems. In this study the focus is instead on 1spatial flexibility as a part of the work situation. More exactly, it deals with people who
travel regularly in their jobs; workers in the transport sector, as well as in business and sales.
In advanced capitalist economies like Sweden, there are many factors which promote work-related travel. In production and business there is an increase in cooperation, communication and networking which often requires personal meetings. The transport sector is on the rise. Technical aids like mobile phones and portable computers make it possible to perform job tasks in hotels and aeroplanes. Some people perform their work at various locations and many categories of employees attend courses and conferences as part of their jobs (Gustafson, 2006).
This means a growing number of work tasks are not place-bound, standardized and timetabled (Hislop and Axtell, 2007). Employees carrying out these work tasks are – to a greater or
lesser extent – mobile in their work and this means, among other things, freedom from direct supervision. In factories and offices management regulate labour through control over time and space. Fixed workstations and regular working hours makes it possible to govern employees‘ activities. The working day is scheduled and many of the work tasks are standardized and governed by routine. There is limited discretion over work performance. This is still the dominating feature of working life. However, this ‗fordist model‘ is not as
dominant today as it was some decades ago, partly due to greater mobility in working life.
Mobile work arrangement can be seen as a resource, which expands autonomy, independence and freedom of action for the worker (Kurland and Baily, 1999; Sirianni, 1991). This poses problems for managers, who strive to control the mobile workers through different types of supervision and surveillance. Employees can be ordered to regularly report their activities or the outcomes of the job, managers can use daily phone calls or different forms of electronic supervision and registration of work performance. (Bennett, 2005; Felstead, Jewson and Walters, 2005: 165-70). Furthermore, in the mobile work context managerial control mechanism can emphasize collaboration with supervisors and different types of group pressure (Adami, 1999; Kinsman, 1987).
1 Spatial flexibility can be seen as one aspect of the broader concepts ‖Flexible work practices‖ or ―Flexible
work arrangements‖. See Powell and Snellman, 2004; Smith, 1997; Kelly and Kavel, 2006.
Many of the mobile workers use advanced technology in their jobs. In a recent study of trends in working conditions within several nations (Green, 2006: 94-110) it was found that discretion and autonomy in work in many countries are decreasing and that this trend can be 2connected to the ever-increasing use of advanced technology.
Thus, mobility in work does not necessary mean autonomy in work. The main question for this paper is therefore how work-related travel affects job conditions and job experiences in terms of autonomy and influence over work conditions. Are mobile employees enjoying greater job autonomy compared to their non-travelling colleagues?
The study is grounded on survey data from Sweden. Because quantitative studies of the subject have been rare, another purpose of the paper is to give a general picture of the extent of work-related travel in working life in contemporary Sweden.
Research on job autonomy and mobility in work
Influence and autonomy at work have appeared to be one of the most significant features of the employment contract (Baldry et al., 2007: 5-11). During the 1980‘s and the early 1990‘s
there was a great deal of optimism about increasing autonomy for employees—at least in
wealthy countries—as a consequence of new technological advances, more qualified work tasks, and a budding interest among managers in high-performance work practices and employee involvement (Kern and Schumann, 1992; Walton, 1985). However, the empirical and theoretical basis for this positive view has been questioned (Appelabaum et al., 2000; Head, 2003; Vallas and Beck, 1996). Thus, it is difficult to reach a definite conclusion on earlier and recent trends for autonomy and discretion in working life, even if there seems to be little empirical support for the more optimistic scenarios (Gallie et al., 2004; Green, 2006: 94-110).
This also holds for trends in mobile work. The hypothesis about increased mobility in work is intuitively reasonable and easy to gain support for in the literature. However, it is difficult to find representative survey studies which puts the hypothesis to a test. Research consists mainly of case studies of workplaces, branches or occupations and there is no agreement about the operational definition of mobile work or if the concept should be used at all.
Thus, researchers trying to estimate the extent and changes of mobility on work often reach different results (Hislop and Axtell, 2007). Nevertheless, there seems to be an agreement that a substantial number of workers are mobile in their jobs and that mobile workers‘ share of the
workforce is growing. At the same time researchers point out that this share is still very small (Dimitrova, 2003; Vilhelmson and Thulin, 2001).
Most empirical research on connections between mobile work arrangements and autonomy has focused on ‖teleworking‖; work performance with the help of laptops, mobile phones and other types of communication technology. Since a great deal of this sort of work is performed outside the traditional workplace and is characterized by spatial flexibility, research on
2 Sweden and Holland deviate to a certain extent from this pattern and in Germany there is a clear trend of increase in autonomy and discretion for employees (Data from second and third European Surveys on Working Conditions).
teleworking has relevance for this study as well; in particular investigations of connections 3between mobile work, supervision and task discretion. The research points towards
contradictory results. There are examples of studies where mobility has hardly any effect on the degree of supervision and control, as well as studies showing a higher degree of autonomy among mobile workers. Others have found that mobile work increased supervision. Finally there are examples of polarisation between less control of high-ranked mobile employees and more supervision of mobile workers in lower positions (Dimitrova, 2003).
The last findings underlines the importance of concepts such as class, gender and age when studying correlations between mobility and autonomy at work Different ways of using new technology, managerial and organizational commitment to employee involvement and especially skill level are often pointed out as important mediating factors for autonomy in mobile contexts.
Social patterns of work-related travel and autonomy
Studies about work-related travel often focus on travelling among well-educated employees with high incomes (Haynes, Vecci and Wickham, 2005; Urry, 2003). These groups travel relatively much in work, especially when it comes to long distant travelling (Frändberg and Vilhelmson, 2003; Presser and Hermsen, 1996). However, it is important to remember that many people with flexible work arrangements do not belong to the upper middle class, for example, truck drivers and some categories within the health occupations. Work-related travel is not an exclusive hallmark for privileged groups in the labour market, it is a common feature of many different occupations.
Conditions and consequences of work-related travel should differ between occupations, when it comes to purposes and contents of job journeys and to what degree workers will be supervised and controlled by management. We know that there is a huge gap in job autonomy related to occupations and authority at the workplace. Consequently, the degree of autonomy among travellers may vary due to their class position, as well as gender and age (Adler, 1993).
Researchers have pointed to the fact that managers‘ control of higher civil servants and salaried employees above all consists of positive incentives, while blue-collar workers often are controlled by intensive supervision and threats (Marshall, 1997: 2; Wright, 1979: 84-86). In that case one can expect that managers choose to compensate for autonomy among blue-collar travellers through more supervision and monitoring, while privileged travellers are rather awarded more freedom and rewards. However, this is not an obvious conclusion because managers are often eager to control the work of professional employees. It also seems to be the most privileged employees who are mobile in their works, at least among higher civil servants. In that case mobility can be an effect variable as well as a cause variable (Clear and Dickson, 2005; Hodson, 2001: 140-170).
The point of departure for this study is that autonomy at work can be defined as the control workers have over their work situation. Since the main interest is in the impact of mobile
3 Earlier research on telework focused above all on home-located work. This has somewhat changed in recent years and more studies have been conducted on mobile work in general (Felstead et al. 2005, Hislop and Axtell 2007).
work on job autonomy, the focus lies on individual autonomy in the work situation—the
worker‘s capacity to decide how and when to perform his or her work tasks – ―task authority‖ 4or ‖task discretion‖. In this case the dependent variable has been operationalised as an index (Cronbach‘s alpha = 0.71), on the basis of survey questions regarding respondents‘ influence
over working speed and how to perform work tasks.
There are six independent variables in the analysis: travel frequency, class, gender, ethnicity, age and sector. Class has been transmitted to socio-economic position in line with categorization by Statistics Sweden, there educational level and belonging to different unions 5are the most important variables for distinguishing between class positions. The variable on
ethnicity leaves much to be desired, as it does not distinguish between different countries of 6origin, just between Sweden and countries of origin other than Sweden. Age is a substitute
for the more desirable variable work experience. I also choose to control for sector, as this variable has turned out to exert influence on workers‘ autonomy (Gallie et al., 2004).
The survey data were gathered during 2005, through a mail questionnaire carried out by Statistics Sweden. In order to achieve an appropriate selection of respondents due to the research questions, the sample was stratified. Respondents who participated in the labour force survey in March 2005 were asked if they travelled 100 km or more in order to perform their work tasks, during a normal working month. Then questionnaires were sent to about 2000 individuals who usually travel at least 100 km during an average working month and to
as many individuals who travelled less or not at all. The total sample consisted of 4041 individuals with a response rate of 70 percent, or 2804 respondents. This report focuses on employed individuals, which means that the sample contains 2587 respondents. All analyses use a weight variable in order to compensate for the stratified sampling design.
The extent and purposes of work-related ravel
4 This means that aspects of autonomy without clear connections to work-related travel will be excluded from the analyses; for example collective autonomy and employees opportunities for participating in organization-wide decisions (Guzzo and Dickson, 1996; Langfred, 2000). 5 Se appendix for definitions of class categories. 6 Of course there are huge differences in discriminating mechanisms in working life who affect workers from Norwegian or other Nordic countries, compared to workers from, for example, Latin America or Africa.
As mentioned earlier, a great deal of research about work-related travel concentrate on elite groups who travel over long distances. However, this is hardly representative for most work-related travellers (van der Land, 1998). Among the respondents in this study, a relatively high proportion of employees – about 40 percent – travel regularly in work (at least a few times a
year). Almost a sixth, 16 percent, is highly mobile which means that they travel several times a week. The great majority of journeys are for short distances. Among those who are
moderately or low mobile, there are just over 40 percent reporting regular travel outside the county and about 10 percent who report foreign travel. Corresponding numbers for highly mobile workers are 37 and 6 percent respectively.
< Table 1 >
Why do people travel in work? The main purposes are meetings, conferences and courses. However, this is not true for daily travellers, where ―manual labour‖ is the most common response. Transport of persons and transport of goods are also frequent answers within this group.
As expected, the purposes of work-related travel vary with respect to class. Among blue-collar workers who often travel in work, many respondents report manual labour (repairs, service work), transports and home-help service, but nearly one fifth also report courses and conferences among the purposes of job journeys. Among higher civil servants and senior salaried employees, work-related travel above all is refers to participation in meetings and conferences, as well as administrative work such as consultants and auditing. These patterns
are similar for both men and women.
Work-related travel within different groups
7Table 2 shows travellers‘ share within different categories of employees. In a sense the
results are not surprising, since privileged groups of employees stand out as the most mobile; higher civil servants travel more often than unskilled workers, men travel more often than women etc. On the other hand, blue-collar workers are relatively frequently among the most mobile. This is partly a consequence of the high frequency of travel among workers within the
transport sector, but also within health service and medical care there are a great share of employees who often travel in their jobs.
< Table 2 >
The most mobile class category is senior salaried employees, and within this group there are above all officials, politicians, managers and heads of enterprises in both the private and public sectors, which are mobile in their work. Among higher civil servants there is relatively high mobility among auditors and public sector employees such as social welfare secretaries and welfare officers. Perhaps surprisingly, there is not so much work-related travel among academic occupations such as university teachers and researchers within the humanities and
7 Se appendix for definitions of class categories.
social sciences. And there are not so many travellers within the IT- and data sector either. These groups‘ share of immobile persons is clearly above the mean share for the population.
Within every class group, men are more mobile than women. However, there are significant differences between the class categories in this respect. Among blue-collar workers, 10 percent of women are highly mobile, compared to 30 percent of men. Corresponding figures for higher civil servants and senior salaried employees are 13 and 20 percent respectively. Among lower level non-manual employees men‘s share of highly mobile respondents is
nearly ten times larger than women‘s; 32,4 percent compared to 3,8 percent. The main
differences between men and women can be found at the lowest and the highest points on the travel scale. 68 percent of women are immobile, compared to 46 percent of men; and 9 percent of women are classified as highly mobile, compared to 24 percent of men.
Among age groups there are little differences in travel frequencies, with the exception of the youngest group with a relatively great share of immobile respondents. Within the other control variables – ethnicity and sector – there are relatively small differences in mobility
We will now turn to the analysis and first examine the affect from mobility on autonomy with control for class, gender, age, ethnicity and sector.
Mobility and autonomy in work
Table 3 shows results from the regression analysis. There are different columns with coefficient values after control for class (2), class and gender (3) and for class, gender and age (4). The table shall be read in the following manner: The figures below the ―Ref‖ sign (reference group = Not mobile) indicate the change of the indexes for one step change in the independent variable mobility. For example, when mobility change from immobile to low mobility, the index for autonomy increases by 0,5 points (left column).
From the results of table 3, it may be concluded that there is a significant effect of mobility on autonomy in work. This effect goes in the expected direction; the more work-related travel, the more autonomy in work. A smaller part of the effect disappears when controlling for socio-economic position—over one fourth of the difference in autonomy between immobile
and low/moderately mobile and 6-7 percent of the difference between immobile and highly mobile. Furthermore, one can notice that the correlation between mobility and autonomy becomes more obvious after controlling for class, that is, the direction of the correlation is the same and the distance between the mobile categories becomes greater. Thus, there should be significant differences in autonomy between mobility categories with respect to class.
< Table 3 >
The independent variable gender reduces the effect even more, with 20-25 percent. Above all, the independent variables class and gender affect the correlation between autonomy and mobility. The other variables—age, ethnicity and sector—have no significant effect. Yet it
can be noticed that there are significant differences between man and women as well as between classes and various age groups regarding correlations between mobility and
autonomy, independent of the effects of the other control variables in the analysis. Therefore, I am showing results only for the analysis with these three variables. The next step in the analysis will be to estimate the effect of class, gender and age separately.
< Table 4 >
The class effect is above all dependent upon the absence of significant correlations between mobility and autonomy within two class groups: unskilled or semi-skilled workers and higher 8civil servants. If we assume that the correlation between autonomy and mobility is based on fewer opportunities for management in controlling mobile workers, the result should mean that control or supervision of mobile workers within these groups is just as great (or small) as is the supervision of their ‖stationary‖ colleagues. In fact, unskilled workers who are highly
mobile have a considerably lower average value on the autonomy index compared to immobile unskilled workers.
Because the variable autonomy is operationalised as task discretion, it is possible that the result partly depends on the fact that the work tasks for mobile unskilled workers are perfunctory and standardized at least to the same degree as they are for their non-travelling colleagues. But this is hardly the whole explanation. Furthermore, the line of argument should not apply at all to higher civil servants.
< Table 5 >
On closer examination the analysis shows that the correlation between autonomy and mobility among unskilled or semi-skilled workers turns out to be negative for occupations such as postal employees and drivers of vehicles, while there is a positive correlation among machine operators. Among higher civil servants, the correlations between mobility and autonomy are negative for many occupations within the educational system, while it is positive for, among others, estate agents and heads of small enterprises. In other words, one explanation for the non-significant or negative correlations among unskilled/semiskilled workers and higher civil servants may be that there is a very large span between different occupations with respect to correlations between autonomy and mobility. However, the number of respondents within each individual occupation is too small to allow more detailed conclusions.
< Table 6 >
As can be seen in table 3, control for gender leads to a substantial decrease of the differences in autonomy between different mobile categories. This effect depends on the fact that the correlation between autonomy and mobility looks different for men and women. Men receive more autonomy in work, the more mobile they are. For women, the correlation is less linear
8 The correlation between work-related travel and autonomy among senior salaried employees is rather weak, which above all depends on the small number of respondents within this group. However, the beta coefficients indicate a linear correlation with clear distinctions between different mobile categories.
and for them, the important distinction is rather that between immobile and mobile in general when it comes to different degrees of job autonomy (table 5).
The results in table 6 show the same pattern among employees in the younger middle ages, 35 to 45 years old, while the correlation among employees who are 45-54 years old is more linear. In addition, it can be noticed that the relationship between autonomy and work-related travel with respect to age is more curvilinear—the correlation is non-existent among the
youngest respondents and becomes stronger among the older groups, to again become weaker within the oldest groups of respondents, employees 55 to 64 years old.
The impact from mobility on differences in autonomy
In the analysis above, autonomy has been treated only in relation to mobility. Therefore, it should be emphasized that findings from other studies on differences in job autonomy—that
higher white collar workers have considerably more job autonomy compared to blue collar workers, that men have more autonomy than women, that older employees have more autonomy than their younger workmates—also holds for this survey data. One can therefore
ask to what extent mobility equalized these sorts of differences.
< Table 7>
As can be seen in table 7, there is a clear reduction of the difference in autonomy between unskilled or semi-skilled workers and other classes when controlling for mobility; on an average the reduction is about 30 percent. Significant differences in autonomy only remain between higher civil servants and senior salaried employees on the one hand and unskilled or semi-skilled workers on the other. Besides, it should be noticed that, even before controlling for mobility, the important difference in job autonomy seems to be that between higher white collar workers and all other groups.
Differences in autonomy between women and men are obvious, before as well as after controlling for mobility this difference is as large as the same difference between unskilled workers and higher civil servants. Furthermore, compared to class, gender ―explains‖ more of the variation in autonomy, both before and after controlling for mobility.
There are also differences in job autonomy between the age groups. The main dividing line seems to go between the youngest respondents (16 to 24 years old)—with considerably lower
value on the autonomy index compared to other groups—and the oldest (55 to 64 years old),
who seem to have clear advantages over the other age groups. Similar with the results of the other analysis, differences in job autonomy are substantially reduced after controlling for mobility. Despite of that, significant differences remain between the youngest respondents and most of the other groups.
< Table 8 >
As expected, respondents with both parents foreign born have a significantly lower degree of autonomy, compared to respondents with Swedish-born parents. However, after controlling for mobility the significance disappears. But there are rather small differences between the estimates; about a 13 percent reduction in the beta coefficient.
Finally, of the control variables which have been used in this study, sector turns out to have the smallest impact on autonomy. However, it can be noticed that, in line with other findings (Gallie et al., 2004), public sector employees seem to have slightly more autonomy at work in general compared to their colleagues in the private sector, even if the differences are not significant.
The point of departure for this paper was the hypothesis—or assumption—that mobile work
arrangements would lead to autonomy in work, in the sense of task discretion. It seems as if this hypothesis has gained at least partial support from the analysis. There is a clear and
positive correlation between work-related travel and autonomy, which turns out to be significant also after controlling for, among other factors, class, gender and age. Therefore, results in this study are supportive of previous research which have found that mobile work arrangements strengthen job autonomy.
However, it must be emphasized that mobility only explains a smaller part of the variation in autonomy and that there are differences in both direction and strength for the correlation with respect to class, gender and age. For example, class differences are substantial, but perhaps not exactly in a way one would expect. The socio-economic groups who, in terms of autonomy, seem to take most advantage of mobility in work are skilled workers, intermediate-level non-manual employees and senior salaried employees. For other groups—lower level
non-manual employees, unskilled and semi-skilled workers and higher civil servants—there
are small effects from mobility on autonomy. For the two latter groups, the effects are negligible.
As mentioned earlier, within these groups there is a vast span between different occupations with respect to correlations between work-related travel and autonomy. Transport workers (unskilled or semi-skilled workers) and compulsory school teachers (higher civil servants) are examples of occupations where mobile employees seem to experience a lower degree of autonomy, compared to their ‖stationary‖ colleagues.
For both transport workers and school teachers there may be a connection between the distance of the journeys and degree of autonomy. The effect from mobility on autonomy seems to be stronger among long distance travellers—such as foreign lorry-drivers—
compared to employees who most often travel over short distances. Among higher civil servants this effect seems to be weaker, but also within this group there is a tendency that long distance travellers have somewhat more job autonomy than other mobile workers and immobile employees. Unfortunately, the data does not allow a more detailed analysis of these correlations.
Women who are mobile have more job autonomy than women who do not travel in work. At the same time, differences between immobile and mobile female employees becomes smaller with increasing mobility. Among men, the correlation is more linear with the strongest