Transportation Research Part A 91 (2016) 251259

Contents lists available at ScienceDirect

Transportation Research Part A

journal homepage: www.elsevier.com/locate/tra

What drives the drivers? Predicting turnover intentions in the

Belgian bus and coach industry
Steven Lannoo a,, Elsy Verhofstadt b
a
Institute for Coach and Bus (ICB), Metrologielaan 8, 1130 Brussels, Belgium
b
Ghent University, Faculty of Economics and Business Administration and Sherppa, Sint-Pietersplein 6, 9000 Ghent, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: The bus industry is characterized by demanding jobs and high turnover rates. In this study
Received 29 September 2015 we gather essential insights that can help companies and industry-level policy makers
Received in revised form 14 April 2016 increase the attractiveness of the profession and design effective retention policies. We
Accepted 29 June 2016
compare the factors that induce Belgian drivers to leave their current organization with
Available online 16 July 2016
those inducing them to leave the industry. Key factors increasing the likelihood to consider
quitting the company are a negative work-life balance, a lack of social support and a tem-
Keywords:
porary contract. Dominant factors to consider quitting the bus driver profession are a lack
Turnover intentions
Bus drivers
of fulfillment, a demanding job environment and a negative work-life balance.
Bus and coach industry 2016 Elsevier Ltd. All rights reserved.
Job search
Separations

1. Introduction

For economists, the phenomenon of voluntary job turnover is important for a variety of reasons. From a macro perspec-
tive, a lot of interest has been expressed in the role of voluntary separations for the efficient allocation of workers and jobs,
triggering research into the relationship between the business cycle and job turnover among other things (Lazear and
Spletzer, 2012). From a managerial perspective, employee replacements involve high costs for the organization. These entail
not only expenses for recruitment and selection, bus also training and (temporary) reduction in productivity. For instance,
cost of replacing drivers for American truckload carriers are estimated to be between $2200 and $21000 pro capita (Suzuki
et al., 2009). Finally, from a micro-economic point of view, voluntary job turnover implies that workers believe the current
balance between job-related rewards and job-related efforts to be suboptimal. After all, the utility maximization theory of
quitting behavior predicts that workers will only leave their job when the balance of expected efforts and rewards in their
current job is less appealing than the one in the alternative job, taking the costs of changing into account (Lvy-Garboua
et al., 2007).
Hence, voluntary job turnover is a subject that has received massive research attention. In the next section we link our
data to the existing, mainly empirical, research on the topic. Attempts have been made to integrate the existing knowledge
into more unified theoretical frameworks (for instance Maertz and Campion, 2004 and Holtom and Interrieden, 2006). These
efforts are of great importance and indeed indispensable when scientific progress has to be made. However, the dynamics of
the job change decision process are enormously multifaceted and depend on the specific characteristics of the sector,

Corresponding author.
E-mail addresses: Steven.Lannoo@icb-institute.be (S. Lannoo), Elsy.Verhofstadt@UGent.be (E. Verhofstadt).

http://dx.doi.org/10.1016/j.tra.2016.06.024
0965-8564/ 2016 Elsevier Ltd. All rights reserved.
252 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259

organization and the job. In order to help (HR-) practitioners develop successful retention policies, these general models
need to be supplemented by industry specific and even job-specific insights (Holtom et al., 2008).
Bus driving is internationally known to be a very demanding profession (Jones et al., 2013) and labor turnover in the
industry is reported to be rather high (Hlotova et al., 2014). In Belgium, the sector has been confronted with a growing scar-
city of drivers caused by expanding activities and high turnover rates. Because of the high average age of the population of
bus drivers, it is expected that the need to replace retired personnel will further intensify that scarcity (Dewit, 2013, VDAB,
2015). As a consequence, increasing the attractiveness of the profession and developing retention policies can be considered
a necessity for the industry. The purpose of this study is to provide the necessary insights for designing such policies. More
specifically, we investigate which variables researched in the literature are of greater and lesser importance for the turnover
intentions of Belgian bus and coach drivers. We distinguish between variables inducing drivers to leave their current orga-
nization in order to be a professional bus driver in another organization and variables inducing them to leave the industry
altogether. Most of the literature on voluntary job turnover does not distinguish between the intention to leave an employer
and the intention to change professions. However, as Delfgaauw (2007) argues, the motivations behind these two decisions
can be rather different. Employees who are dissatisfied with job domains that are rather organization specific are more
inclined to change employers. On the other hand, workers dissatisfied with job domains that are function-specific will be
more inclined to leave the industry.
The Belgian private bus and coach industry counts about 450 different companies, most of which are small to medium
sized. In total, about 12,000 drivers are working in the sector, which comes down to about 10,000 full time equivalents. These
are distributed over three subsectors, i.e. public transport (subcontracts from the regional public transport companies), spe-
cial regular services (mostly school and staff transport) and occasional services (national and international tourism). Collec-
tive bargaining is strongly vested in the industry, making rewards and benefits very similar across firms, although they do
differ between different subsectors. Public transport in Belgium is a legally enforced monopoly of the regional transport
companies that are entirely controlled by their respective regional government. In Flanders, about half and in Wallonia about
a third of the lines are subcontracted out to private companies. Despite reforms in the sector, this situation is not about to
change in the near future since in both regions the public companies have been appointed internal operators in the context
of legislation 1370/2007. In Wallonia subcontracting so far has been based on direct concessions and in Flanders on compet-
itive tendering.1 The companies operating these lines are mostly large or medium-sized companies. So far competition between
the subcontractors has been very weak, preventing pressure on working conditions in this sector. In both regions regulations on
pay and working conditions mimic the ones from the public sector to a very large extend, making them the best paid drivers of
the sector. For practical reasons, drivers working directly for the private companies are not included in our sample. However,
because working conditions are so similar, our conclusions should also be of value for drivers of the public companies. Of course,
we are unable to answer the question whether working in a private or a public company as such is a factor influencing turnover
intentions. This is a limitation of our research, but it doesnt affect the value of our other conclusions.
In the subsector of special regular services contracts are also awarded through tendering. Both large and small companies
are competing on this market that is characterized by a lack of quality requirement and very strong price competition.
Returns are under pressure resulting in less attractive pay and working conditions. Wages are lower, drivers often have tem-
porary contracts and work part-time jobs (which may or may not be voluntary).
Coach drivers (occasional services) have very specific working conditions. The job itself is more divers: whereas in other
sectors drivers almost always do the same circuit, these drivers have very different assignments every day. Often they work
longer hours but earn relatively high wages partly because they frequently get tipped on the job.
The remainder of this paper is organized as follows. We start by explaining our data and econometrics and link our pos-
sible determinants of turnover to the available literature. Subsequently, we present an elaborate discussion of our estimation
results. We conclude with discussing the consequences of our findings for retention policies in the private bus and coach
industry.

2. Data and model

2.1. Sample and survey

The European Union Working Conditions Survey (EWCS) covers almost every ground of the concept of job quality and
would therefore constitute the ideal data source for our purpose (Muoz de Bustillo et al., 2009, 2011). However, its sample
size in Belgium is limited to about a 1000 respondents, making a sector specific analysis of job quality impossible. We there-
fore decided to collect our own data using a questionnaire that was to a large extent based on the one used for the EWCS. The
survey was administered to a random sample of 2915 workers drawn at random from the database of the sectoral social
fund. This fund is an administrative body controlled by a joint committee of workers and employers representatives and
charged with several duties such as the organization of the sectoral pension plan. The database of the fund contains every
worker that has been payed as a bus driver in the Belgian private sector, which makes it an almost perfect reflection of the
sector and thus an unbiased data frame for our sample selection.

1
In 2020, also in Wallonia subcontracting will be based on competitive tendering.
 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259 253

The data collection process was based on the Tailored Design Method (Dillman, 2000). This means, among other things,
that mailings were personalized and that two reminders were sent to the workers that hadnt responded within ten days. In
total 1709 useful responses were received which means a response rate of almost 60% was attained. This is rather high since
average response rates for mail surveys are closer to 50% (Anseel et al., 2010). However, given the effort invested in the col-
lection process and the strong involvement of respondents with the subject of the survey, such a high rate is not unexpected.
We performed a non-response analysis by comparing the sample obtained and our sampling frame on sex, age and
nationality. Differences in nationality and sex were not statistically significant. Differences in age were significant with
the younger age groups (2635 and 3645) being slightly underrepresented. However, these differences were smaller than
2 percentage points which makes us conclude that the sample obtained was representative for the sector.
In order to acquire a homogenous sample of bus drivers we deleted the mechanics and workers for whom being a bus
driver is not their main profession. In the ultimate sample 1280 respondents were retained.

2.2. Determinants

Holtom et al. (2008) offer an overview of turnover models from 1995 to the present. They distinguish individual charac-
teristics, job characteristics, attitudes (such as job satisfaction, involvement and stress), the organizational context and the
person-context interface (e.g. interpersonal relations) as determinants for turnover intentions and actual quit behavior.
Other general studies use different labels, separating economic (objective) from psychological (subjective) determinants:
e.g. Sousa-Poza and Henneberger (2004) in their cross-national comparison, and Bckerman and Ilmakunnas (2009) for Fin-
nish data. Specific for truck drivers Suzuki et al. (2009) summarize (based on an overview of the literature) that turnover rate
is mainly affected by driver demographics, work environment, supervisors and recruiting methods.
Given our data for the Belgian bus and coach industry, we distinguish between three groups of possible determinants
(descriptive statistics are presented in Table 1).

2.2.1. Individual characteristics

Obvious individual characteristics are gender (dummy for women), ethnicity (dummy for native Belgian)2 and age.
Educational level (introduced by three dummies, the reference category being higher secondary education) is included as an
explanatory variable because higher educated workers are assumed to have better labor market alternatives and thus higher
turnover intentions (Sousa-Poza and Henneberger, 2004). The data also provided information about the family situation
(e.g. dependents) but these variables were not significant and left out of the presented results in Section 3.

2.2.2. Objective job characteristics

Objective job characteristics (including job characteristics and organizational context variables from the review of
Holtom et al., 2008) are our second group of explanatory variables. First we control for the different subsectors using a
dummy for special regular services and occasional services (with public transport as reference category). Since some drivers
are active in different subsectors, the subsector variable refers to the activity the drivers exercises most of the time. Further-
more, as Sousa-Poza and Henneberger (2004) point to more job security as a universal determinant of lower turnover inten-
tions (and thus also for the transport sector), we included a dummy for an indefinite contract (reference is a fixed term or day
contract). The dynamic study of two carriers in the US of Suzuki et al. (2009) suggest operational work variables (such as
weekly miles) are predictors of drivers turnover intentions. Therefor we entered the number of hours worked per week,
number of times working on non-conventional hours per month, and dummies for getting informed on work schedules
changes (reference is no changes in work schedule) as explanatory variables in our model. Haines et al. (2010) found that
employer-provided training increases voluntary turnover in Canada. Our data also has information about training but these
variables were not significant for this specific sector and left out the presented results in Section 4.

2.2.3. Subjective job characteristics

Finally, we take a number of subjective job characteristics (including attitude and person fit variables from the review of
Holtom et al., 2008) into account. Violence against transport workers is a growing problem (both more incidents and increas-
ing levels of violence) (Essenberg, 2003; Tse et al., 2006) and also perceived incivility is harmful (Morrow et al., 2011). All
these factors might trigger turnover behavior. Therefore we included a dummy for having been a victim of physical violence
or (sexual or other) harassment.
Many studies consider job satisfaction to be a very important factor (e.g. Clark (2001). We do not include a general job
satisfaction scale, but factor scores from a battery of questions on satisfaction with specific elements of the job environment.
The factors were obtained through Principal Components Analysis with Varimax rotation. Three of the obtained factors can
be related to the Job Demand-Control (JDC) model (Karasek, 1979), one of the most dominant and influential paradigms on
job quality in labor psychology (Wang et al., 2014) which has gained influence within other disciplines such as economics
(Green, 2006). Our first factor relates to the demand dimension (high factor loadings on the following items: experience

2
In the survey respondents were asked whether they considered themselves as native Belgians, as Belgians with an immigration background, as immigrants
or as something else. Because the two non-native Belgian groups were very small, these two groups were merged.
254 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259

Table 1
Descriptive statistics.

Mean/proportion SD Range
Dependent variable
Doesnt consider quitting company nor profession 66.0%
Considers quitting company 16.4%
Considers quitting profession 17.6%
Individual characteristics
Gender
Woman 16.6%
Men 83.4%
Ethnicity
Native Belgian 89.2%
Immigrant 10.8%
Age 49.5 9.1 2177
Education
Lower education 40.0%
Lower secondary education 44.8%
Higher secondary 6.2%
Higher education 8.9%
Objective job characteristics
Subsector
Special regular services 25.1%
Occasional services 16.1%
(public transport = ref) 58.8%
Contract type
Indefinite contract 86.0%
Temporary contract 14.0%
log2 (number of hours) 5.0 0.7 1.06.6
Working non-conventional hours (times/month) 12.6 10.0 0.052
Work schedule changes
Work schedule changes on the day itself 12.0%
Work schedule changes one day in advance 28.4%
Work schedule changes week/month in advance 16.3%
(No changes = ref) 43.4%
Subjective job characteristics
Physical violence or harassment
Victim 37.6%
No Victim 62.4%
FAC_demands (higher = more demanding) 0.0 1.0
2.4 to 3.5
FAC_control (higher = more control) 0.0 1.0
2.3 to 3.6
FAC_fulfillment (higher = better fulfullment) 0.0 1.0
4.5 to 3.0
FAC_worklife (higher = better balance) 0.0 1.0
2.6 to 2.8
FAC_lack-of-support (higher = less support) 0.0 1.0
3.4 to 2.7

stress, enough time to get job done, work at high speed, painful or tiring positions and dealing with angry customers- the last
two items are very specific for the transport sector). An impact of this factor would support the results of Jou et al. (2013) for
(a higher) workload in increasing turnover tendency for air traffic controllers in Taiwan and Tse et al. (2006) for poor cabin
ergonomics as important stressors for bus drivers. The second factor control relates to autonomy in work organizations and
involvement in decisions.
The Karasek models simplicity has been criticized (e.g. Johnson and Hall, 1988), because it disregards some crucial job
dimensions. In a subsequent elaboration of the model, a third dimension was added to the two previous ones: the amount
of social support within the job environment (Karasek and Theorell, 1990). This social support aspect is related to our fifth
factor.
The third factor fulfillment loads high on knowing what is expected and doing useful and gratifying work. This relates to
the intrinsic aspect of work, already identified in 1959 by Herzberg et al. as an important facet. The fourth factor worklife
relates to the work life balance, one of the sub-indices of the European Job Quality index (Leschke and Watt, 2014) and iden-
tified by Jou et al. (2013) as important for air traffic controllers in Taiwan. Because Eurofound (2015) indicates that in the
transport sector the work-life balance is worse than in other sectors, we expect worklife to be an important driver for turn-
over in the Belgian bus and coach industry. The factor loadings (rotated factor matrix) are shown in Table 2.

2.3. Turnover intentions

The survey provided us with information about turnover intentions, not about actual turnover behavior. Studying
turnover intentions is important for a number of reasons. In their meta-analysis Steel and Ovalle (1984) point to a weighted
 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259 255

Table 2
Rotaded components matrix (loadings <0.3 are suppressed).

F1 F2 F3 F4 F5
Demanding Control Worklife Fulfillment Lack-of-
support
Does your job involve tiring or painful positions 0.751
Does your job involve working at very high speed? 0.735
Does your job involve handling angry clients? 0.731
Do you experience stress in your work? 0.693
Do you have enough time to get the job done?
0.563 0.379
Can you influence decisions that are important for your work? 0.857
Are you able to apply your own ideas in your work? 0.838
Are you involved in improving the work organization or work processes of your 0.695 0.363
department or organization?
Would you say that for you arranging to take an hour or two off during working hours 0.772
to take care of personal or family matters is . . . ?
In general, do your working hours fit in with your family or social commitments 0.740
outside work?
Do you have the feeling of doing useful work? 0.808
Does your job gives you the feeling of work well done?
0.309 0.325 0.673
Do you know what is expected of you at work? 0.659 0.330
Do your colleagues help and support you? 0.857
Does your manager help and supports you? 0.329 0.441 0.495

average correlation of 0.5 between intentions and actual turnover. A more recent study of Bckerman and Ilmakunnas (2009)
finds that 42% of those who wanted to quit also actually leave their job during the four subsequent years. In other words,
turnover intensions are a good proxy for actual turnover behavior. Moreover, investigating intentions rather than retrospec-
tively assessing quitting behavior has the advantage of tapping the here and now factors that determine this critical deci-
sion (Knudsen et al., 2003, p. 129).
We measure turnover intention by asking to what extent, in the course of the last six months, the drivers considered
changing jobs. For those who did (30% of the sample), we asked whether they would want to do the same job in another
company (60%) or a different kind of job (in the same or another company - 40%). Our dependent variable was constructed
as a combination of these two variables. We distinguish between three types of respondents, i.e. those who did not consider
changing jobs nor companies (66% of the sample), those who considered changing companies but wanted to remain bus dri-
vers (16.4% of the sample) and those who considered quitting the bus driver profession (17.6% of the sample).

2.4. Model

In Table 3 we present the results of a multinomial logit model. This type of model is used since there are more than two
alternatives without natural ordering (Greene, 2003). Alternative models provide the same intuitions.3 The relative impor-
tance of the predictors from the multinomial model are determined by calculating the absolute value of the pseudo partial cor-
relation coefficient using the formula suggested by Bhatti et al. (2006):
r
Wald statistic
2K
jRj

2LL
where K refers to the degrees of freedom associated with the Wald statistic. If the Wald statistic is less than two times K, the
value is set to zero. These values estimate the unique contribution of an independent variable to the variation of a dependent
variable, and are therefore much lower than standard r-values. Since higher values are indicative of a stronger marginal asso-
ciation between the predictor and the outcome, the value can be used to rank the predictors in order of most-to-least impor-
tance. The results of the relative importance analysis are presented in Table 4.

3. Estimation results

In Table 3 we find the result of the multinomial logit analysis. The overall explanatory power of our model was satisfac-
tory, with Nagelkerke R-squared = 38.6%. In the table we notice that women are almost twice (1.862) as likely as men to con-
sider quitting their company. However, men and women are about equally as likely to consider quitting their profession.
Likewise, workers with an immigration background are almost twice (1/0.528) as likely as native Belgians to consider
quitting their company, but there is no statistically significant difference between both groups in considering to quit the

3
We also estimated (1) an ordinal logit model to explain the extent to which the worker considers changing jobs (never, occasionally or regular) and (2) a
binary logit model considering only those who want to change, whether they want to change company or job. The presented multinomial logit model
incorporates the results of the alternative models: it indicates who wants to change and what the different drivers are for changing job/company. The additional
results are available on request from the authors.
256 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259

Table 3
Multinomial logit results for considering to company or profession (ref. cat.: doesnt consider quitting company nor profession).

Consideres quitting company Consideres quitting profession

B p OR B p OR
Intercept
0.031 0.981
0.219 0.867
Individual characteristics
Gender
Woman 0.622 0.021 1.862 0.008 0.979 1.008
(Men = ref.)
Ethnicity
Native Belgian
0.639 0.037 0.528 0.246 0.481 1.278
(Immigrant = ref.)
Age
0.047 0.000 0.954
0.059 0.000 0.943
Education
Primary education 0.351 0.303 1.421 0.328 0.341 1.389
Lower secondary education 0.067 0.760 1.069 0.114 0.592 1.120
(Higher secondary = ref.)
Higher education 0.255 0.539 1.291 0.234 0.560 1.264
Objective job characteristics
Subsector
Special regular services 0.356 0.304 1.427 0.719 0.034 2.053
Occasional services 0.460 0.194 1.584 0.314 0.382 1.368
(public transport = ref)
Contract type
Indefinite contract
1.431 0.000 0.239
0.697 0.073 0.498
(Temporary = ref)
log2 (number of hours) 0.331 0.152 1.392 0.290 0.205 1.337
Working non-conventional hours (times/month)
0.011 0.363 0.989
0.012 0.294 0.988
Work schedule changes
Work schedule changes on the day itself 1.116 0.000 3.053 0.286 0.372 1.331
Work schedule changes one day in advance 0.497 0.056 1.644 0.091 0.715 1.095
Work schedule changes week/month in advance 0.212 0.516 1.237
0.069 0.811 0.933
(No changes = ref)
Subjective job characteristics
Physical violence or harassment
Victim 0.519 0.014 1.681 0.450 0.028 1.569
(No victim = ref.)
FAC_demands (higher = more demanding) 0.409 0.000 1.506 0.800 0.000 2.225
FAC_control (higher = more control)
0.554 0.000 0.575
0.425 0.000 0.654
FAC_fulfillment (higher = better fulfillment)
0.458 0.000 0.633
0.884 0.000 0.413
FAC_worklife (higher = better balance)
0.636 0.000 0.530
0.649 0.000 0.523
FAC_lack-of-support (higher = less support)
0.581 0.000 0.559
0.258 0.011 0.772

Table 4
Relative importance of predictors in the multinomial logit model.

Consideres quitting company Consideres quitting profession

Wald df |r| Rank Wald df |r| Rank
Individual characteristics
Gender 4.694 1 0.065 10 0.007 1 0.000 12
Ethnicity 2.926 1 0.038 11 0.408 1 0.000 12
Age 16.780 1 0.152 5 27.999 1 0.196 4
Education 2.530 3 0.000 12 0.626 3 0.000 12
Objective job characteristics
Subsector 2.465 2 0.000 12 5.376 2 0.045 8
Contract type 15.818 1 0.147 6 2.875 1 0.036 10
log2 (number of hours) 1.653 1 0.000 12 3.278 1 0.043 9
Working non-conventional hours (times/month) 0.505 1 0.000 12 2.181 1 0.016 11
Work schedule changes 12.142 3 0.098 8 1.945 3 0.000 12
Subjective job characteristics
Physical violence or harassment 6.749 1 0.086 9 4.346 1 0.059 7
FAC_demands (higher = more demanding) 12.010 1 0.125 7 44.854 1 0.252 2
FAC_control (higher = more control) 21.972 1 0.177 3 12.662 1 0.126 5
FAC_fulfillment (higher = better fulfillment) 19.916 1 0.167 4 69.075 1 0.315 1
FAC_worklife (higher = better balance) 30.743 1 0.212 1 37.465 1 0.229 3
FAC_lack-of-support (higher = less support) 26.071 1 0.194 2 6.395 1 0.081 6
 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259 257

profession. In other words, women and immigrants are not more likely to consider quitting their job as a driver, but they are
more likely to consider being a driver in another company. It is possible that in some companies female drivers and immi-
grant drivers experience problems because of their gender and migration background and are therefore more likely to con-
sider changing companies. Since there is a shortage of drivers in the industry (VDAB, 2015) they are probably rather likely to
be hired again since they already have the appropriate licenses and certificates. The fact that they are not inclined to quit the
profession is in line with their lower access to the formal Belgian labor market: in 2010 the unemployment rate for EU
nationals (24.7%) was five times higher than for Belgian nationals (4.7%) (VDAB, 2012).
Both the likelihood of considering to change companies and considering to change professions is lower for older workers.
The effect is stronger for quitting the profession: when drivers get a year older, their odds for considering to leave their com-
pany go down with 4.8% (1/0.954) and their odds for considering to change professions with 6.1% (1/0.943). This finding is
consistent with the literature (Holtom et al., 2008) and reflects the fact that the expected benefits of changing jobs and/or
companies diminish as the expected time one will remain active on the labor market goes down (whereas the expected costs
remain equal or increase).
Although many employers believe higher educated drivers have better job alternatives and are thus more inclined to
change jobs, there was no evidence of an effect of educational level on turnover intentions within our sample of Belgian
bus drivers. This finding is in line with research findings on the population of employees in general that couldnt demonstrate
a link between education and turnover either (Sousa-Poza and Henneberger, 2004).
Drivers that are mainly active in special regular services are twice as likely to consider quitting their job as a bus driver. As
explained above, this subsector is the one with the highest price competition and the least attractive working conditions.
Since this is a general fact over all companies, it is logical drivers in special regular services do not differ in their inclination
to leave the company. There also was no significant difference between public transport and occasional services, both for
leaving the profession and leaving the industry. Both subsectors offer advantages and disadvantages, and most of the times
drivers do seem to find the subsector that fits them best.
We did observe a significant effect of contract type. People with temporary contracts are more than four times (1/0.239)
as likely to consider leaving their company and two times (1/0.498) as likely to consider quitting the profession. The asso-
ciation between (the binary log of) the number of hours worked and the likelihood to consider quitting the company or leav-
ing the profession is not at all significant. Also, the degree to which one works non-conventional hours does not seem to
influence our dependent variable. On the other hand, work schedule changes do have an effect on the likelihood to consider
leaving the company (in line with the importance of operational work variables as pointed to by Suzuki et al. (2009)). How-
ever, not the changes themselves but the communication about these changes seem to be paramount. Workers whose sched-
ule changes the day itself are more than three times as likely to consider leaving their employer whereas those whose
schedule changes one day in advance are only about one and a half times as likely. There is no difference between workers
confronted with work schedule changes occurring at least a week in advance and workers experiencing no changes.
Although last minute schedule changes do affect the drivers likelihood to change companies, they dont affect the likelihood
of considering to quit the bus driver profession.
Having been a victim of physical violence or harassment significantly affects the odds of considering to quit the company:
the odds of quitting are more than one and a half times as high for victims as for non-victims. As far as the odds for consid-
ering to quit the profession are concerned, the difference between victims and non-victims is slightly smaller.
The five factors are all significantly related to both the odds of considering to change employers and considering to change
professions. Workers perceiving their job to be more demanding are more likely to consider quitting their company and even
more so to consider quitting the bus driver profession. Drivers experiencing more control in their job are less likely to quit
their profession and even less likely to quit their company. Experiencing more fulfillment diminishes the likelihood to con-
sider leaving the profession stronger than it diminishes the likelihood to consider leaving the company. The effect of the
work-life balance is about equally as strong for quitting the employer as for quitting the profession. Finally, experiencing
a lack of support has the strongest effect on the likelihood to consider changing employers.
Table 4 presents the results of the relative importance analysis. The most important drivers of the likelihood to consider
changing companies seem to be the subjective job characteristics. The number one driver is the work-life balance, followed
by the lack-of-support factor. The fulfillment factor ranks forth, whereas the control factor ranks third. The subjective char-
acteristics are also the most important drivers for considering to quit the profession, but the order is different. The most
important factor here is the fulfillment factor, followed by the demands and the work-life balance. Job control ranks fifth,
the support factor seventh. The physical violence factor seems to be less important as a driver, ranking only ninth and sev-
enth for considering to leave the company and the profession respectively. Both education and working non-conventional
hours are ranked last because for both variables the Wald statistic was smaller than two times the degrees of freedom, which
results in a pseudo partial correlation coefficient of (exactly) 0 (cf. supra).
Of all the objective job characteristics, the contract type is the most important driver. For considering to change employ-
ers the contract type comes on place six. For considering to leave the profession it is only ranked tenth. Work schedule
changes are ranked eight for considering to change employers, but are not important for considering to leave the profession.
Non-conventional hours and the number of hours are of lesser importance.
Although we saw earlier that women and migrants had a bigger chance of leaving their company, these personal char-
acteristics do not seem to be the most dominant drivers of company turnover. Of all the personal characteristics, age seems
to be the most important one.
258 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259

4. Discussion and conclusions

In this paper we identify the key determinants of driver turnover intentions in the Belgian private bus and coach industry.
Our study is innovative in that sense that we distinguish between bus drivers intentions to turn away from their employer
and their intention to turn away from their profession. So we encompass the findings of Delfgaauw (2007), who has found
diverging factors predicting company versus profession turnover for the public sector, to a specific private industry.
Our research results show that individual characteristics do play a role in turnover intentions, although the explanatory
power of these factors is not very large. As has been shown for other jobs and industries, when people grow older their will-
ingness to change companies and jobs diminishes. In our sample the odds of women to consider changing companies was
twofold that of men. Likewise, the likelihood of drivers with an immigration background to consider leaving their employer
was twice as large as that of native Belgians. However, there was no effect of gender or ethnicity on drivers inclination to
change professions. So, although women and migrants like their job as a driver as much as men and native Belgians, they
seem to be less attached to their company. A possible explanation of this finding is that drivers attribute potential experi-
ences with racism or sexism to the company environment. However, this does not suggest that discrimination (solely) comes
from colleagues and/or employers. Since different companies tend to drive in different regions, difficult costumers on diffi-
cult lines may also be regarded as a part of the company environment.
Educational level did not have a significant effect on turnover intentions. Most of the drivers in the industry did not finish
secondary education and can thus in a Belgian context be regarded as lower educated. However, a minority of the drivers did
complete secondary or even tertiary education. Often employers expect those drivers to have higher odds of wanting to
change jobs because they have better alternatives on the job market. However, our analysis did not confirm this idea. The
reasons why higher educated workers end up doing the job of a driver are divers. Some have experienced difficulties finding
a job at their educational level, some have had personal problems or just discovered that a desk job didnt fit with their aspi-
rations and personality. Many higher educated driver have chosen their job for a good reason (whether positive or negative)
and are therefore not more likely than others to change jobs.
Most of the objective job characteristics were not strongly related to turnover behavior, with the notable exception of
contract type (ranking third in the relative importance analysis for considering to change companies). As a general rule,
workers are inclined to look for job security and their contract type can play an important role in that regard. Last notice
changes in working schedules were related to the likelihood to consider quitting the company. In an increasingly competitive
environment, many companies are obliged to accept last minute demands from clients. Especially in smaller companies
where the pool of drivers is not that large, this may demand considerable flexibility on the part of the drivers. Certain time-
table management systems might be a part of the solution. However, giving the difficult market environment in which a lot
of companies have to operate, this will not be easy to implement. Although time table changes were related to the likelihood
of changing companies, they were not associated with the likelihood of changing professions. We interpret this result as an
illustration that drivers are convinced these very last minute changes do not need to be inherent to the profession.
The subjective job characteristics were the most important predictors of turnover intentions, both regarding changing
employers and changing jobs. A positive work-life balance and the experience of social support from the working environ-
ment (employers and co-workers) are the key factors in convincing drivers to stay in their current company. The degree to
which the job is experienced as rewarding (in the sense of having the idea to be doing good and useful work) seemed to be
the dominant factor in determining a drivers intention to stay in the profession. Experiencing a demanding job environment
(i.e., physical burden, pressure from time constraints, etc.) was the second important factor. The job control factor only came
in fifth for both considering to change companies and considering to change jobs. In the general literature this factor is often
given a more prominent role (e.g. Karasek, 1979; Clark, 2001; Origo and Pagani, 2008). However, the job of a bus driver is to a
large extent bound to limit the control of the drivers over their working environment. This well-known fact is probably dis-
couraging people with a relatively higher desire to control their working environment to enter the sector in the first place.
In their paper, Jones et al. (2013) ask themselves the question whether bus driving can ever be a good job. Although, as we
outlined above, the profession is confronted with a lot of inherent difficulties, we believe opportunities for improvement do
exist. The importance of the fulfillment factor shows that drivers have a strong desire for being appreciated for their work.
Our analysis suggest that initiatives on this front from the company management and from other actors in the sector can
make a big difference. Additionally, innovative policies, such as the new way of dealing with time tables in congested areas
analyzed by Hlotova et al. (2014), can render the most difficult parts of the job more bearable. We hope our analysis will
inspire the different actors in the sector, both in Belgium and abroad, to start experimenting with such endeavors.

Acknowledgement

This research project was supported by a grant from the European Social Fund (ESF).

References

Anseel, F., Lievens, F., Schollaert, E., Choragwicka, B., 2010. Response rates in organizational science, 19952008. A meta-analytic review and guidelines for
survey researchers. J. Bus. Psychol. 25, 335349.
Bhatti, I.P., Lohano, H.D., Pirzado, Z.A., Jafri, I.A., 2006. A logistic regression analysis of ischemic heart disease risk. J. Appl. Sci. 6 (4), 785788.
 S. Lannoo, E. Verhofstadt / Transportation Research Part A 91 (2016) 251259 259

Bckerman, P., Ilmakunnas, P., 2009. Job disamenities, job satisfaction, quit intentions and actual separations: putting the pieces together. Indust. Relat. 48
(1), 7396.
Clark, A.E., 2001. What really matters in a job? Hedonic measurement using quit data. Labour Econ. 8, 223242.
Delfgaauw, J., 2007. The effect of job satisfaction on job search: not just whether, but also where. Labour Econ. 14, 299317.
Dewit, H., 2013. De Toekomstige Competenties in De Arbeidsmarkt Van De Sector Van De Autobus- En Autocarondernemingen. Fonds Sociaal Fonds
Bus&Car, Brussels.
Dillman, D.A., 2000. Mail and Internet Surveys: The Tailored Design Method, second ed. John Wiley Co., New York.
Essenberg, B., 2003. Violence and Stress at Work in the Transport Sector. Working Paper 205, International Labour Office Geneva.
Eurofound, 2015. Transport and Storage Sector: Working Conditions and Job Quality. Information sheet EF138433.
Green, F., 2006. Demanding Work: The Paradox of Job Quality in the Affluent Economy. Princeton University Press, Princeton and Oxford.
Greene, W.H., 2003. Econometric Analysis, fifth ed. Prentice Hall, Upper Saddle River, NJ.
Haines, V.Y., Jalette, P., Larose, K., 2010. The influence of human resource management practices on employee voluntary turnover rates in the Canadian non
governmental sector. Ind. Labor Relat. Rev. 63 (2), 228246.
Herzberg, F., Mausner, B., Snyderman, B., 1959. The Motivation to Work. John Wiley, New York.
Holtom, B.C., Interrieden, E., 2006. Integrating the unfolding model and job embeddedness to better understand voluntary turnover. J. Manage. Issues 18 (4),
435452.
Holtom, B.C., Mitchell, T.R., Lee, T.W., Eberly, M.B., 2008. Turnover and retention research: a glance at the past, a closer review of the present, and a venture
into the future. Acad. Manage. Ann. 2 (1), 231274.
Hlotova, Y., Cats, O., Meijer, S., 2014. Measuring bus drivers occupational stress under changing working conditions. Transp. Res. Rec. 2415, 1320.
Johnson, J., Hall, E., 1988. Job strain, work place social support, and cardiovascular disease: a cross-sectional study of a random sample of the Swedish
working population. Am. J. Public Health 78 (10), 13361342.
Jones, W., Haslam, R., Haslam, C., 2013. Bus driving can it be a good job? Contemporary Ergonomics and Human Factors. In: Contemporary Ergonomics and
Human Factors: Proceedings of the International Conference on Ergonomics & Human Factors 2013. Taylor & Francis, Cambridge, UK, p. 69, 1518 April
2013.
Jou, R., Kuo, C., Tang, M., 2013. A study of job stress and turnover tendency among air traffic controllers: the mediating effects of job satisfaction. Transp. Res.
Part E 57, 95104.
Karasek, R., Theorell, T., 1990. Healthy Work: Stress, Productivity and the Reconstruction of Working Life. Basic Books, New York.
Karasek, R.A., 1979. Job demands, job decision latitude and mental strain: implications for job redesign. Admin. Sci. Quart. 24, 285308.
Knudsen, H.K., J. Johnson, J.A., Roman, P.M., 2003. Retaining counseling staff at substance abuse treatment centers: effects of management practices. J. Subst.
Abuse Treat. 24, 129135.
Lazear, E., Spletzer, J., 2012. Hiring, Churn and the Business Cycle. NBER Working Paper.
Leschke, W., Watt, A., 2014. Challenges in constructing a multi-dimensional European job quality index. Soc. Indic. Res. 118, 131.
Lvy-Garboua, L., Montmarquette, C., Simonnet, V., 2007. Job satisfaction and quits. Labour Econ. 14 (2), 251268.
Maertz, C.P., Campion, M.A., 2004. Profiles in quitting: integrating process and content turnover theory. Acad. Manag. J. 47 (4), 566582.
Morrow, P.C., McElroy, J.C., Scheibe, K.P., 2011. Work unit incivility, job satisfaction, and total quality management among transportation employees.
Transp. Res. Part E 47, 12101220.
Muoz de Bustillo, R., Fernndez-Macas, E., Ignacio Antn, J., Esteve, F., 2009. Indicators of job quality in the European Union. Study on behalf of the
European Parliaments Committee on Employment and Social Affairs. Brussels: European Parliament.
Muoz de Bustillo, R., Fernndez-Marcas, E., Fernando, E., Ignacio Antn, J., 2011. E pluribus unum? A critical survey of job quality indicators. Socio-Econ.
Rev. 9, 447475.
Origo, F., Pagani, L., 2008. Workplace flexibility and job satisfaction: some evidence from Europe. Int. J. Manpower 29 (6), 539566.
Sousa-Poza, A., Henneberger, F., 2004. Analyzing job mobility with job turnover intentions: an international comparative study. J. Econ. Issues 38 (1), 113
137.
Steel, R.P., Ovalle, N.K., 1984. A review and meta-analysis of research on the relationship between behavioral intentions and employee turnover. J. Appl.
Psychol. 69 (4), 673686.
Suzuki, Y., Crum, M.R., Pautsch, G.R., 2009. Predicting truck drive turnover. Transp. Res. Part E 45, 538550.
Tse, J.L.M., Flin, R., Mearns, K., 2006. Bus driver well-being review: 50 years of research. Transp. Res. Part F 9, 89114.
VDAB, 2012. Kansengroepen in Kaart, Allochtonen op de Vlaamse arbeidsmarkt Retrieved from: <https://www.vdab.be/sites/web/files/doc/trends/KiK_
allochtonen_2012.pdf>.
VDAB, 2015. Knelpuntberoepen: Kansenberoepen Retrieved from: <https://www.vdab.be/trendsdoc/vacatureanalyse/vacatureanalyse_2015.pdf>.
Wang, M., Mykletun, A.K., Myner, S., et al, 2014. Job strain, health and sickness absence: results from the hordaland health study. PLoS ONE 9, e96025.