Multiple jobholding in Europe: features and effects of primary job quality

Abstract

German

French

This article investigates the relationship between job quality features of the primary job and the propensity to engage in multiple paid activities. The analysis covers workers from 28 European countries using data from the EWCS (European Working Conditions Survey) 2010-2015. The results show that workers experiencing economic and job insecurity in their primary job are more likely to engage in additional paid employment. Multiple jobholders (MJHs) report higher work pressure and more unsocial hours in their main jobs, but also more control over and flexibility in working hours, more autonomy and a wider scope for exercising skills, the latter largely explained by compositional factors. Moreover, the evidence shows that experiences of work diverge among multiple jobholders, and they can be classified into six clusters based on the quality of their primary job. This points to a variety of motivations and factors that encourage multiple jobholding. Finally, we find a considerable cross-country variation in job quality among multiple jobholders, with worse outcomes in more segmented labour markets with a higher proportion of non-standard employment.

Keywords

Multiple jobholding moonlighting job quality

Introduction

Multiple jobholding, i.e., holding more than one paid job simultaneously (Campion et al., 2020) – in the literature also called moonlighting or dual jobholding – is a common and growing phenomenon in many European countries (Conen, 2020; Piasna and Drahokoupil, 2017). Its expansion has been linked to a general decline in the standard employment relationship following the flexibilisation of labour markets and an increasing fragmentation of work (Rubery, 2015), with future growth fuelled by digitalisation and the rise of the gig economy. However, reasons for engaging in multiple paid activities are complex, and experiences of work among multiple jobholders (hereafter – MJHs) are far from clear. Additional paid employment is generally thought to be taken up in order to compensate for poor-quality features of the main job. Such job quality deficits may manifest as low income or an insufficient number of working hours to maintain the expected standard of living (e.g. Panos et al., 2014), but may also involve job insecurity (Bell et al., 1997), a poor skills match between a worker’s competences and the requirements of a current job, or a lack of task variety (Dickey et al., 2011). All these factors may coexist, possibly with different degrees of importance depending on worker and work characteristics, but this strand of literature leads to a general expectation that MJHs are more often found among workers with poor quality jobs, at least in some respect. On the other hand, it is also plausible that multiple jobholding is fuelled by a motivation to enrich one’s work experience (Wu et al., 2009), build a CV or develop personal interests, thus factors not directly related to negative characteristics or the poor quality of the main job. On the contrary, main jobs that offer employee-oriented flexible working hours, are not overly stressful or allow enough autonomy, in other words jobs that are of better quality, might facilitate juggling them with another paid employment. Existing studies support these presumptions to varying degrees, yet their predominantly single-country focus or case-study character limits the generalisability of the findings. We thus lack knowledge about experiences and working conditions among MJHs across countries and different groups of workers.

This study adds to existing knowledge by examining job quality among MJHs across the entire labour force and from 28 European countries (EU-27 and the UK). The objective is twofold. First, we examine what job quality features of the main paid job are associated with multiple jobholding. To that end, we analyse differences in job quality between MJHs and single jobholders (hereafter – SJHs), i.e., workers who have only one paid job. Second, we investigate differences in job quality within the group of MJHs in order to determine whether they share similar work experiences or instead form several heterogeneous groups, or job quality clusters. We approach job quality as a multi-dimensional concept and analyse it along six dimensions: income, job security and prospects, work pressure, skills and autonomy, unsocial hours, and working time control and flexibility. Such a comprehensive measurement of job quality allows us to identify a very broad range of factors that may encourage multiple jobholding and to assess their respective impacts. Finally, the wide geographic coverage of our data enables us to explore differences in the prevalence of MJH job quality clusters across 28 European countries.

The article begins with a review of previous studies that investigated motivations for holding multiple jobs, with a special emphasis on theory and research bridging job quality and multiple jobholding. Next, we present the methodology of the empirical analysis, which applies multi-level modelling and hierarchical cluster analysis to the data from two waves of the European Working Conditions Survey (2010-2015). The presentation of the empirical results is structured in two parts: a comparison of differences in job quality between MJHs and SJHs, and an analysis of job quality profiles among MJHs. Job quality profiles are compared on the basis of various worker and work characteristics, and across countries. The article closes with a summary of findings and some recommendations for policy and further research.

Job quality and multiple jobholding: a review of the literature

In recent years, the issue of job quality has been high up on the research agenda for various reasons. In general, they all relate to a growing imbalance between capital and labour, with the latter losing the bargaining power, which widens the possibilities for the former to reduce labour costs, and ultimately creates the risk of job quality deterioration. There have been attempts to counterbalance those processes with public policy measures and recommendations (e.g. the ILO’s Decent Work Agenda or the EU’s European Employment Strategy); however, the expansion of ‘bad jobs’ in the sense proposed by Kalleberg et al. (2000) continues across the EU (Broughton et al., 2016; Eurofound, 2017; Grimshaw et al., 2016; Piasna, 2017). A growing number of workers are engaged in activities characterised by low pay, short-term contracts, job insecurity, a lack of skills development opportunities or limited access to social protection. It is important to note that both non-standard and standard employment (permanent and full-time) can be precarious (Dörre, 2019; Grimshaw et al., 2016). Indeed, the rise of low-quality jobs has resulted from both the growth of non-standard and insecure forms of work (Broughton et al., 2016; Eichhorst and Marx, 2015; Kretsos and Livanos, 2016) and the declining quality of standard jobs (Dekker and Van der Veen, 2017; Grimshaw et al., 2016).

Deteriorating job quality and the expansion of bad jobs also feature in the ever-present debate on labour market segmentation (Emmenegger et al., 2012). There is a stream of literature written from a variety of perspectives (e.g. Guillén and Dahl, 2009; Piasna and Myant, 2017) that are all critical of the one-sided, neoclassical view of the labour market where quantitative measures (mostly unemployment and employment rates) overshadow substantive characteristics of work, and they are likewise critical of policies based on this view which ultimately deepen labour market segmentation and inequalities, leading to the ‘precarisation’ of growing parts of the workforce (Doellgast et al., 2018; Kalleberg, 2009; Standing, 2011).

In general, negative tendencies in the area of job quality are quite common, even if variations across countries persist (Antón et al., 2012; Green et al., 2013; Prosser, 2016). Some tendencies emerged – paradoxically – in the years that saw the development of affluent and growing economies (Green, 2006), while others are more recent (Isidorsson and Kubisa, 2018); all intensified with the outbreak of the 2008 financial crisis and the introduction of austerity policies (Myant et al., 2016; Vaughan-Whitehead, 2015). Finally, the COVID-19 outbreak and measures taken by national governments to contain the spread of the pandemic may have a profound impact on employment and work quality. They may lead to lasting changes in the functioning of workplaces, work organisation and employment relations, as well as job losses, reductions in working hours and the deteriorating situation of vulnerable workers.

Prior literature has frequently linked multiple jobholding to levels of and developments in job quality, particularly that of the primary job. However, various research studies have shed light on the different aspects of job quality that can come into play. As a starting point, several contributions argue that multiple jobholding is dominant among workers with a low-quality primary activity, and they highlight financial necessity and declining earnings as a main motive. Since low earnings can put workers in a precarious situation and create socio-economic vulnerability (Bamberry and Campbell, 2012; Sliter and Boyd, 2014), moonlighting emerges from financial stress as a strategy to tackle financial constraints, to make ends meet or simply to raise standards of living (Wu et al., 2009). In line with such an interpretation, Kimmel and Conway (2001) found that, among prime-age men in the US, MJHs were, on average, poorer than other workers. A financial motivation for engaging in multiple jobholding was found even among a relatively well-paid group of oil and gas workers in the UK (Dickey et al., 2011), with those facing financial difficulties or more financial commitments more likely to engage in an additional paid activity. A more recent investigation by Panos et al. (2014), using panel data on UK workers, showed that the incidence of multiple jobholding closely followed developments in unemployment, and it increased in response to financial shocks. This suggests that growing financial uncertainty triggers multiple jobholding.

The financial motive has also been found to influence self-employed persons whose business does not generate sufficient or stable income: in such cases, taking a second job may be a strategy for sustaining the primary self-employment (Henley, 2007; Wu et al., 2009). This tendency increased after the 2008 financial crisis, exerting a strong pressure on the financial viability of many own-account businesses (Atherton et al., 2016). Moreover, it can be expected to increase further still in the aftermath of the COVID-19 crisis, with the self-employed and micro-enterprises in the service sector being particularly hard hit.

Financial motivation, caused by low or insufficient income in the primary job, is often the result of working-hour constraints or hours-based underemployment (Paxson and Sicherman, 1996). Workers decide to hold a second job (or more) where they want or need to work more hours but their primary job does not offer such a possibility (Dickey et al., 2011; Wu et al., 2009). In line with this interpretation, multiple jobholding was found to be common among involuntary part-timers in the UK (Green and Livanos, 2015). Similarly, Panos et al. (2014) showed that, among male adult workers in the UK, those holding two paid jobs were significantly more likely to want to work more hours in their primary job than single jobholders.

Moreover, several studies have emphasised the relationship between multiple jobholding and job insecurity, where the latter is defined as the perception of a threat to the continuity of one’s employment situation. Indeed, an analysis of 29 European countries between 1998 and 2011 revealed that both the incidence and intensity of multiple jobholding increased as labour markets became more volatile and precarious jobs more prevalent (Zangelidis, 2014). The job insecurity model suggests that individuals may engage in a second job (or in multiple jobs) if they believe, for various reasons, that they might lose their primary job (Bell et al., 1997). Multiple jobholding is believed to increase the probability of securing a new job and to decrease the probability of becoming unemployed (Panos et al., 2014; Pouliakas, 2017). In other words, it is felt that taking on more than one paid job serves as ‘insurance’ to offset the risk of losing the primary employment (Bell et al., 1997) or as a means of exploring and diversifying options of remaining in the labour market, including through business start-ups (Folta et al., 2010).

Recent studies also show a higher propensity to multiple jobholding among workers with non-standard (primary) employment (see, for instance, Conen, 2020). Although both standard and non-standard jobs can be insecure, there is a strong association between atypical forms of employment and insecurity (Broughton et al., 2016; Kretsos and Livanos, 2016). Therefore, according to this strand of the literature, the above-mentioned recent developments in job quality, especially the deterioration in job security and earnings, are likely to be contributing significantly to the increase in moonlighting.

However, the relationship between job security in the main job and multiple jobholding is probably more complicated and context-specific than this. For instance, the hypothesis about multiple jobholding as a hedge against insecurity in the primary job was tested and refuted by two studies using British large-scale panel data (Böheim and Taylor, 2004; Wu et al., 2009), especially when considered in connection with subjective evaluations of job security. Wu et al. (2009) even suggest that multiple jobholding tends to be more prevalent among public sector employees and workers with standard and secure forms of employment. A possible explanation for this is that the stability offered by the primary job encourages the pursuit of entrepreneurial activities as a secondary activity (Kimmel and Conway, 2001).

Nevertheless, it is likely that multiple jobholding does not arise solely from job quality deficits in the primary job, in particular a precarious financial situation or job insecurity for which an additional paid activity is meant to compensate. Indeed, other, non pecuniary reasons were found to gain in importance once people have a more secure financial situation (Dickey et al., 2011). Accordingly, holding multiple jobs can be a strategy aimed at improving or enhancing skills, gaining new competences and expertise, or exploring alternative occupations, new career paths or opportunities, including a self-employed career (Paxson and Sicherman, 1996; Panos et al., 2014). One reason for taking up additional jobs can be to increase task variety or acquire new credentials (Dickey et al., 2011). The incentive for moonlighting in the UK, apart from financial pressures, has been shown to stem from the desire for heterogeneous tasks (Wu et al., 2009). Engagement in more than one paid job also serves to increase job satisfaction, enhance professional status or derive additional non-monetary utility not provided by the first job (Böheim and Taylor, 2004). Bamberry and Campbell (2012), in their Australian small-scale qualitative study, reported that, at least for some workers, a second job was a means to improve their satisfaction with work and their capacity to balance the demands of paid work with family or social obligations. Moreover, MJHs are also likely to be found among workers with good-quality primary jobs. For instance, Wu et al. (2009) argue that workers who are satisfied with the working hours in their main job are more likely to take up additional jobs.

Based on the reviewed evidence, which indicates many possible links between job quality and multiple jobholding, yet without providing definitive answers, we formulate the first research question:

RQ1: What job quality features of the main job are associated with multiple jobholding?

The reviewed studies provide a multitude of – at times conflicting – explanations for the motives behind multiple jobholding, but the comparability of the results is limited, as each study analysed only one specific socio-demographic group of workers or one particular country. Thus, it is very likely that the varied reasons for multiple jobholding and the diverse experiences of MJHs all coexist and that they depend on the individual work situation. Bouwhuis et al. (2018), in their study of older MJHs in the Netherlands, found that, even among those workers with relatively similar demographics, experiences ranged from positive to negative and were influenced by factors such as the type of contract, financial situation or flexible work arrangements, hence aspects of job quality in both their primary and secondary jobs. This suggests a heterogeneity of working conditions and socio-economic status among MJHs (Conen, 2020).

The job quality literature has long argued that jobs can be classified into distinct clusters, or types, according to various job quality characteristics. While a classic study by Karasek and Theorell (1990) distinguished four types of jobs based on trade-offs between the psychological demands of work and a combined measure of task control and skill use, more recent studies using a wider range of job quality features identified five (Eurofound, 2016) or six (Holman, 2013) job quality profiles among European workers. Building on this literature, we formulate the second research question:

RQ2: Do MJHs share similar work experiences, or do they instead form several heterogeneous job quality clusters?

Finally, since working conditions and experiences of work differ across countries, we explore the final research question:

RQ3: Do patterns of job quality among MJHs (i.e., uncovered job quality clusters) display any cross-country variations?

Previous studies have revealed striking differences in the quality of work between European countries, driven by differences in production regimes, institutional systems of employment regulation, industrial relations or levels of prosperity (Antón et al., 2012; Eurofound, 2016; Gallie, 2007; Green et al., 2013; Prosser, 2016). Cross-county differences in job quality among MJHs could thus broadly reflect variations in the national levels of job quality and be driven by similar factors. However, some institutional features that vary across countries and labour markets can also have a specific impact on the structures and dynamics of moonlighting (Conen, 2020; Hirsch et al., 2016). For instance, based on the literature on varieties of capitalism, a relatively high prevalence of poor job quality among MJHs can be expected in countries with highly segmented labour markets, such as in southern European countries. In such countries, a high proportion of precarious jobs and vulnerable workers, weak intervention of the state in supporting innovation and high-road paths to competition, and less effective policies and social protection systems can all increase the importance of multiple jobholding as a compensatory strategy for precariousness in primary jobs (Molina and Rhodes, 2007). Similar outcomes might be expected in highly dualised capitalism models (Thelen, 2014), as well as countries where the prevalence of non-standard employment is linked to higher ‘protection gaps’, particularly in terms of social protection and integration (Grimshaw et al., 2016). However, a more detailed explanation of the drivers of cross-country differences is beyond the scope of this article, as it necessitates a specific comparative and in-depth analysis.

Data and methods

Our analysis uses data from the European Working Conditions Survey (EWCS) carried out by Eurofound. The EWCS sample is representative of persons in employment who worked for at least one hour during the week that preceded the interview (employees and self-employed). Country samples contain typically around 1000 respondents. To ensure sufficient sample sizes for small categories of workers, we combine two most recent waves of the EWCS, with data collected in 2010 and 2015. The analysis includes 28 European countries (EU-27 and the UK), with 72,237 respondents in total.

Multiple jobholders (MJHs) are identified in the EWCS with the question ‘Besides your main paid job, do you have any other paid job(s)?’ The distinction is made between workers who have a single job (single jobholders – SJHs) and MJHs who work in additional jobs on a regular (at least 30 minutes per week) or occasional basis. In total, 5604 respondents in our sample indicated that they had an additional job or jobs, with 41 per cent of them juggling more than one paid job on a regular basis.

Job quality is a multi-dimensional concept and, accordingly, is measured on several dimensions which characterise features of jobs linked to positive outcomes for workers (Burchell et al., 2014; Felstead et al., 2019), commonly used in previous studies of job quality (Eurofound, 2016; Green and Mostafa, 2012; Piasna, 2017). Table 1 provides a detailed description of six job quality dimensions and their sub-dimensions used in the analysis: (1) income; (2) job security and prospects; (3) work pressure; (4) skills and autonomy; (5) unsocial hours; and (6) working time control and flexibility. They have been selected for the analysis based on their relevance for multiple jobholding, as concluded from the literature review.

Table 1.

Job quality dimensions.

Dimensions	Short label	Detailed description
1. Income	Income	Net monthly earnings from the main paid job in euros (trimmed and adjusted for PPP)
2. Job security and prospects	Prospects	2a. Job security: might lose job in the next 6 months. 2b. Prospects: job offers good prospects for career advancement.
3. Work pressure	Pressure	Working to tight deadlines; factors constraining pace of work: colleagues, customer demands, production/performance targets, machine speed, boss; not enough time to get the job done; experiencing stress at work; working during free time to meet work demands.
4. Skills and autonomy	Skills	4a. Cognitive dimension: solving unforeseen problems; complex tasks; learning new things; able to apply own ideas at work; working with computers. 4b. Decision latitude: able to choose or change order of tasks, methods of work, speed or rate of work; having a say in the choice of work colleagues or partners. 4c. Agency: consulted before objectives are set for own work; can influence decisions that are important for own work.
5. Unsocial hours	Unsocial	Night-time working; working long days (more than 10 hours); working at weekends.
6. Working time control and flexibility	Flexibility	Changes in work schedules: initiative and notice given; short-term flexibility (taking an hour or two off for personal reasons).

No arbitrary weighting was introduced in the calculation of job quality indices (see Leschke and Watt, 2014), with all survey items contributing equally to the final score within each dimension. All dimensions and sub-dimensions of job quality used in the analysis have been scaled to range from 0 to 100, with higher values always indicating better job quality. This necessitated inversion of some of the scores (e.g. for work pressure and unsocial hours). Income is measured in euros and adjusted by the purchasing power parity (PPP) index obtained from Eurostat. Moreover, a small number of outliers (top and bottom 0.25 per cent of the income distribution) were removed.

All questions about job quality, as well as about sector or occupation, refer to the main paid job of a respondent. Therefore, in the analysis, we do not capture the (average) quality of all paid activities, but only the characteristics of the primary job. Based on previous findings (e.g. Holman, 2013; Smith et al., 2013), we expect to find some trade-offs between certain elements of job quality and therefore analyse each dimension separately, as opposed to combining them into one composite index of job quality.

In the first stage of the analysis, we investigate differences in the job quality features of main paid jobs between SJHs and MJHs. The objective is to determine what job quality features of the primary job are more often found among MJHs and can thus be assumed to encourage multiple jobholding. To this end, random intercept multi-level (ML) regression models with workers grouped within countries are estimated. The choice of the ML models is motivated by the expected similarity in working conditions among workers from the same country (see, for example, Gallie, 2007), violating classical OLS regression assumptions about the independent error terms.

Job quality dimensions are dependent variables in the regression analysis; thus, in each stage, six models are computed for each dimension of job quality separately. In the base models, only a binary distinction of SJHs and MJHs is introduced as an explanatory variable, thus revealing average gaps in job quality between workers with one or more paid activities. In the next stage, we test whether such job quality differences are due to the over-representation of MJHs in certain socio-demographic groups of workers. To this end, control variables accounting for a number of job and worker characteristics are introduced, including employment status (self-employed, employee with an open-ended contract, employee with another type of contract), sector (13 groups based on NACE classification), occupation (nine groups based on ISCO classification), education (up to lower secondary, upper secondary, any tertiary level), age (in years), gender and year of survey.

The second stage of the analysis aims to investigate differences between MJHs, i.e., whether they form a homogeneous group, are polarised between ‘top’ and ‘bottom’ quality jobs, or whether they can be classified in a number of clusters with distinct job quality profiles. To that end, a hierarchical cluster analysis is carried out using nine sub-dimensions of job quality¹ and according to the Ward method (minimising within-cluster variation). The sample consists of MJHs who have non-missing information on all sub-dimensions of job quality (n = 4 284, unweighted). The optimal number of clusters (that best fits the data) was chosen using the Duda-Hart index. Identified clusters of MJHs are then characterised in terms of job quality, with a focus on economic and job insecurity, and socio-demographic and occupational characteristics of workers. In the final stage, cross-national differences in job quality clusters of MJHs are explored. To ensure sufficient cell sizes, countries are grouped according to employment and welfare regime-type typologies. All descriptive analysis is carried out using post-stratification and cross-country weights.

Results

Differences in job quality between SJHs and MJHs

A comparison of the quality of main paid jobs between workers who have only one job and those who engage in more than one paid activity reveals significant differences in all dimensions of job quality. However, the results shown in Table 2 do not indicate a consistent job quality penalty, but rather trade-offs between different job features. On average, MJHs report lower net monthly earnings and lower job security and prospects, experience higher work pressure and work more unsocial hours (models 1, 2, 3, and 5, Table 2). In this respect, the evidence is consistent with previous studies pointing to economic and job insecurity as factors encouraging additional paid employment.

Table 2.

Job quality differences between MJHs and SJHs (ML regression results).

	Income	Prospects	Pressure	Skills	Unsocial	Flexibility
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
MJH	−35**	−1.199**	−1.644***	2.618***	−2.257***	1.549***
Constant	1266***	57.358***	65.140***	61.047***	81.867***	52.004***
Deviance (-2*LL)	800 425.4	603 871.0	573 767.3	612 935.2	596 018.0	630 341.4
	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
MJH	−71***	−2.457***	−1.439***	0.619*	−2.151***	1.051**
Control variables	yes	yes	yes	yes	yes	yes
Constant	1456***	69.543***	54.805***	63.645***	81.847***	42.024***
Deviance (-2*LL)	783 286.5	595 416.0	568 258.5	588 450.4	579 382.7	607 061.9
No. of observations	49,951	64,966	66,887	66,887	66,687	66,848

* p < .05, ** p < .01, *** p < .001

Notes: Number of groups: 28. Higher values on all dependent variables indicate better job quality. Control variables: employment status, sector, occupation, education, age, gender and year of survey.

On the other hand, we find that the primary jobs of MJHs are of better quality in the other analysed dimensions. On average, MJHs work in jobs characterised by a wider scope for exercising skills and discretion, and a higher degree of control over and flexibility in working hours (models 4 and 6, Table 2). All these differences in job quality between MJHs and other workers are statistically significant (p values ≤ .004).

These job quality gaps between MJHs and SJHs cannot be explained by their different position in the labour market or personal characteristics. To test the role of compositional factors, we include a range of control variables, but results remain significant even after accounting for sectoral, occupational and socio-demographic differences between SJHs and MJHs (models 7–12, Table 2). Once worker and job characteristics are accounted for, the job quality penalty among MJHs in terms of income and job security becomes even more pronounced, while the premium in terms of skills and discretion, and working time control and flexibility is somewhat reduced.

Job quality profiles among MJHs

In the first stage of the analysis, we investigated what job quality features of the primary job are, on average, more likely to be found among MJHs. To that end, we examined whether MJHs face, overall and accounting for other work and worker differences, a job quality premium or penalty when compared to SJHs. In the next stage, the objective is to explore whether MJHs form a uniform group characterised by high or low job quality for each analysed dimension, or whether they can instead be classified in a number of different job quality clusters.

The cluster analysis identified a solution with six clusters as the optimal fit to the analysed data (a solution with a relatively high Duda-Hart index and a relatively low associated pseudo-T-squared value). Within these clusters, workers share similar patterns of job quality, and such a grouping maximises dissimilarity between clusters. Table 3 summarises the prevalence of the six clusters. The clusters are then interpreted in terms of their prominent job quality characteristics (Table 4), the professional profiles of workers (Table 5 and Table 6) and their socio-demographic characteristics (Table 7).

Table 3.

Job quality clusters among MJHs.

Cluster No.	Share among MJHs	Frequency in the sample	Label
1	19%	735	Autonomous
2	11%	431	Under pressure
3	23%	912	Balanced
4	16%	644	Insecure
5	16%	642	Low discretion
6	14%	548	Vulnerable

Table 4.

Differences in job quality between MJH job quality clusters.

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable
Income	1669	1523	1316	1093	1110	1024
Prospects	82.8	81.2	54.5	37.4	54.5	31.5
Pressure	66.0	59.0	66.1	60.5	59.8	57.4
Skills	82.8	61.2	71.9	70.4	36.7	33.4
Unsocial	82.4	77.4	82.0	80.4	79.2	79.1
Flexibility	70.8	40.1	61.6	62.7	33.0	26.4

Note: Job quality measured on a scale of 0-100, income in euro.

Table 5.

Occupational structure of job quality clusters (column %).

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable
Managers	13	1	4	3	1	1
Professionals	37	32	24	29	16	10
Technicians	18	21	17	15	12	6
Clerks	8	10	8	10	9	9
Service and sales	11	17	15	19	27	21
Skilled agricultural	2	1	4	3	1	2
Craft workers	7	8	11	9	10	17
Operators	1	5	4	3	11	9
Elementary occupations	4	5	12	8	12	25

Table 6.

Sectoral structure of job quality clusters (column %).

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable
Agriculture	4	1	6	4	1	3
Manufacturing	10	8	10	13	15	20
Construction	4	5	6	5	6	8
Commerce	14	13	16	15	22	26
Transport	2	5	3	3	8	6
Financial	6	3	3	2	2	1
Public admin	7	8	5	7	4	2
Education	18	15	15	14	11	6
Health	12	26	13	10	15	5
Other services	23	15	23	27	16	22

Table 7.

Socio-demographic characteristics by MJH job quality clusters (column %).

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable
Age < 35	33	38	28	42	38	39
Age 35–49	42	44	41	36	37	35
Age 50+	25	18	31	22	26	26
Women	43	52	50	49	47	46
Education: primary, lower secondary	9	9	15	15	17	29
Education: upper secondary	38	50	45	42	55	54
Education: tertiary	53	41	40	43	28	17

Autonomous

Covering one in five (19 per cent) MJHs, this cluster exhibits very good outcomes in all dimensions of job quality. Jobs in this cluster offer high levels of cognitive demands combined with high decision latitude and autonomy, both in organising work and in scheduling working hours. These features are combined with relatively low work pressure and a low incidence of unsocial hours. It mainly includes high-skilled white-collar occupations, thus professionals and managers, with a fair representation of technicians and associate professionals, in sectors such as education, other services and commerce. This cluster has the lowest share of women (43 per cent) and by far the highest share of workers with tertiary education (53 per cent).

Under pressure

The second cluster consists of jobs with a relatively high income, high levels of job security and good prospects for career advancement. However, these positive features are offset by relatively high work pressure, little scope for exercising skills and autonomy, and working hours that are unpredictable for workers and allow for little control over scheduling. This is combined with the highest number of unsocial hours among all clusters. This is the smallest cluster (11 per cent of MJHs), consisting of occupations mainly found in health care, followed by education, other services and commerce. Compared to the ‘autonomous’ cluster, there is a slightly higher proportion of low-skilled white-collar professions, and virtually no managers. It has the highest share of women (52 per cent) and a relatively young workforce (only 18 per cent of workers aged 50 or over).

Balanced

This cluster scores average in terms of earnings, skills and autonomy, as well as working time control and flexibility. It has moderate levels of job security and prospects. Nevertheless, jobs in this cluster are worker-friendly with the lowest levels of work pressure and a very low incidence of unsocial hours. This is the largest cluster in our sample, accounting for 23 per cent of MJHs. Compared to the first two clusters, it has a larger share of blue-collar workers, especially in elementary occupations, but is similarly concentrated in the service sector. Workers in this cluster are older, with the highest proportion of the 50+ age bracket (31 per cent), and have lower educational attainment compared to the previous two clusters.

Insecure

The fourth cluster combines a mixture of job insecurity, poor prospects and low incomes, with a fair amount of scope for exercising skills and autonomy and a fair level of working time control and flexibility. It accounts for 16 per cent of MJHs. The occupational profile is similar to that of the ‘balanced’ cluster, with slightly more professionals and service and sales workers. Its occupations are mostly found in other services and commerce, followed by education and manufacturing. This cluster has the highest proportion of young workers below the age of 35 (42 per cent), and has a relatively polarised educational structure.

Low discretion

The fifth cluster has similarly low levels of earnings as the ‘insecure’ cluster but presents a reverse picture in terms of autonomy and security. It is characterised by fair levels of job security and prospects, yet a very low level of skills and autonomy, and little control over and flexibility in working hours. It represents 16 per cent of MJHs in our sample. It is dominated by service and sales workers, with a substantial share of low-skilled manual occupations and an over-represented proportion of occupations in commerce and manufacturing. The majority of workers (55 per cent) have secondary education.

Vulnerable

The last cluster groups jobs with the poorest job quality, on average, in all dimensions except unsocial hours, where its score is the second-worst. This cluster accounts for 14 per cent of MJHs. It is over-represented in elementary occupations, service and sales workers, and craft workers who work in commerce, manufacturing and other services. This group has a relatively low share of women (46 per cent) and the highest proportion of workers with primary education (29 per cent).

In sum, the analysis revealed a considerable heterogeneity among MJHs in terms of the quality of their primary jobs. The six uncovered clusters cannot be easily positioned on a continuum from good to bad quality, but they rather display various trade-offs between the analysed job quality dimensions.

Economic and job insecurity

As suggested by earlier studies, one of the factors weighing on the decision to look for additional paid work is that the main paid job fails to meet all of an individual’s needs from work. Among these unmet needs, time-related underemployment, closely related to insufficient income and an overall assessment of one’s employment situation as insecure, is likely to play a major role.

This is broadly confirmed by the analysed data, albeit with a considerable divergence between job quality clusters. As shown in Table 8, MJHs in all clusters work, on average, shorter hours in their main paid job compared to SJHs. However, the ‘autonomous’ and ‘under pressure’ clusters are characterised by the longest weekly hours (around 35 h/week) and the lowest share of part-time work (35 per cent and 34 per cent, respectively, work less than 35 h/week), although still well above the 25 per cent found among SJHs. Workers in these two clusters are also the only MJHs to prefer, on average, to work fewer hours in their primary job. Their motivation to take up additional employment is thus not directly or mainly explained by hours-related underemployment or insufficient income. The economic situation of the ‘autonomous’ cluster is the most secure, with only 3 per cent of MJHs reporting that they face difficulties in making ends meet and 50 per cent (compared to 66 per cent among SJHs) tending to have low employment security, measured as a perceived difficulty in finding a job with a similar salary should they lose their current employment. This, combined with a high share of self-employed in this cluster (19 per cent) and their occupational profile, might suggest that additional paid activities are undertaken primarily to increase the variety of work, enrich a CV or expand a client base.

Table 8.

Economic and job security among MJH job quality clusters and SJHs.

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable	SJHs
Working time
Weekly working hours in main job	34.8	35.1	32.4	31.6	31.2	34.0	37.1
Wants to work more (+) or fewer (−) hours	−0.1	−1.2	+0.3	+1.3	+0.8	+1.3	−1.7
Works 1-34 h/week (%)	35	34	45	46	45	40	25
Works 41 or more h/week (%)	21	22	19	17	13	15	24
Employment type
Self-employed (%)	19	4	14	18	6	2	15
Employee with open-ended contract (%)	67	78	69	42	71	48	68
Employee with other contract (%)	14	18	17	40	23	50	17
Perceived insecurity
(Great) difficulty to make ends meet (%)	3	10	10	15	14	24	11
Low employment security (%)	50	52	57	55	62	72	66

In contrast, the ‘under pressure’ cluster has a very low share of self-employed workers (4 per cent) and combines additional jobs mainly with dependent employment on open-ended contracts (78 per cent). The ‘under pressure’ cluster also reports an above-average level of economic security. In this case, motivation to take up additional employment might be intrinsic, e.g. to increase task variety or develop skills, or it might be linked to professional practice in the health-care sector, combining public and private provision of services.

MJHs in the other four job quality clusters would prefer to work more hours than they currently do in their main job, with the ‘balanced’, ‘insecure’ and ‘low discretion’ clusters having the highest share of part-time work (around 45 per cent of jobs with less than 35 h/week). The ‘balanced’ and ‘low discretion’ clusters have a relatively high share of employees on open-ended contracts (69 per cent and 71 per cent respectively) and below-average levels of self-employment (14 per cent and 6 per cent respectively). In these clusters, additional jobs might thus be a means to compensate for low hours in the main job.

However, in the ‘insecure’ and ‘vulnerable’ clusters, it seems that an additional consideration is the temporary nature of the main job, which might not guarantee continuous income in the long run; it thus follows that additional jobs are taken up as a strategy to ‘smooth out’ unstable earnings. Indeed, the ‘insecure’ and ‘vulnerable’ clusters have the highest share of employees on non-standard contracts (40 per cent and 50 per cent respectively). The situation of the ‘vulnerable’ cluster is the least favourable. Relatively long actual working hours (34 h/week) in the main job coincide with a strong preference for even more working hours. This is indicative of very low hourly wages, resulting in the highest share of workers who find it difficult to make ends meet (24 per cent). At the same time, for as many as 72 per cent of MJHs in the ‘vulnerable’ cluster, it would be difficult to find a job with similar earnings should they lose their current main job. This points to a combination of time-related underemployment and high job and employment insecurity as factors pushing workers in this cluster to juggle several paid activities.

Cross-country differences in the prevalence of job quality profiles

In the final stage of the analysis, we explore whether the distribution of multiple jobholders between the job quality clusters varies across countries. As can be seen in Table 9, job quality clusters show a regional variation. To begin with, the size of the ‘vulnerable’ cluster is particularly large in the Mediterranean countries, confirming the prevalence of poor-quality jobs among MJHs in highly segmented labour markets. The ‘low discretion’ cluster is more often found in the UK and Ireland, and in the group of Continental countries. Among the latter, Germany, often considered an example of a highly dualised capitalism model, stands out with a particularly high share of ‘vulnerable’ and ‘low discretion’ clusters. Our findings suggest that, in these countries, moonlighting is more frequently a behaviour driven by necessity, a strategy to tackle financial constraints and to make ends meet. The ‘insecure’ cluster is relatively large in the Mediterranean and CEE countries, thus regions featuring a high incidence of workers in non-standard forms of employment, including the bogus self-employed.

Table 9.

Distribution of job quality clusters across countries (row %).

	Autonomous	Under pressure	Balanced	Insecure	Low discretion	Vulnerable
Nordic	28	15	24	16	13	5
UK and Ireland	24	12	27	15	16	6
CEE	19	12	21	22	13	14
Continental	17	11	24	13	20	15
Mediterranean	13	5	19	23	13	26
Total	19	11	23	16	16	14

Notes: Nordic: DK, FI, SE; Continental: AT, BE, DE, FR, LU, NL; Central and Eastern European: BG, CZ, EE, HR, HU, LT, LV, PL, RO, SI, SK; Mediterranean: CY, EL, ES, IT, MT, PT.

The proportion of multiple jobholders with a relatively good-quality primary job is by far the highest in the Nordic countries, which also record the lowest incidence of the ‘vulnerable’ cluster. Such cross-country differences closely reflect the variations in the national levels of job quality and the distribution of job quality profiles among all workers (see, for example, Gallie, 2007). The UK and Ireland are also among those countries with an above-average share of the ‘autonomous’ cluster. The ‘under pressure’ cluster, which is most commonly found in the Nordic countries, is practically non-existent in the Mediterranean region. Finally, there is some geographical clustering in the prevalence of the ‘balanced’ cluster, which is generally less common in the southern European and CEE countries but more prevalent in northern Europe.

Discussion and conclusions

This article investigated the relationship between the job quality features of the primary job and the propensity to engage in multiple paid activities, with the objective of identifying potential push and pull factors that might lead to multiple jobholding. Using data on workers from 28 European countries (EWCS 2010-2015) and very detailed information on their work and employment conditions, we examined what job quality features of the primary job are associated with having multiple jobs. Subsequently, we explored whether, within the group of multiple jobholders (MJHs), there are any systematic differences in job quality, i.e., whether they form distinct job quality profiles; and, finally, we analysed cross-country variations in their distribution.

The results revealed that, on aggregate, MJHs work in primary jobs that are of poorer quality in most of the analysed dimensions compared to workers who have only one job. This suggests that multiple jobholding is largely a compensatory strategy for job quality deficits in the primary employment. In particular, MJHs have significantly lower net monthly earnings, higher job insecurity and poorer career prospects – differences which are not driven by compositional factors. Thus, the results confirm earlier findings (e.g. Bell et al., 1997; Dickey et al., 2011; Panos et al., 2014; Wu et al., 2009; Zangelidis, 2014) that economic and job insecurity in the primary job are key factors in encouraging additional paid employment. As more jobs fail to provide sufficient and stable income – due to labour market liberalisation, a declining wage share, increases in casual work, or a catastrophic effect on working hours and unemployment of the COVID-19 pandemic – an increasing number of workers are likely to engage in multiple jobholding as a survival strategy.

Moreover, MJHs report, on average, higher work pressure and a greater frequency of working unsocial hours in their primary jobs. While this might indicate that additional employment is undertaken with the aim of moving from a high-strain primary job to another one, it seems more likely that higher work pressure and unsocial hours are a consequence of juggling several paid activities. Finally, we find that MJHs report better job quality outcomes compared to SJHs in terms of wider scope for exercising and developing skills, as well as more control over and flexibility in working time. These job quality premiums, however, are largely accounted for by compositional differences, with MJHs found more often among professionals or the self-employed.

The results also indicate that experiences of work diverge among MJHs. They can be classified into six distinct job quality clusters based on the quality of their primary job (‘autonomous’, ‘under pressure’, ‘balanced’, ‘insecure’, ‘low discretion’ and ‘vulnerable’). This heterogeneity among MJHs leads us to the conclusion that motivations and factors encouraging multiple jobholding are also heterogeneous (Dickey et al., 2011; Panos et al., 2014). For instance, an additional paid activity is more likely a survival strategy, stemming from a necessity to compensate for low earnings and high job insecurity among more vulnerable workers in the ‘vulnerable’, ‘insecure’ or ‘low discretion’ clusters. On the other hand, MJHs in managerial and professional primary jobs of good quality can rely on the financial security and stability provided by their main employment and pursue additional paid jobs to enrich their skills and work experiences, build social capital, or explore alternative career avenues and entrepreneurial activities. Nevertheless, the latter group constituted a minority of the analysed MJHs, with ‘autonomous’ jobs held by only one in five MJHs in our sample and a relatively high level of job security (‘autonomous’ and ‘under pressure’ clusters) reported by one in three MJHs. A considerable share of MJHs, over-represented among the self-employed and non-standard workers, held jobs that, despite having some disadvantageous job quality features, were not precarious in all respects.

Finally, the evidence shows a considerable cross-country variation in job quality among MJHs, with worse outcomes prevailing in more segmented labour markets, particularly in countries with more widespread protection gaps and a higher proportion of non-standard employment. The findings thus suggest that the institutional context influences the structure and dynamics of moonlighting, as well as its potential individual and social impact, a topic that merits investigation in future research.

The findings of this study should be considered in the light of its limitations. First, correlation is not causation, and thus the association between certain job quality features of primary jobs and multiple jobholding can be interpreted in many ways. It can, indeed, show that certain job quality features encourage multiple jobholding, but also that they simply coexist or come as a package in some professions, or even that the causation runs in the opposite direction. For instance, juggling more than one job might have negative consequences for the primary activity in terms of putting constraints on time availability, increasing stress and pressure, or even limiting career progression due to lower availability for overtime at short notice, or simply causing depleted energy levels as a result of excessive work demands. Our interpretation of the results is thus largely based on previous research and theoretical considerations. A second limitation is the lack of detailed information about the second job, such as whether it has similar skills requirements, tenure or earnings. This would certainly provide a more nuanced picture, yet the objectives of this study pertained to the relationship between the quality of a primary job and the propensity to engage in multiple jobholding. Finally, the small numbers of MJHs in our sample prompted a decision to perform a joint analysis of two waves of the EWCS, admittedly at different points in the economic cycle (2010 and 2015), which might have had an impact on the propensity for or feasibility of multiple jobholding. While the use of a single survey wave would be preferable, it nonetheless would not guarantee the same macroeconomic conditions in all of the 28 analysed countries. Controlling for country and year fixed effects in the regression analysis is thus deemed a satisfactory solution, one which was applied in this study.

From a policy point of view, the findings suggest that the heterogeneity of MJHs and specificity of national (or regional) contexts require divergent policy responses. Nevertheless, despite the uncovered heterogeneity, the majority of MJHs experience economic and job insecurity in their primary jobs. Therefore, public policies targeting MJHs ought to acknowledge their above-average vulnerability by offering special measures to provide social protection and address job insecurity or unpredictable income fluctuations. Where multiple jobholding is an individual strategy of compensating for economic and job insecurity, which are largely a result of labour market liberalisation over the past few decades, then a reversal of such policies is necessary. This is particularly important in countries with more widespread protection gaps and where policies based on individual responsibility for one’s employment outcomes prevail (Grimshaw et al., 2016; Keune and Serrano, 2014). Furthermore, issues related to high work pressure and a high number of unsocial hours among MJHs deserve particular attention in policy terms, since individuals forced to undertake multiple jobs out of economic and job insecurity may be more at risk of compromising their health and safety. It is also more difficult to enforce working time limits or manage work stress when a worker has more than one employer. Finally, as this study demonstrates, job growth concentrated in low-wage work and insecure jobs might not translate fully into growth in the employment rate, given that workers tend to combine more than one such job in order to make a living.

Footnotes

Funding

This work was supported by a travel grant from the Hans Böckler Foundation as part of the research network Hybridisation of Work in Europe.

Notes

References

Antón

J-I

Fernández-Macías

Muñoz de Bustillo

(2012) Identifying bad jobs across Europe. In: Warhust

Carré

Findlay

Tilly

(eds) Are Bad Jobs Inevitable? Trends, Determinants and Responses to Job Quality in the Twenty-First Century. London: Palgrave Macmillan, pp. 25–44.

Atherton

Faria

Wheatley

(2016) The decision to moonlight: Does second job holding by the self-employed and employed differ? Industrial Relations Journal 47(3): 279–299.

Bamberry

Campbell

(2012) Multiple job holders in Australia: Motives and personal impact. Australian Bulletin of Labour 38(4): 293–314.

Bell

Hart

Wright

(1997) Multiple Job Holding as a ‘Hedge’ against Unemployment. CEPR Discussion Paper No. 1626. London: CEPR.

Böheim

Taylor

(2004) And in the Evening she’s a Singer in the Band: Second Jobs, Plight or Pleasure? IZA Discussion Paper No. 1081. Bonn: IZA.

Bouwhuis

De Wind

De Kruif

, et al. (2018) Experiences with multiple job holding: A qualitative study among Dutch older workers. BMC Public Health 18(1054). Available at: https://doi.org/10.1186/s12889-018-5841-7 (accessed 1 February 2020).

Broughton

Eichhorst

Magda

, et al. (2016) Precarious Employment in Europe: Patterns, Trends and Policy Strategies. Brussels: European Parliament.

Burchell

Sehnbruch

Piasna

Agloni

(2014) The quality of employment and decent work: definitions, methodologies, and ongoing debates. Cambridge Journal of Economics 38(2): 459–477.

Campion

Caza

Moss

(2020) Multiple jobholding: An integrative systematic review and future research agenda. Journal of Management 46(1): 165–191.

10.

Conen

(2020) Multiple Jobholding in Europe: Structure and Dynamics. Amsterdam: AIAS-HSI, University of Amsterdam.

11.

Dekker

Van der Veen

(2017) Modern working life: A blurring of the boundaries between secondary and primary labour markets? Economic and Industrial Democracy 38(2): 256–270.

12.

Dickey

Watson

Zangelidis

(2011) Is it all about money? An examination of the motives behind moonlighting. Applied Economics 43(26): 3767–3774.

13.

Doellgast

Lillie

Pulignano

(eds) (2018) Reconstructing Solidarity: Labour Unions, Precarious Work, and the Politics of Institutional Change in Europe. Oxford: Oxford University Press.

14.

Dörre

(2019) Precariousness in the Eurozone: Causes, effects and developments. In: Schmalz

Sommer

(eds) Confronting Crisis and Precariousness: Organized Labour and Social Unrest in the European Union. London; New York: Rowman & Littlefield Publishers, pp. 15–32.

15.

Eichhorst

Marx

(eds) (2015) Non-Standard Employment in Post-Industrial Labour Markets. Cheltenham: Edward Elgar.

16.

Emmenegger

Häusermann

Palier

Seeleib-Kaiser

(eds) (2012) The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies. Oxford: Oxford University Press.

17.

Eurofound (2016) Sixth European Working Conditions Survey – Overview report. Luxembourg: Publications Office of the European Union.

18.

Eurofound (2017) Occupational Change and Wage Inequality: European Jobs Monitor 2017. Luxembourg: Publications Office of the European Union.

19.

Felstead

Gallie

Green

Henske

(2019) Conceiving, designing and trailing a short-form measure of job quality: A proof-of-concept study. Industrial Relations Journal 50(1): 2–19.

20.

Folta

Delmar

Wennberg

(2010) Hybrid entrepreneurship. Management Science 56(2): 253–269.

21.

Gallie

(ed.) (2007) Employment Regimes and the Quality of Work. Oxford: Oxford University Press.

22.

Green

Livanos

(2015) Involuntary non-standard employment and the economic crisis: Regional insights from the UK. Regional Studies 49(7): 1223–1235.

23.

Green

(2006) Demanding Work: The Paradox of Job Quality in the Affluent Economy. Princeton: Princeton University Press.

24.

Green

Mostafa

(2012) Trends in Job Quality in Europe. Luxembourg: Publications Office of the European Union.

25.

Green

Mostafa

Parent-Thirion

, et al. (2013) Is job quality becoming more unequal? ILR Review 66(4): 753–784.

26.

Grimshaw

Johnson

Rubery

Keizer

(2016) Reducing Precarious Work: Protective Gaps and the Role of Social Dialogue in Europe. Manchester: European Work and Employment Research Centre.

27.

Guillén

Dahl

S-Å

(eds) (2009) Quality of Work in the European Union: Concept, Data and Debates from a Transnational Perspective. Brussels: PIE Peter Lang.

28.

Henley

(2007) Entrepreneurial aspiration and transition into self-employment: evidence from British longitudinal data. Entrepreneurship & Regional Development 19(3): 253–280.

29.

Hirsch

Husain

Winters

(2016) Multiple job holding, local labor markets, and the business cycle. IZA Journal of Labor Economics 5(4): 1–29.

30.

Holman

(2013) Job types and job quality in Europe. Human Relations 66(4): 475–502.

31.

Isidorsson

Kubisa

(eds) (2018) Job Quality in an Era of Flexibility: Experiences in a European Context. Oxon; New York: Routledge.

32.

Kalleberg

(2009) Precarious work, insecure workers: Employment relations in transition. American Sociological Review 74(1): 1–22.

33.

Kalleberg

Reskin

Hudson

(2000) Bad jobs in America: Standard and nonstandard employment relations and job quality in the United States. American Sociological Review 65(2): 256–278.

34.

Karasek

Theorell

(1990) Healthy Work: Stress, Productivity and the Reconstruction of Working Life. New York: Basic Books.

35.

Keune

Serrano

(eds) (2014) Deconstructing Flexicurity and Developing Alternative Approaches: Towards New Concepts and Approaches for Employment and Social Policy. London: Routledge.

36.

Kimmel

Conway

(2001) Who moonlights and why? Evidence from the SIPP. Industrial Relations 40(1): 89–120.

37.

Kretsos

Livanos

(2016) The extent and determinants of precarious employment in Europe. International Journal of Manpower 37(1): 25–43.

38.

Leschke

Watt

(2014) Challenges in constructing a multi-dimensional European job quality index. Social Indicators Research 118(1): 1–31.

39.

Molina

Rhodes

(2007) The political economy of adjustment in mixed market economies: A study of Spain and Italy. In: Hancke

Rhodes

Thatcher

(eds) Beyond Varieties of Capitalism: Conflict, Contradictions, and Complementarities in the European Economy. Oxford: Oxford University Press, pp. 223–252.

40.

Myant

Theodoropoulou

Piasna

(eds) (2016) Unemployment, Internal Devaluation and Labour Market Deregulation in Europe. Brussels: ETUI.

41.

Panos

Pouliakas

Zangelidis

(2014) Multiple job holding, skill diversification, and mobility. Industrial Relations 53(2): 223–272.

42.

Paxson

Sicherman

(1996) The dynamics of dual job holding and job mobility. Journal of Labor Economics 14(3): 357–393.

43.

Piasna

(2017) ‘Bad Jobs’ Recovery? European Job Quality Index 2005-2015. Working Paper 2017.06. Brussels: ETUI.

44.

Piasna

Drahokoupil

(2017) Gender inequalities in the new world of work. Transfer: European Review of Labour and Research 23(3): 313–332.

45.

Piasna

Myant

(eds) (2017) Myths of Employment Deregulation: How it Neither Creates Jobs nor Reduces Labour Market Segmentation. Brussels: ETUI.

46.

Pouliakas

(2017) Multiple job-holding: Career pathway or dire straits? IZA World of Labor: 356. Available at: https://wol.iza.org/uploads/articles/356/pdfs/multiple-job-holding-career-pathway-or-dire-straits.pdf (accessed 1 February 2020).

47.

Prosser

(2016) Dualization or liberalization? Investigating precarious work in eight European countries. Work, Employment and Society 30(6): 949–965.

48.

Rubery

(2015) Change at work: Feminisation, flexibilisation, fragmentation and financialisation. Employee Relations 37(6): 633–644.

49.

Sliter

Boyd

(2014) Two (or three) is not equal to one: Multiple jobholding as a neglected topic in organizational research. Journal of Organizational Behavior 35(7): 1042–1046.

50.

Smith

Piasna

Burchell

, et al. (2013) Women, Men and Working Conditions in Europe. Luxembourg: Publications Office of the European Union.

51.

Standing

(2011) The Precariat: The New Dangerous Class. London: Bloomsbury.

52.

Thelen

(2014) Varieties of Liberalization and the New Politics of Social Solidarity. Cambridge: Cambridge University Press.

53.

Vaughan-Whitehead

(ed.) (2015) The European Social Model in Crisis: Is Europe Losing its Soul? Cheltenham: Edward Elgar.

54.

Baimbridge

Zhu

(2009) Multiple job holding in the United Kingdom: Evidence from the British Household Panel Survey. Applied Economics 41(21): 2751–2766.

55.

Zangelidis

(2014) Labour market insecurity and second job-holding in Europe. SSRN Papers. Available at SSRN: https://ssrn.com/abstract=2615268 or https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.2615268 (accessed 1 February 2020).