Beyond Income: Why We Want to Keep on Working Even if We Don’t Need the Money

This research focuses upon non-financial work motivation against the background of the debate about the introduction of a basic income. We focus on work commitment; that is the question what binds workers to the employment system except for the wage. We argue that work commitment measures intrinsic work motivation, and that intrinsic work motivation is dependent on the extent to which paid work satisfies the human needs of autonomy, competence and relatedness. Empirical analyses on ESS 2010 for 26 countries show that 55% of the European employees would go on working if means would allow not to and about 25% not. Workers who answer affirmatively work in jobs with high levels of autonomy, good development opportunities, and have co-worker support. Greater autonomy, not only in the job, but also in decisions about work times and the number of work hours, is associated with greater work commitment. Workers in temporary jobs and workers in financial problems have low work commitment. We conclude by arguing that the introduction of a basic income will increase work commitment, because it will relieve workers’ strains and stresses, and will be an incentive for employers to improve the quality of work.


Introduction
During the last decades employment relationships in Western countries have become more flexible and more precarious (Kalleberg 2011;Standing 2011). The expectation is that due to automation the coming years the number of paid hours of work per worker will decrease and that the share of flexible employment relationships will further increase. Against this backdrop it is argued that a basic income should be provided as the alternative for the present system of employment relationships (e.g. Standing 2017 ;Beck 1999;Bregman 2017;Graeber 2018). Proponents of the basic income argue that the decoupling of sustenance and work provides workers of the necessary income security and better opportunities to organize their life.
'Money for free', however, is contested. In the present labor market the wage is the main allocation and motivation device, and it is feared that the decoupling of work and income will withhold workers from working. During the last decades the dominant view in human resource management on work motivation has been that to ensure a strong work motivation, performance and pay should be directly coupled, since it was shown that performance pay can have strong positive effects on performance (Lazear 2018). It has become clear, however, that in the longer run the direct coupling leads to problems of overconcentration on measured performance indicators at the expense of other valuable, but less well measurable work activities. There is therefore good reason to reconsider the claim that work motivation is not only financial (Cassar and Meier 2018). The argument that workers have goals and motivations besides earning a good income was and is generally accepted, but has not got much attention lately due to the prominence of the focus on pay in labour market and hr-policies.
Against the backdrop of this debate about work incentives, we seek to answer the intriguing question what binds workers to the paid work employment system, besides the wage. To answer this question we focus on 'non-financial employment commitment', in the following abbreviated to work commitment. Work commitment is a near approximation of the non-monetary meaning of paid work, since it measures to what extent the worker would enjoy having a paid job if he does not need the money. We will argue that the work commitment question measures intrinsic motivation, and that intrinsic motivation is associated with well-being. The same work conditions that contribute to well-being, also contribute to intrinsic motivation, and, in addition, intrinsic motivation is itself a source of well-being. To explain the attachment to the paid employment system, we elaborate on theories of intrinsic motivation and selfdetermination to develop hypotheses about differences in work commitment between workers. We test the hypotheses on the European Social Survey 2010.

What is Work Commitment?
Work commitment is a measurement construct, of which it is not fully theoretically clear what kind of motivation it measures. The concept has been used to measure different kinds of work motivation. In this section, we describe what is known about work commitment. We first describe the history of the concept, the research in which it was used and its answer patterns. To get a deeper understanding of the concept, we compare work commitment to the related concepts of work ethic and intrinsic work orientation.
The work commitment question has been introduced in the empirical literature as a reformulation of the famous lottery question. The lottery question confronts workers with their routine work motivation (Morse and Weiss 1955). It asks whether the worker is willing to do his paid job after having won a fortune in a lottery or as an inheritance. Response categories 'yes ', 'no' or 'yes, under different conditions' force respondents to take a clear stand. The lottery question opens up a frame of mind that reflects the meaning of work, and for that reason has been used by consultants and work psychologists to discuss their client's work motivation. The usual response to the lottery question is that people are willing to continue doing paid work (Warr 1982;Harpaz 1989), with many workers, however, wanting to reduce their number of work hours or to change jobs (Paulsen 2008). Research of actual behavior after the lottery shows that most winners keep on working, but this depends on the size of the prize (Arvey et al. 2004). Not much is known about the determinants of the answer. It has been argued and shown that answers depend on the satisfaction with the job (Harpaz 1988) and are related to work ethic and work orientation (Harpaz 2002). The work commitment question improves on the lottery question for research goals. The unlikely event of winning the lottery is replaced by the more likely condition that sufficient means for sustenance are available. In addition, the respondent is no longer forced to answer black or white, but is given the opportunity to provide a more nuanced answer. The question is reformulated as a statement, and the respondent indicates whether and to what extent he agrees or disagrees.
In the 1980s and 1990s the concept of work commitment was introduced in research on the work motivation of especially the unemployed workers (Warr 1982;Gallie et al. 1994). The suggestion is that work commitment measures the moral commitment to work, that is the extent to which the unemployed worker feels obliged to work, that is his work ethic. Work ethic is the extent to which people consider work for themselves and for others a social obligation. The work ethic therefore is an internalized extrinsic motivation (Deci and Ryan 2000). The worker identifies with the norm and is motivated by it, but the motivation is not fully self-determined nor fully internally regulated. The concept of the work ethic stems from Max Weber, who argued that Protestants in the 17th and 18th centuries had come to extraordinary economic activity, because they believed that hard work was their duty to God. In the wealth accumulated, as the consequence of their hard work and ascetic life, they saw a finger point of being predestined to eternal grace. For many workers the obligation to work is no longer the main work motivation in today's secularized society (Bell 1976;Muirhead 2004), and measurements show that in Western countries support for work ethic statements is declining (Halman et al. 2012;Wielers and Raven 2013). It is argued that the work ethic has developed into a 'wealth ethic': disposal of sufficient means to be economically independent releases citizens from the obligation to work (Kelvin and Jarrett 1985). It is, however, questionable that work commitment indeed measures the work ethic.
Another notion in the literature is that work commitment is a measure of intrinsic work motivation (Hult and Svallfors 2002;Gallie 2007;Kittel et al. 2019). Intrinsic work motivation is that people work because they enjoy doing so. Workers are intrinsically motivated when the work in itself is rewarding and when an extrinsic reason is not the main motivation. Since all paid work has an external reward, the work commitment construct is a close approximation to intrinsic motivation in paid work, and for that reason has been used as a measure of intrinsic motivation. There is evidence in the empirical literature that work commitment indeed measures intrinsic motivation and not work ethic. Kittel et al. (2019) show that the answer pattern to the typical work ethic question 'Work should always come first even if it means less spare time' shows a different spread across European countries than that of the work commitment question. Dunn (2013) argues and shows that lower educated workers have a stronger work ethic, and better educated workers have a higher work commitment. This suggests that better educated workers work for intrinsic reasons, whereas for many lower educated workers work is a moral obligation. The reverse associations indicate that work ethic and work commitment are different concepts.
The argument that work commitment measures intrinsic work motivation needs qualification. In the empirical literature different concepts of intrinsic work motivation can be found. The most usual measurement asks workers what they find important in a job. Workers indicate different aspects, for instance the extent to which the job enables them to use their own initiative, offers them good training opportunities, is secure or offers them a high income. Factor analysis then shows that the intrinsically motivating aspects such as use of own initiative and good training opportunities are highly correlated, just as extrinsic aspects (Halman and Müller 2006;Turunen 2011;Gallie 2007). The first factor is labelled intrinsic work orientation, the second extrinsic work orientation. The literature does not offer a clear picture about how intrinsic work orientation relates to work commitment. Both concepts measure the valuation of work abstracting from income, but there are also differences. Intrinsic work orientation seems to measure how the respondent values work content, whereas work commitment refers to the broader meaning of the value of doing paid work as part of a good life. A goal of our research is to illuminate the relationship between work commitment and intrinsic and extrinsic work orientations.

Determinants of Work Commitment
In this section we elaborate hypotheses about the determinants of work commitment. Our starting point is that work commitment measures intrinsic work motivation, more specifically the belief that doing paid work adds to the quality of life. We use selfdetermination theory (SDT; Deci and Ryan 2000) and the empirical research literature about how work conditions affect well-being to develop hypotheses about how and when doing paid work adds to work commitment. Starting point is the argument from SDT that well-being is the result of the satisfaction of fundamental human needs of autonomy, competence and relatedness, and that the satisfaction of these needs fosters intrinsic motivation. We elaborate how work conditions and their spillover effects affect well-being and work commitment. We also elaborate how worker and country characteristics affect work commitment.
Autonomy Autonomy is the extent to which people feel free to decide about what they do and how they do a task. Autonomy in the job is the discretion of the worker to determine the content of the work and the work pace. Jobs differ considerably in the level of autonomy, and a high level of autonomy has been shown to increase happinesss (Van der Meer and Wielers 2013). A first test is whether a high level of autonomy is related to a high level of work commitment (H1a).
An issue related to autonomy is work pressure (e.g. Karasek 1979), because high pressure diminishes autonomy. The empirical literature does not give a clear picture when and how work pressure affects well-being. A high level of work pressure may indicate that the job is challenging, but may also be a threat to well-being and health. It is not clear from the literature whether workers generally experience too high or too low work pressure. Green (2006) and Gallie (2005) have shown that work pressure has been increasing and suggest that this increase has a negative effect on well-being. Van der Meer and Wielers (2013) show the existence of such a negative effect. Paulsen (2014) and Graeber (2018), however, argue that many workers feel that they can do the work in fewer hours and are unhappy about the lack of challenge in their work. We therefore do not formulate a hypothesis about the level of work pressure itself, but, instead, focus on the regulation opportunities of work pressure.
The literature argues that the extent to which workers are able to regulate their work load has strong effects on the extent to which they experience their work load as a burden (Karasek 1979). This ability to regulate is captured in the discretion to determine the work pace in the autonomy measure. Another regulation mechanism is the number of work hours. Workers may increase their number of work hours to get the work done, but a high number of work hours can affect well-being and health negatively. This leads to the expectation that workers who work many hours and often work overtime are less committed to work than workers who work fewer hours (H1b).
A slightly different argument about work pressure and work hours is that the extent to which workers are free to decide autonomously about their work hours has strong effects on their well-being. Workers may voluntarily work more hours to do a good job (Beckers et al. 2008;Van der Meer and Wielers 2015). These 'engaged workers' should be distinguished from 'workaholics', who are addicted to work and feel an urge to work constantly, since they are not able to relax (Van Beek et al. 2012). Accordingly, it depends on the autonomy of the worker whether a large number of work hours leads to mental burdening, unwanted overtime hours and lower work commitment. We test whether workers who work more hours than wanted show lower work commitment (H1c).
A well-known argument in SDT is that autonomy is dependent on the reward system, because extrinsic motivators tends to crowd out intrinsic motivation (Ryan and Deci 2000b). Many work organizations use financial rewards as incentives to increase performance. Performance pay is a strong incentive to increase output, but also forces the worker into a narrow focus on measured output. Workers play the game of 'making out' (Roy 1952) and feel that their autonomy is restricted. For that reason, we test whether performance pay is related to low work commitment (H1d).
Competence A job may offer good opportunities to show and develop skills, but not all jobs do. Van der Meer and Wielers (2013) show that career and development opportunities in the job have a strong positive effect on happiness. We expect that better opportunities to improve competences are associated with greater work commitment (H2a).
Relatedness Relatedness needs in the workplace affect well-being and work commitment in different ways. A first hypothesis is that the quality of the relationships in the workplace affects well-being. The quality of relationships with co-workers and superiors varies for personal reasons, for the nature of the job, or because of the organization culture, but good relationships add to well-being, and it is therefore fair to expect that better relationships with co-workers affect work commitment positively (H3a).
In addition, we expect an effect of job insecurity, since job security affects feelings of belonging. The prospect of losing the job and entering unemployment implies exclusion. Research shows that job insecurity is a source of mental problems that can have strong negative effects on health and well-being (Sverke et al. 2002). This research does not show a strong relationship between the flexible employment contract and job insecurity. The negative effects of job insecurity are stronger for older workers on a permanent contract than for younger workers on a temporary contract (Klandermans and Van Vuuren 2010;De Witte and Näswall 2003). We therefore expect that the experience of job insecurity is related to lower work commitment, with no or only a small effect of the flexible employment contract (H3b). In addition, we expect that work security, that is the perceived opportunity that after the loss of the present job an equally valued job can be found, has a positive effect on well-being. Research shows that work security partly compensates for the detrimental effects of job insecurity (Green 2011;Knabe and Rätzel 2011). We expect that people who assess that a comparable job is hard to find will show lower work commitment (H3c).
So far, we have argued that paid work, to a greater or lesser extent, can satisfy fundamental needs. Work, however, also has spillover-effects, that may be positive or negative for well-being. We focus on two spillover effects, namely financial problems, as the consequence of a too low income from work, and work-life conflict, due to time and effort spent on the work and not on other life interests.
Financial Stress Financial stress diminishes freedom of action, and therefore negatively affects quality of life. Paid work is the main way to solve or relieve the problem of financial stress. Mullanaithan and Shafir (2013) have shown that financial scarcity narrows the focus to short-term solutions of acute stress. We therefore expect that workers in financial problems more often consider giving up paid work when sufficient financial means would be available (H4a). The narrow focus on financial rewards may not only apply to people in acute financial problems, but also to people who live on low incomes. Living on little money makes the wage the main motivation to work (H4b).

Work-Life Conflict
In addition, paid work has spillover effects on relationships outside the workplace. A high level of work-life conflict implies that important life interests, such as the care for children or own parents, do not get the time and attention needed, and this will affect work motivation negatively (H5a). The alignment of paid work and private life, however, may also increase work motivation, because the combination of the benefits of paid work and of family life may enrich both (Greenhaus and Powell 2006). Research indicates that workers may dispose of instruments that enable them to bring work and private life into the balance (Lyness et al. 2012;Hofäcker and König 2013): discretion in decisions about work times and about number of work hours. We expect that workers who are free to determine their own start and end times show higher work commitment (H5b). We also expect that workers who have discretion about work hours show higher work commitment. The full-time working hours are the standard number in all countries, but during the last decades in several western countries, workers have acquired the right to deviate from this standard. In these countries, the share of part-time workers is high and most of them are part-time workers voluntarily (OECD 2010). We expect that these part-time workers show a high level of work commitment, because part-time work allows them to combine paid work with their other life interests (H5c).
Worker Characteristics Revenues in well-being and costs of effort differ between workers. We expect the balance to be especially positive for workers for whom doing paid work is more than a routine activity, because they experience the participation in paid work as a free choice that brings independence, development opportunities and inclusion. These are workers who, individually or as a social category, have obligations outside the labor market or often are discriminated against. Paid work offers them independence, development and full participation in society. We will refer to this as the 'emancipation effect' of paid work. Such an emancipation effect may apply to women, younger workers, members of ethnic minorities, and, if scarred, formerly unemployed workers (H6a). The other side of the coin is that paid work may burden workers to such an extent that it does not add to, but harms well-being. Workers who are older or less healthy often consider to reduce hours or to stop working, if financial means allow (H6b).

Country Characteristics
The available literature does not show strong country effects. An effect of the generous welfare state was reported (Van der Wel and Halvorsen 2015), but further research did not confirm the effect. Testing for a number of cultural, institutional and economic differences between countries, the only significant effect was a relatively small effect of Gross National Income (Turunen and Nätti 2017). This indicates that differences between countries do not so much stem from differences in national institutions and culture, but are the consequence of country differences in job structure and composition of the labor force. The literature provides several explanations how national institutions and policies affect the job structure and the composition of the labour force (Gallie 2009), but these differences are not reduced to country-level effects in statistical multi-level models. The general picture from the multi-level statistical models is that the experience of work is much more dependent on variables on the individual level than on the country level (Pichler and Wallace 2009). It is therefore probable that differences in work commitment between countries are to a large extent explained by variables on the individual level.

Work Commitment and Well-Being
So far, we have focused on the determinants of work commitment, elaborating on the argument that what in work fosters well-being also fosters work commitment. We now shift focus to the effect of work commitment on well-being. The literature suggests two reasons why a high level of work commitment itself adds to well-being. These reasons stem from SDT and originate in arguments about the associations of well-being with intrinsic and extrinsic motivations respectively.
The first reason is that work commitment indicates that workers feel that they do paid work for intrinsic reasons, that is because they themselves want to, and not for an extrinsic reason (Ryan and Deci 2000a). Doing what you want to do adds to feelings of well-being. A high level of work commitment indicates that the worker feels that paid work adds to his well-being, and that doing paid work is to large extent self-determined. The self-determination is not only restricted to the practice of work, but, because of the central position of work in the life, extends to the organization of life. Such intrinsic motivation is associated with greater efficacy, more self-esteem and greater life satisfaction (Deci and Ryan 2000). We test whether this higher level of work commitment is itself a resource for well-being. We expect work commitment and well-being to be associated due to common fostering conditions, but we furthermore expect that work commitment itself is an additional resource for the generation of well-being (H7a).
The second reason is that a high level of work commitment indicates a relatively weak orientation towards the materialistic rewards of work. Kasser (2003) has argued and shown that people with a strong materialistic orientation reach lower levels of wellbeing than people focused upon intrinsic rewards. Materialistic values are associated with more strains and stresses, because material needs never are fully satisfied, since meeting the goal leads to an upward adaptation of the aspiration level. Kasser argues that materialism stems from uncertainty, but also that a materialistic orientation does not remove the uncertainty, but enhances it. We therefore expect that people with a strong orientation on material rewards show less work commitment. We test whether a strong extrinsic work orientation affects work commitment negatively, and thus has an indirect negative effect upon well-being (H7b).

Data and Measurements
Our argument has led to hypotheses that are tested on the European Social Survey 2010. ESS 2010 contains a special module with questions about work, work conditions and work-life balance. We use 26 countries in our analysis. 1 We restrict analysis to employees between 18 and 65 years who had indicated that paid work was their main activity, and who had no missing values on the variables in the analysis. There are 13,185 respondents in the analysis.
Work Commitment is measured as the adherence to the statement: 'I would enjoy having a paid job even if I did not need the money'. Responses were measured on a 5point-scale, answer categories varying from Agree strongly (5) to Disagree strongly (1).
Well-Being is measured by the question: 'How satisfied are you with your life as a whole', which has a ten points scale.
Job Autonomy is measured as the mean adherence to three items: 'being allowed to decide how daily work is organized', 'being able to influence policy decisions about activities of the work organization', and 'being allowed to choose/change pace of work'. Adherence is measured on 11-point scales, varying from No influence (0) to Complete control (10). Reliability of the items is .86.
Work Pressure is the average of support to two statements: 'My job requires that I work hard' and 'In my job there is never enough time to get all things done'. Work pressure is measured on a 5-point scale, with answers varying from Agree strongly to Disagree strongly. Answers are recoded such that higher values indicate greater work pressure.
Hours Excess indicates that the respondent spends more hours on his work than he wants to. Hours excess is computed as the difference between the number of actually worked hours and the preferred number of work hours per week. There is hours excess, when this difference exceeds 2. Hours excess is coded as a dummy variable. The reference category is hours balance.
Performance Contract is measured as support for the statement: 'In my job the reward depends on the effort', with answer categories varying from Not true at all (1) to Very true (4).
Development Opportunities is measured as adherence to two items: 'There is variety in my work'; and 'My job requires learning new things'. Adherence is measured on 4point scales varying from Not at all true (1) to Very true (4). The variable used is the mean of these items. Support Co-workers is measured as adherence towards the statement that the respondent can get help or support when needed from co-workers. Adherence is measured on a 4-point scale, varying from Not at all true (1) to Very true (4).
Job Uncertainty respondents indicated to what extent their current job is secure. Answers vary on a 4-point scale from Not at all true (1) to Very true (4). Answers are recoded, so that higher values indicate greater uncertainty.
Work Uncertainty respondents indicated how difficult or easy it would be for them to find a comparable or better job if they would have to give up their present job. Answer categories vary on an 11-points scale from Extremely difficult (0) to Extremely easy (10). Answers were recoded so that a higher value indicates greater work uncertainty.
Temporary Contract respondents indicated the duration of their employment contract as unlimited, limited or no contract. The variable temporary contract indicates workers with a contract of limited duration and workers without employment contract. Reference category are workers with a permanent contract.
Financial Problems is measured as the extent to which the respondent can live well on the present family income. Answer categories vary from Living comfortably on present income (1) to Very difficult on present income (4).
Household Income using the respondents' information, ESS estimates the net household income. These household incomes per country are divided over deciles. The variable net household income thus measures the relative income on a 10 points scale, in which a higher value indicates a better income.
Work-Life Conflict is measured as the average of the adherence to three statements: the extent to which the respondent 'keeps worrying about work problems when not working'; 'feels too tired after work to enjoy the things he would like to do at home', and 'finds that his job prevents him from giving the time he wants to partner or family'. Adherence is measured on a 5-point scale, answer categories varying from Never (1) to Always (5). These items usually are part of an extended work -family conflict scale, to which items about the effect of the family on work performance are added. We have not included these items because this would exclude workers without family responsibilities. Reliability of the three items is .598.
Flexible Work Hours is measured as the adherence to the statement that the respondent is free to decide about start and end times of his work. Answer categories vary from Not at all true (1) to Very true (4).
Number of Work Hours is measured as the answer to the question how many hours the respondent usually works per week, overtime hours included. We recoded the answers into dummies that indicated one of three categories: marginal part-time workers: 0-12 h; part-time workers: 13-29 h; full-time workers: 30-44 h; and overtime workers: >45 h. We use full-time workers as the reference category.
Worker Characteristics For age, we distinguish between young, middle-aged and older workers. Young workers are in the age category 15-25 years, older workers in the category 55-64. Respondents assessed their subjective general health, with a score varying from Very good (5) to Very bad' (1). Gender is coded 0 for Man, 1 for Woman. If the respondent lives with a partner, Partner = 1, otherwise 0. If respondent has at least one child at home, Child = 1, otherwise 0. Respondents indicated whether they belong to an ethnic minority (if so = 1, not = 0). Unemployment experience is measured as adherence to the statement: 'Ever been unemployed and seeking work for a period more than 3 months'. If Yes = 1; if No = 0. Educational level is coded according to ISCED, a 7-point scale, varying from Lower than secondary level (1) to Ma-level or higher (7), which we use as an interval scale.
Intrinsic and Extrinsic Orientation As argued, previous research has shown two different underlying dimensions of intrinsic and extrinsic orientation when different job characteristics are included in a factor analysis (Halman and Müller 2006;Turunen 2011;Gallie 2007). ESS includes two items per orientation. For intrinsic orientation respondents were asked about the importance of being able to show own initiatives and the importance of good schooling opportunities. For extrinsic orientation, the respondents were asked to assess the importance of job security and of a high income. Respondents indicated their preferences on a 5-points scale. For each respondent, values on both dimensions were computed as the average of the two items. A higher value indicates a higher intrinsic or extrinsic work orientation.

Research Strategy
We estimate multi-level structural equation models to test the hypotheses. The models test simultaneously for the effects of the independent variables on both work commitment and well-being, and for the effect of work commitment on well-being. Before estimating the structural equation models, we estimated separate multi-level models for the determinants of work commitment and well-being respectively. Results did not differ much from those of the equations in the structural model.
In Table 1 we report four models. The first model is empty in both the equations for work commitment and well-being, and serves as the reference for improvements in model fit. In the second model, we add for both equations all variables except those that indicate motivation. In the third model we add measures for intrinsic and extrinsic motivation to the work commitment equation and work commitment to the well-being equation. In the fourth model we add intrinsic and extrinsic motivation also to the wellbeing equation. The point estimates in the text refer to this model. Models 2-4 are full models, thus variables measuring related quantities are entered into the same model and effects may not be robust. We use correlation analysis to interpret and disentangle effects. The correlation matrix is presented in an Online Appendix. We did some additional analyses to assess the relative size of the worker effects in comparison to the job effects.
We report multi-level models to account for country differences. We did some additional analyses to test for the effect of country level variables, but found no significant effects of country variables on work commitment when all individual level variables were included. A linear structural equation model in which countries were added as a fixed effect and in which the standard errors were clustered by country showed similar effects as the models reported. We do not present the results of these analyses. Results of analyses not reported are available upon request.

Results
We first focus on the results for work commitment and then proceed to well-being.

Distribution of the Work Commitment
Variable Most respondents answer the question about their work commitment affirmatively. The average is 3.34 on a scale varying from 1 to 5. About 55% of the respondents indicate to keep on working, more than 25% do not. Table 1 shows that about 10% of variance in the model is explained at the country level. Subsequent models show significant improvement in model fit. Variance explained, calculated as the change in variance at the individual and country level, is low. At the individual level about 6.4% of variance is explained in Models 3 and 4; at the country level 15.9%.

Model Fit ICC of Model 1 in
Autonomy Autonomy shows a strong positive effect on work commitment (H1a: b = 0.025; se = 0.004). The models show a positive effect of work pressure (b = 0.026; se = 0.012). The effect of overtime work is not significant (H1b: b = 0.042; se = 0.026), but   there is a significant negative effect of excess hours (H1c: b = −0.075; se = 0.021). There is not a significant effect of performance contract (H1d: b = −0.010; se = 0.010).
Relatedness Support of co-workers has a significant positive effect (H3a; b = 0.035; se = 0.012). There are negative effects of job uncertainty (H3b: b = −0.051; se = 0.010) and work uncertainty (H3c: b = −0.012; se = 0.004), and there is not an additional effect of temporary contract (b = 0.023; se = 0.028). These effects correspond to the significance of the associations reported in the correlation table.
Financial Scarcity Financial problems have a negative effect on work commitment (H4a: b = −0.040; se = 0.015). There is not a significant additional effect of the relative household income (H4b: b = 0.001; se = 0.005). Additional correlation analysis shows that lower income households have a greater chance of having financial problems. The effect of household income in the model is affected by its correlation with financial problems. The correlation between household income and financial problems is −.40. Additional analyses show that the AIC and BIC of models containing only job characteristics are lower than the AIC and BIC of models containing only personal characteristics. This shows that work commitment in these models is dependent more on the quality of the job than on worker characteristics.

Intrinsic and Extrinsic Motivation
Model 3 shows additional effects for intrinsic and extrinsic orientation. Intrinsic orientation has a positive association with work commitment (b = 0.147; se = 0.015), and extrinsic orientation shows a negative association (b = −.0139; se = 0.017). The effects are relatively strong in comparison to other independent variables, but not very strong. Addition of these variables does not change the effects of the other independent variables substantially.
Effects on Well-Being The second equation tests the additional effect of work commitment and of intrinsic and extrinsic work orientation on well-being.
The independent variables in the models mostly show the well-known and expected effects on well-being of personal characteristics such as the strong positive effects of having a partner (b = 0.385; se = 0.038) and of good health (b = 0.480; se = 0.021). In addition, they show strong effects of work-related variables. There are strong positive effects of having autonomy (b = 0.039; se = 0.007), development opportunities (b = 0.126; se = 0.021) and support of co-workers (b = 0.136; se = 0.019), and there are negative effects of financial problems (b = −0.556; se = 0.024), job uncertainty (b = −0.073; se = 0.017), work uncertainty (b = −0.030; se = 0.006) and work-life conflict (b = −0.282; se = 0.021). These results support our argument that the work-related variables that add to work commitment also add to well-being. A significant exception is flexible work times, that shows a positive effect on work commitment (b = 0.023; se = 0.010), and a negative effect on well-being (b = −0.057; se = 0.016). The correlation table, however, shows a positive correlation between flexible begin and end times and well-being. Work pressure, hours excess, work-life conflict and part-time work are significant in one equation, but not in the other.
Models 3 and 4 show a strong effect of work commitment on well-being (H7a: b = 0.050; se = 0.014). The effect is hardly reduced when intrinsic and extrinsic orientation are added to the equation. Intrinsic orientation (b = 0.038; se = 0.024) and extrinsic orientation (H7b: b = 0.005; se = 0.027) do not show significant effects. The correlation table shows about equally strong correlations between work commitment, intrinsic orientation and extrinsic orientation with well-being, but the only significant effect in the model is that of work commitment. Work commitment is a better predictor of wellbeing than intrinsic or extrinsic motivation.

Level of Work Commitment
Our results show that workers in European countries have high work commitment. More than 50% of employees in western countries would continue doing paid work even when the money is not needed. The default answer to the work commitment question thus is affirmative. About 25% of workers indicate to drop out. The empirical evidence indicates that low work commitment is related to strains and stresses, such as less good health, financial problems and job and work insecurity.
Meaning of Work Commitment We find that a high level of work commitment is associated with a high level of well-being. Workers who are committed to doing paid work have a higher level of well-being than workers who do not. There are two complementary reasons for this association. The first reason is that generally the same work conditions that foster work commitment also foster well-being. For many workers doing paid work contributes to the satisfaction of their fundamental needs of autonomy, competence and relatedness. The work adds to their well-being and thereby fosters their work motivation. In addition, a high level of work commitment itself adds to wellbeing. The apparent reason is that committed workers, in contrast to less committed workers, feel that doing paid work is their free choice, and not an external obligation or external regulation. The implied free choice makes high work commitment itself a resource of well-being. We did not find confirmation for an alternative explanation that a less materialistic work orientation explains the higher well-being. The positive association of work commitment with well-being explains the high level of work commitment.
These results support our argument that the work commitment question measures to what extent doing paid work is intrinsically motivated, thus that committed workers feel that doing paid work adds to their quality of life. Intrinsic motivation to work is fostered in conditions that satisfy the fundamental human needs of autonomy, competence and relatedness, and the motivation itself is subsequently a resource for wellbeing. It follows that work commitment is to be distinguished from the internalized but externally regulated work motivation that is referred to as the work ethic. Our research adds to the research that shows that work commitment and work ethic are different phenomena (Dunn 2013;Kittel et al. 2019). We find that work commitment is especially dependent on the quality of the work, whereas the literature reports that the work ethic is mainly dependent on worker characteristics that capture internalized norms, such as membership of a religious denomination, age and education (Furnham 1990;Wielers and Raven 2013). An implication is that when research reports a declining work ethic, this does not necessarily mean a decline of work motivation. Workers high on work commitment feel that paid work should and indeed does add to their own well-being, and do not wish to work because it is an external obligation. They therefore do not adhere to the norm of work as an obligation for themselves nor for other workers.
Our results, furthermore, show that the meaning of work commitment differs from that of intrinsic work orientation. We found a significant positive association between work commitment and intrinsic work orientation, but this association is not very strong. In addition, we found that in a regression analysis work commitment has a significant positive effect on well-being, whereas intrinsic work orientation has not. These results indicate that work commitment and intrinsic work orientation measure related, but different phenomena. Work commitment measures the intrinsic motivation to do paid work. This intrinsic motivation is associated with feelings of well-being and free choice. Intrinsic work orientation measures the valuation of job characteristics that foster intrinsic motivation, thus not the intrinsic motivation itself. It measures to what extent workers think that an intrinsically motivating job adds to their well-being in comparison to an extrinsically motivating job, thus tactical considerations in job choice. They measure different coping strategies to deal with work conditions in which needs are thwarted and demand a trade-off (cf. Deci and Ryan 2000).

Determinants of Work Commitment
We have tested a number of variables to explain differences in work commitment. These variables explain only small part of the variance. The full model explains about 6.5% of the variation at the individual level. The low level of explained variance may be the result of incomplete models. It is, however, also possible that respondents differ in base rates, that is that differences in work commitment are related to personality or misery traits. The relatively strong associations with other, more subjective measures, such as well-being, and intrinsic and extrinsic orientation support such an interpretation.
Nevertheless, we find considerable effects of the quality of the present job. Autonomy, development opportunities, social contacts and job security have significant positive effects on work commitment and well-being. An additional effect is that workers in financial problems, often workers with low household incomes, have lower work commitment. We do not find an effect of work-life conflict, but there are positive effects of the freedom to determine begin and end times of the work, the own determination of overtime hours and of part-time work on work commitment. Since part-time work is most frequent in countries where workers have the legal right to choose the own number of work hours, our results indicate that not the extent of worklife conflict itself harms work commitment, but that the autonomy in decisions about work times and work hours is decisive for work commitment.
Not all job characteristics that add to work commitment also add to well-being. Flexible begin and end times has a positive effect on work commitment, and shows a positive correlation with well-being, but in the regression model the effect is negative. Work pressure, hours excess, work-life conflict and part-time work show a significant effect in the one equation and a non-significant effect in the other. These variables have in common that they refer to the acts of balancing work and private life. The results suggest that the way the balance is kept or not kept affects work commitment and wellbeing differently.
Our results raise concern about the effect of the development of the job structure on work commitment. The trend in western countries during the last decades has been that the quality of jobs is deteriorating, that is that jobs have become less autonomous, offer fewer development opportunities, and have become more insecure (Kalleberg 2011;Gallie 2013). Lower job quality affects work commitment negatively, and may lead to orientations on alternative opportunities to satisfy needs outside the workplace.
Worker Characteristics We find some effects of worker characteristics on well-being. Workers in less good health and older workers show lower work commitment, probably because work is more of a burden for them. Better educated workers show greater work commitment. The plausible explanation is that they have better and therefore more intrinsically motivating jobs. We also find a strong positive effect for women, which may be the result of an emancipation effect. We do not find effects of other worker characteristics, such as membership of an ethnic minority, having been unemployed, nor of having partner or a child at home.

Work Commitment and the Basic Income
In the debate on the basic income the effect on the work motivation is a delicate issue. Antagonists of the basic income, often inspired by neoclassical economic theory, expect that the basic income will infirm work motivation and thereby make the basic income unaffordable. Proponents have more differing views. The usual reference is the affirmative answers to the work commitment and lottery questions as empirical support for the belief that a basic income will not or hardly affect work motivation negatively (e.g. Graeber 2018). There are, however, proponents of the basic income who doubt that the present quality of paid work is sufficient to motivate workers intrinsically. Reanalysing the answer patterns to the lottery question, Paulsen (2008) argues that many workers are not satisfied with their job and prefer to seek another job or reduce hours.
Our research results add to the understanding of work commitment as measurement of work motivation and its determinants. We have argued and presented empirical evidence that work commitment indeed measures the intrinsic motivation to do paid work. Our research is one of a score that shows that the willingness to do paid work, also when the money is not needed, is high. We add to this literature the argument that doing paid work satisfies fundamental needs, and is therefore positively associated with well-being. The same conditions that foster well-being also foster work commitment, and, in addition, work commitment itself is a source of well-being. Workers low on commitment often have the problem that doing paid work is too much of a burden for them to bring well-being. For these workers paid work is mainly a source of strains and stresses.
The introduction of a basic income will affect work commitment in direct and in more indirect ways. The direct effect is that a basic income gives workers financial leeway to alleviate strains and stresses. Our research shows that work commitment is low for older workers, workers in bad health and for workers with financial problems. For older workers and workers with bad health, the basic income will open up new opportunities to balance paid work with capabilities. The basic income may help to find a better fitting job or to reduce work hours. A basic income will also relieve financial problems. We find a relatively strong negative effect on work commitment. Financial problems are a burden, that harms health, well-being, and work commitment. Workers with low incomes have an increased risk of financial problems. Mullanaithan and Shafir (2013) have shown that problems of financial scarcity are associated with diminished cognitive capacities. The focus is on short-term solutions to solve acute problems, at the cost of more sustainable solutions for the longer term, so that the financial problems return. A basic income provides low income workers with some income security, and thus contributes to the prevention of financial problems. It also increases these workers' cognitive bandwidth, such as their work commitment.
In addition, we expect indirect effects, via the quality of work. Our results show that it is the quality of the work that mainly determines to what extent needs are satisfied in the work. The high level of work commitment in European countries indicates that for most workers the quality of work is sufficient to be intrinsically motivated. There are, however, jobs that are not intrinsically motivating, and mainly are done for their monetary rewards. After the introduction of a basic income especially workers in low quality jobs will reconsider their participation in paid work. This work offers not only low intrinsic rewards, but, often, also low monetary compensation and insecurity. Workers in these jobs relatively often feel that their work is a burden that eventually affects their well-being and health negatively, and therefore are not committed to paid work. The plausible scenario is that workers on a basic income who do not find an adequate job in paid employment will seek and find better ways to satisfy their needs in unpaid employment outside the labor market. Volunteer work becomes a feasible alternative for paid work, because volunteer work may better satisfy the fundamental needs than low quality jobs (Thoits and Hewitt 2001;Wu and Li 2019). Experiments show that volunteers with social security payment are relatively satisfied with their work and their life despite their low income (Kampen et al. 2013). The plausible scenario therefore is that after the introduction of the basic income the now marginal segments of the labour market of voluntary workers and working benefit recipients will grow into a larger and more vital alternative labor market. In addition, it can be expected that the basic income will permit to shift focus to more meaningful work (Graeber 2018;Cassar and Meier 2018). It is probable that workers with a basic income more often will turn to sectors that produce socially useful services, such as care and community work, because of the intrinsically motivating work. The expectation thus is that the introduction of a basic income will bring an increase of intrinsic work motivation, but the probable consequence is also a decrease of supply of paid labor. Low supply will force employers to improve the quality of work. The basic income is an incentive for employers to improve the quality of work in the worst jobs.
We conclude that a basic income will increase rather than decrease work commitment. A reason is that strains and stresses that now affect work commitment negatively, will diminish. We, furthermore, expect that the introduction of a basic income will improve the quality of work, and that this will increase work commitment. Greater intrinsic work motivation, however, does not imply greater labor supply. Workers may seek refuge into intrinsically satisfying volunteer jobs, may shift focus to more intrinsically rewarding work, or may reduce their number of work hours.

Shortcomings and New Directions
This research suffers from the usual shortcomings of cross-sectional research. We have argued and presented evidence that good quality jobs lead to high work commitment, and we were not able to account for the effect that high work commitment leads to better quality jobs. There are several shortcomings in the data. The focus on the work commitment question enabled us to penetrate in the meaning and determinants of the answers, but in the end, it is only a single question about a for many workers hypothetical situation. Not much is known about the relationship between intrinsic work motivation and ownership, except for the behavior after winning the lottery. Further research is wanting.
As for the independent variables, we worked with measurements available in ESS 2010. Several concepts deserve better measurement, for instance, we measured the satisfaction of the need of competence with development opportunities instead of the more direct measurements whether the worker is able to show his competences and feels valued for this. Several variables that we would have wanted to include were not available, for instance whether the worker is in a supervisory position, and how her relationship with her supervisor is. Also missing in ESS 2010 are questions about the culture of the workplace, for instance to what extent it is a competitive culture.
Nevertheless, starting from the issue of the basic income and the question of work commitment, our research has opened up broader issues of the nature of the motivation for paid work, and the measurement and determinants of the intrinsic motivation for paid work. Especially the issue of the controlled and the intrinsic regulation of work motivation deserves further attention. Our research has shown that the traditional measurement instruments of extrinsic and intrinsic work orientations have their limitations, and are better not used for the measurement of intrinsic work motivation. A developed and calibrated measurement instrument of intrinsic work motivation would add to the improvement of research on the quality of work, and thus on the quality of life.

Conclusion
Against the backdrop of the debate about monetary work incentives, this article has investigated what binds workers to the paid employment system, except for the wage. We found that the extent to which paid work satisfies the needs of autonomy, competence and relatedness, and thus contributes to their well-being also adds to their work commitment. In European countries, most workers feel that the quality of the work is good enough to bind them to the employment system. The workers who would stop working when having enough financial means, often have strains and stresses that make the work for them a burden. We expect that the introduction of a basic income will increase work commitment, thus the willingness of workers to do paid work. Paid work for many workers satisfies fundamental needs, and therewith contributes to their well-being. The basic income will relieve the strains and stresses of workers who now feel that work is an obligation, and it will force employers to improve the quality of work.