Introduction
Over the past two decades many OECD countries, Israel included, have adopted welfare policies aimed at reducing receipt of means tested social security benefits and increasing self-sufficiency. Means tested programs also known as Social Assistance Benefit, Income Support Benefit, Safety-net program and Welfare payment (US), hereafter referred to as welfare. These policy shifts, also called “welfare reforms”, have substantially changed the trade-offs women face when considering labor and welfare-use behavior, making employment – almost any employment – far more attractive than welfare receipt. Related to these changes, employment and earnings have risen substantially while welfare use has declined in many countries (e.g. Achdut and Stier, Reference Achdut and Stier2016; Blank, Reference Blank, Auerbach, Card and Quigley2006; Finn and Gloster, Reference Finn and Gloster2010; Shannon, Reference Shannon2009). However, many welfare recipients and leavers experience multiple spells of unemployment, have low earnings, circulate in and out of the welfare system or become disconnected from both work and welfare (e.g. Achdut and Stier, Reference Achdut and Stier2016; Reference Achdut and Stier2018; Blank and Kovak, Reference Blank, Kovak, Heinrich and Scholz2009; Lightman et al., Reference Lightman, Mitchell and Herd2010; Loprest and Nichols, Reference Loprest and Nichols2011). Despite this limited success and the low employment stability of many welfare recipients only little empirical work has examined their job quality and its consequences for welfare exit.
Based on a national panel survey of 2,800 single mothers who were receiving Income Support Benefit, when the Israeli welfare reform was implemented (2003) and were followed up until 2007, this study focuses on their objective job attributes and the extent to which the quality of their employment affects their ability to exit welfare. The study joins a wide literature on job quality and goes beyond wages and working-hour measures to cover other monetary attributes of compensation, some mandatory (paid annual vacation and paid sick leave), others not mandatory (employment-sponsored pension and training fund), and non-monetary aspects of the job held (use of own skills and occupation, existence of standard employment contract, irregular work schedule). The study examines the contribution of these job quality components to welfare exit, on the assumption that having a good job affects work continuity and lessens the benefit of staying on welfare. To our knowledge few empirical studies have focused on this issue in the post-reform era. Two exceptions are Johnson and Corcoran (Reference Johnson and Corcoran2003), who analyzed the relation of schooling, skill content of work experience, and different types of employment patterns to the welfare recipient’s job quality outcomes; and Foley and Schwartz (Reference Foley and Schwartz2003), who examined the effect of an earnings supplement on welfare recipients’ job quality. Other studies provide more descriptive results on this aspect (Loprest, Reference Loprest2003; Loprest and Zedlewski, Reference Loprest and Zedlewski2006). Our study extends this work by including a variety of job characteristics and examining their role, beyond wages, in determining welfare exit.
Background
Theoretical framework
According to economic theory, women make decisions about whether to participate in paid work based on the costs and benefits of entering the labor market (Becker, Reference Becker1981; Killingsworth and Heckman, Reference Killingsworth, Heckman, Ashenfelter and Layard1986). To maximize their utility, welfare recipients make choices about labor supply and welfare participation subject to a budget constraint that takes into account wage opportunities and welfare benefits. Welfare participation is chosen if the indirect utility of the welfare choice is greater than the indirect utility of being off welfare.
Additionally, human capital theorists (Becker, Reference Becker1975; Mincer and Polachek, Reference Mincer and Polachek1974) argue that women will work and work more, thereby exiting welfare through work when they have the levels of education, job skills and work experience demanded by employers and when the expected income from work exceeds the expected income from welfare. This model predicts that women with low levels of human capital are at higher risk for long-term unemployment and welfare receipt. Moreover, sociodemographic characteristics, such as marital status, age, number of children and their age, will play an important role in determining the labor supply of mothers. Children, particularly young children, may substantially limit the number of working hours and the economic viability of employment when childcare cost is relatively high and must be paid for more than one child. Based on this theoretical framework, the labor decision of single mothers depends on the cost of participation, availability of other sources of income and on their ability to replace welfare income with some other source for supporting their family.
While wage offers are important for both the decision to take a job and the scope of employment, other monetary and non-monetary conditions of employment, which determine the quality of a job, matter as well. Some aspects of the jobs have long-term implications for the economic standing of individuals and their families. These include pension benefits, health insurance, training funds that increase employability or other forms of saving that are sponsored by the employer. Having these attributes is expected to affect welfare exit as they increase the attractiveness of employment compared with welfare receipt.
Other monetary aspects enter the utility function of all employees, but they might be of great importance for single mothers. These include paid annual vacation and sick leave that help mothers to combine work and family and allow flexibility and security that affect work continuity and thereby increase the likelihood of welfare exit.
Non-monetary attributes might also be important for a welfare-labor decision. A good match of workers’ skills and job demands is expected to enhance welfare exit. This is because an adequate skill match, where workers can demonstrate and use their occupation and skills, provides an opportunity for further training on the current job – firm-specific human capital acquisition – and increase work experience which are central determinants of employment and earnings (Becker, Reference Becker1975; Mincer and Polachek, Reference Mincer and Polachek1974), hence of welfare exit (Achdut, Reference Achdut2016; Achdut and Stier, Reference Achdut and Stier2018; Moore et al., Reference Moore, Wood and Rangarajan2012; Seefeldt and Orzol, Reference Seefeldt and Orzol2005). A better skill match has also been shown to increase employment stability (Kalleberg, Reference Kalleberg2008), post-unemployment stability (Tatsiramos, Reference Tatsiramos2009) and job satisfaction (Allen and van der Velden, Reference Allen and van der Velden2001; Eurofound, 2018). Thus, a better skill match is expected to enhance welfare exit. In addition, working under a standard employment contract (versus a temporary contract) almost inherently affects work continuity. This again is associated with the accumulation of work experience and might also be associated with wage growth and as a result with self-sufficiency. Finally, while workers’ ability to manage their working hours may help parents, more particularly mothers, to better balance work-family conflict (e.g. Allen et al., Reference Allen, Johnson, Kiburz and Shockley2013; Munsch, Reference Munsch2016), irregular work schedules, which imply working on official rest days and holidays, including evening shifts, when child-care facilities and schools are closed, may worsen the work-family conflict of low-income single mothers. Nonstandard schedule is a particular concern for parents of young children and more so for low-income families. Many such jobs include irregular schedules varying daily or weekly (Enchautegui et al., Reference Enchautegui, Johnson and Gelatt2015). Workers may receive little advance notice of their schedule and are expected to be available at any time, forcing last-minute adjustments of family and childcare responsibilities (Henly and Lambert, Reference Henly and Lambert2014). Taking care of young children may be the hardest with jobs with nonstandard schedules because these often require parents to be at work while their children are out of daycare and school (Enchautegui, Reference Enchautegui2013).
A word of caution. Conceptualizing work and welfare as competing choices for many welfare recipients, particularly single mothers, tends to ignore the characteristics and life circumstances that lead many of them to become welfare reliant. Mothers with multiple and chronic barriers to employment, who have already been on welfare for long periods, face major difficulties in achieving self-sufficiency. This reality has only little to do with the welfare policy that increased the incentive to work. Accordingly, in their critique of the economic rational choice model Duncan and Edwards (Reference Duncan and Edwards1996) argued that national policy is not the only factor in determining employment and benefit receipt among single mothers; the assumption of rationality and maximization of financial gains tends to ignore substantial differences in single mothers’ uptake of paid work by social group and geographical settings, such as local labor market opportunities. Furthermore, while economic calculations of benefit levels and wage rates are one important factor in single mothers’ decisions about entering the labor market, they are not the only form of rationality at work. Other rationalities may affect this choice and act independently of “economic pressures”, such as gender role perceptions, belief systems and their perceptions of motherhood and child care responsibilities.
Having said this, contemporary welfare policy in many western countries, although to a varying extent, leaves only little choice for many low-income single mothers whether to engage in paid employment. Extremely low public support combined with stringent income tests greatly strengthen the necessity of these mothers to work in order to provide for their families. Given this work-based regime and the accumulated empirical knowledge we have every reason to believe that the most prominent rationale at work is that suggested by standard economic theories. In fact, this was exactly the purpose of these new policy measures.
Job quality: Concept and measurement
Job quality is a central aspect of employment and inequality and thus important for labor market and welfare policies (Eurofound, 2012; 2018). It is commonly accepted nowadays that wages, while viewed as the main indicator of job quality, do not capture all aspects of work and other measures of job characteristics must be considered for a better grasp of what constitutes a “good job” (Dahl et al., Reference Dahl, Nesheim, Olson, Guillén and Dahl2009; Stier and Yaish, Reference Stier and Yaish2014; Stier, Reference Stier2015). Job quality, defined as a multidimensional phenomenon, embraces both objective and subjective concepts (Dahl et al., Reference Dahl, Nesheim, Olson, Guillén and Dahl2009; Green, Reference Green2006). The subjective point of view suggests that job quality is the ‘utility’ that a worker derives from his or her job. That utility depends on job features, such as wage, hours, and type of work, but it is subjective in that each worker has preferences regarding the different job features. On the other hand, from the objective standpoint job quality is constituted by the features of jobs that meet workers’ needs from work. As such, any objective concept stems ultimately from a theory of what human needs are and proceeds to investigate how far jobs meet those needs (Eurofound, 2012).
An additional commonly used differentiation in the domain of job quality is intrinsic versus extrinsic work attributes. This differentiation is primarily based on Herzberg’s two-factor theory of job satisfaction and motivation (Herzberg et al., Reference Herzberg, Mausner and Snyderman1959). The theory proposes that employee satisfaction derives most often from factors intrinsic to the work done, such as autonomy, achievements, responsibility, job recognition, and work that is challenging and interesting. These factors are called “motivators” because they are believed to be effective in motivating employees to perform better. On the other hand, factors extrinsic to the work itself, also called “hygiene factors”, include supervisory practices, salaries and other tangible benefits and working conditions.
Many studies have used this theoretical framework to investigate job quality (Olsen et al., Reference Olsen, Kalleberg and Nesheim2010). For example, Gallie et al. (Reference Gallie, Felstead and Green2012), examining job preferences, use factor analysis to demonstrate the distinction between intrinsic and extrinsic attributes of job quality. They found that those with higher levels of education were markedly more intrinsically orientated than those with less education. This may indicate the primacy of extrinsic attributes in the case of workers with low education.
However, views on the definition and measurement of job quality are quite divergent (Dahl et al., Reference Dahl, Nesheim, Olson, Guillén and Dahl2009; Piasna, Reference Piasna2017). The OECD framework for assessing job quality considers three objective job dimensions. These are earnings quality; labor market security; and quality of the working environment (OECD, 2017). The EU also developed a six-dimension measure of job quality, which mostly relies on the European Working Conditions Survey (EWCS), namely the Job Quality Index (JQI). The dimensions are wages; forms of employment and job security; working time and work-life balance; working conditions; skills and career development; and collective interest representation (Piasna, Reference Piasna2017). The European Foundation for the Improvement of Living and Working Conditions employed an objective concept of job quality, which means that only features that characterize jobs are included, leaving out those applying to individuals’ preferences. These measures include earnings, prospects, intrinsic job quality, and working time quality (Eurofound, 2012).
In this paper we consider objective measures of the job held, including monetary elements such as paid annual vacation, paid sick leave, pension savings and training fund; and non-monetary elements, namely use of the worker’s own skills or occupation in the workplace, existence of a standard employment contract (versus a temporary contract), and an irregular work schedule. However, we focused on a narrower conception of job quality than that offered by the international indexes detailed above, as we do not have further information on other objective measures or on subjective dimensions, such as work-life balance and working-hours preferences, let alone intrinsic attributes.
Holding a good and suitable job affects workers’ satisfaction, productivity, job continuity and stability, general health and wellbeing (Allen and van der Velden, Reference Allen and van der Velden2001; Arends et al., Reference Arends, Prinz and Abma2017; Dupre and Day, Reference Dupre and Day2007; Henseke, Reference Henseke2018; Milner et al., Reference Milner, Butterworth, Bentley, Kavanagh and LaMontagne2015; Stansfeld and Candy, Reference Stansfeld and Candy2006). Studies have found a link between long work hours (e.g. Caruso, Reference Caruso2006), working irregular hours or shifts (Chung et al., Reference Chung, Wolf and Shapiro2009; Presser, Reference Presser2005), working hours mismatch (Wooden et al., Reference Wooden, Warren and Drago2009) and job insecurity (Caroli and Godard, Reference Caroli and Godard2016; De Witte et al., Reference De Witte, Pienaar and De Cuyper2016; Henseke, Reference Henseke2018) to adverse physical and psychological well-being outcomes. A strong job demand combined with low job control leads to high-strain jobs which are associated with a higher propensity to cardiovascular disease (Kivimaki et al., Reference Kivimaki, Virtanen, Elovainio, Kouvonen, Vaananen and Vahtera2006) and depression (Bonde, Reference Bonde2008). A poor match between workers’ skills and duties are related to dissatisfaction and reduced well-being (Green, Reference Green2011).
Recent labor market processes include a rise in knowledge-based technologies, resulting in a growing demand for formal skills (Kalleberg, Reference Kalleberg2008; Piasna, Reference Piasna2017); globalization, with a higher incidence of migrant workers increasing international competition; and a general decline in the manufacturing sector (Burtless, Reference Burtless1995), worsen the labor market position of low-skilled or unskilled workers (Stier, Reference Stier2015). Furthermore, alternative work arrangements and non-standard employment, which has become more prevalent in the last two decades (Piasna, Reference Piasna2017), greatly increases workers’ exposure to bad job characteristics in the form of low pay and absence of health insurance and pension benefits, working time and the flexibility to manage it, and poorer prospects for career advancement (Eurofound, 2018; McGovern et al., Reference McGovern, Smeaton and Hill2004). These non-standard forms of employment almost automatically suggest low job security, especially for low-skilled workers and offer less predictability in terms of income and working hours (Drahokoupil and Piasna, Reference Drahokoupil and Piasna2017; Rubery and Piasna, Reference Rubery and Piasna2016). These changes put welfare recipients in a precarious position: jobs open to them might offer only minimum wage, part-time and temporary employment contracts and limited or no employer-sponsored benefits. Previous studies have demonstrated that low-skilled workers lag behind those with higher skills in every aspect of their employment. They have lower job security, and their jobs provide fewer opportunities for achievement, worse job content and less flexibility, fewer benefits and less regular hours (e.g. Acs and Loprest, Reference Acs and Loprest2005; Clark, Reference Clark2005; Stier, Reference Stier2015). Also, international evidence suggests a general decline in non-wage dimensions of job quality, such as forms of employment and job security, working conditions and union representation over the past decade with some variation across countries (Piasna, Reference Piasna2017). Based on the EWCS, relative stability was found in 1995-2010 in measures of ‘intrinsic job quality’. Examples are use of skills and discretion, good physical environment and work intensity as regards labor effort at work. By contrast, there was a rise over time in the work-time quality measure, referring to a set of job features that are generally conducive to a good work–life balance (Eurofound, 2012). An unequal trend was found in job discretion across clusters of countries with different institutional regimes and across groups of employees. However, no evidence was found for divergence of job discretion between high- and low-skill occupations (Holman and Rafferty, Reference Holman and Rafferty2018).
Job quality of welfare recipients
Scant empirical research has examined job quality – beyond wages and working hours – of former and current welfare recipients. Studies that have done so usually consider hours of work, hourly wage and employer-sponsored health insurance (Danziger and Johnson, Reference Danziger and Johnson2004; Johnson and Corcoran, Reference Johnson and Corcoran2003). Some of these studies include other components of job quality, among them pension plan, paid sick leave and paid vacation (Acs and Loprest, Reference Acs and Loprest2001); occupation and industry, night schedule and multiple jobs (Loprest, Reference Loprest2003; Loprest and Zedlewski, Reference Loprest and Zedlewski2006); availability of employer-sponsored benefits, union membership, and occupation and industries (Foley and Schwartz, Reference Foley and Schwartz2003); possible promotion (Holzer et al., Reference Holzer, Stoll and Wissoker2004). However, most of the above studies focused on welfare leavers – that is, had exited the welfare system, and are descriptive in nature. Exceptions are Johnson and Corcoran (Reference Johnson and Corcoran2003), who tested whether the skill and task content of work experience affected movement from bad jobs to good jobs; Foley and Schwartz (Reference Foley and Schwartz2003), who estimated the effect of the Self-Sufficiency Project on job quality; and Holzer et al. (Reference Holzer, Stoll and Wissoker2004), who examined various determinants of job performance and retention rates, among them job characteristics. Overall, these studies suggest that the jobs held by current and former welfare recipients can generally be characterized as low-wage with minimal benefits and with little or no prospects for improvement over time. For example, Johnson and Corcoran (Reference Johnson and Corcoran2003) indicated that although continuous employment improved welfare recipient’s chances of working at a good job, most women who had worked continuously over roughly a five-year period did not hold a good job. With respect to use of skills, the use of hard skills on the job improved the likelihood of a transition from a bad job to a good job. Using the same sample and a similar definition of a “good job,” Danziger and Johnson (Reference Danziger and Johnson2004) showed that welfare mothers who worked at jobs that required cognitive skills commanded higher wages and a greater return on work experience than women with jobs requiring only soft skills. Based on a sample of recently hired welfare recipients, Holzer et al. (Reference Holzer, Stoll and Wissoker2004) indicated that a job’s characteristics, such as jobs with reported promotion potential, showed a somewhat lower exit rate and appeared to affect retention rates independently of observed worker attributes.
The current study provides an opportunity to examine the job quality of one of the weakest segments in the labor market: single mothers who are also welfare recipients. The latter group is also associated with low-skills, and, because of their multiple disadvantages, with other employment barriers as well. Similar to other low-skilled workers (Dumont et al., Reference Dumont, Rayp and Willemé2012) welfare recipients, in particular single mothers, have low bargaining power in the labor market due to their low human capital, poor health, and heavy family burden, and therefore are expected to have low-quality jobs in terms of material rewards (i.e. wages) but also in other job conditions.
Although there is substantial research regarding the determinants of welfare exit, little is known about the effect of various job quality indicators, beyond wages, on the likelihood of leaving welfare. Aiming to fill this gap, we ask whether monetary and non-monetary aspects of job quality are crucial for welfare exit. The Israeli context is appropriate here in two respects: the characteristics of Israeli labor legislation and the characteristics of Israeli welfare policy. The following section addresses these issues.
The Israeli context
Employment benefits and work conditions in Israel
Israel implemented several laws and regulations that are relevant for the current study, including the Minimum-Wage Law (1987), the Annual Leave Law (1951) referring to annual vacation, and the Sick Pay Law (1976). In 2007, the minimum wage level was 42 percent of the average wage and 56 percent of the median wage for a salaried worker employed fulltime (compared with 24% and 31% in the US, respectively) (OECD, 2010). Some 13 percent of salaried workers in Israel earned an hourly wage not exceeding the minimum wage. This rate was higher for women than for men (15% and 11%, respectively) (Bank of Israel, 2007); 40 percent of salaried workers had total earnings below the monthly minimum wage (National Insurance Institute, 2008). These figures indicate the high prevalence of low-wage employment in Israel.
As for annual vacation and sick leave pay, both relate to the scope of employment, and the minimum number of days guaranteed is for full-time employees (14 days’ annual vacation and 18 days’ sick leave: 1.5 days for each month of employment). Annual vacation days also relate to the employee’s tenure at the same workplace, and entitlement is enjoyed only by employees who have worked a full year for the same employer (Ministry of Labor, Social Affairs and Social Services, 2018). Thus, those with fragmented employment, as is the case with many welfare recipients, do not necessarily benefit from it. Sick leave pay is subject to a medical certificate, and unlike vacation days, accrued and unused sick days are not redeemable. Finally, and importantly, employees are entitled to use a proportion of their accrued sick days for the illness of a child, parent or a spouse, all subject to certain requirements specified under the law. In addition, the Hours of Work and Rest Law (1951) regulates employees’ working hours, overtime, night labor and rest days (including holidays), as well as the payment for each category of work. With respect to Jewish employees, Saturday (starting Friday eve) is the official weekly rest day. An employee is entitled to no less than 150% of his or her regular salary for every hour of work during the weekly rest hours (Ministry of Labor, Social Affairs and Social Services, 2018).
These laws and regulations apply to all workers, full- or part time, in standard or non-standard employment contracts.
Beyond these labor laws, there is a National Health Insurance law (1995) in Israel that covers the entire population. Employers and employees are co-obliged to sponsor it, while the insurance premiums are relative to wages and are in accordance with a progressive scheme. This sum is automatically collected from employee and deducted from the workers’ wages. Unemployed people pay a reduced sum directly to the National Insurance Institute (NII). Income Support Benefits (ISB) are adjusted to that sum. Given this, in our estimates we did not include a measure of employer-sponsored health insurance as a measure of job quality. Finally, pension plans were not obligatory in the years covered by this study (2003-2007) but became so in 2008. Accordingly, we included this measure as a non-mandatory compensation attribute. In both pension plans and training funds usually the workers contribute 2.5% of their gross earnings and the employer contributes an additional 7.5%.
Welfare policy in Israel
The Income Support Program (ISP) administered by the NII in Israel is a selective program aimed to guarantee minimum income for families with limited means. The level of ISB for single mothers depends on the number of dependent children, and the amount of benefits paid to the family is determined by an income test.Footnote 1 In addition, eligibility for ISB in Israel is contingent on a work test. Welfare recipients are required regularly to report to the local employment service and to accept any job offered them by the employment officer. Recipients who fail the work test and do not work are subject to benefit deprivation for two months. Work-test exemption is given to certain groups of recipients, among them mothers of young children and recipients defined as “unplaceable” (temporarily or permanently) due to their own or their children’s health condition, or other personal problems.
In 2003 the public assistance programs were restructured in Israel to increase employment and decrease welfare receipt: ISB and the disregarded income level were reduced; the previous exemption from the work test for mothers of preschool-age children (younger than age seven) was limited to mothers of children younger than age two (National Insurance Institute, 2002–2003). While more single mothers were propelled into the labor market, the new legislation did not set specific working-hour requirements. Sanctions were imposed only on those refusing to work at all, so in practice recipients could work very little and still receive supplemental ISB, which remained extremely low after the reform. This reform was implemented throughout Israel.
More generally, Israeli policy differs from that of many other countries that introduced welfare reforms, particularly from the TANF program under the US federal reform, in two respects: first, it is a uniform nationwide reform with no local discretion in designing and operating cash assistance and other work-related programs as in the US (Bentele and Nicoli, Reference Bentele and Nicoli2012) or in Sweden (Bergmark and Bäckman, Reference Bergmark and Bäckman2004). Secondly, and importantly, several policy components exist that enable combining work and welfare for long periods; thus, we can easily examine the job quality of welfare stayers – those who combine welfare receipt with employment – as opposed to welfare leavers. Among these are the absence of minimum working-hour requirements, employment exemptions for mothers with very young children, and no time limit for benefit receipt. This policy makes combining part-time work and welfare for long periods quite common. Furthermore, we found no evidence that the two most common measures of job quality among welfare recipients, namely hourly wage and working hours, which are also the most influential factors for welfare exit, correlate with the measures suggested in this study. This means that welfare recipients working at part-time minimum-wage jobs – which is associated with being a combiner – do not necessarily suffer from other bad-job conditions. Against this labor market and policy context, the current study examined the contribution of various job characteristics to welfare exit.
Method
Data and sample
The sample was drawn from National Insurance Institute administrative data, which covered all single mothers, 46,000 in number, receiving ISB in May 2003. From this file the Ministry of Industry, Trade and Labor randomly selected a national sample of 2,800 single mothers and interviewed them in summer 2003 and again in summer 2007. Our follow-up covers the 51-month period from May 2003 to July 2007. The response rate was relatively high: 79 percent of the sample attended both interviews. At the time of the second interview 2,350 (83%) mothers were employed. Since we were interested in the job quality of former and current welfare recipients, we limited the sample to this group, but we further accounted for selection for paid employment assuming that only women who receive wages equal to or higher than their reservation wage participate in paid employment (Heckman, Reference Heckman1979) (see Data Analysis, footnote 5). From this point onward, we refer only to mothers employed at the time of the second interview.
Based on self-report, the two surveys provide detailed information on respondents’ demographics, work history, education level, health condition, employment and earnings, current working hours, grant of mandatory and non-mandatory fringe benefits in the current workplace. We characterized job stability – switching jobs – between the two interviews through retrospective questions from the second interview. Wages, work hours and monetary and non-monetary job characteristics were for the current job as recorded in the second interview. Since our dataset did not include information on all aspects of job quality at the first measurement point, we could not estimate job transitions.
Measurement
Outcome variable - Welfare exit
The outcome variable was coded as 1 when a mother was in employment at the second observation point in 2007 and had been off welfare for at least three months (May-July, 2007). All other cases were coded as 0. This category includes mothers who were in employment at the second observation point in 2007 and combined employment with benefit receipt and a few mothers who had been off welfare for only one or two months prior to the observation point.
Explanatory variables – 1. Job quality
To estimate recipient’s job quality, respondents were asked to state if each of the following conditions applied to their main job at the time of the second interview (2007)Footnote 2: (a) use of own skills or occupation to some extent at the workplace (=1); (b) existence of standard employment contract (i.e. traditional direct employment and non-temporary job contract) (=1); (c) irregular work schedule (i.e. working on weekends and holidays); (d) receiving both annual paid vacation and paid sick leave (=1); (e) receiving at least one non-mandatory monetary benefit: pension or training fund (=1). In a preliminary analysis we applied a Varimax exploratory factor analysis (EFA) (Henson and Roberts, Reference Henson and Roberts2006) for the different aspects of job quality in order to create a job quality index, and two factors emerged.Footnote 3 The first was mandatory and non-mandatory benefits; the second was use of skills and the existence of standard employment contract. Unexpectedly, hourly wage did not correlate at all with any of the other components tested. The same applied to working an irregular schedule. We then estimated a regression model for welfare exit using the two factors variables as a measure of job quality. This model yielded similar results to those of a model that considered each job quality component separately. Since these job characteristics were not highly correlated with each other, we decided to present a regression model that considered the five aspects separately (for correlation matrix see Appendix, Table A.1).Footnote 4
2. Control variables
We measured each explanatory variable for the baseline year (2003), including socio-demographic characteristics (marital status; mother’s age; number of children; age of youngest child; immigration); chronic health condition measured through two questions: those who reported being in fair or poor health were further asked about their specific health problems and whether they suffered from one chronic condition or more, such as arthritis, asthma, emphysema, back problems, diabetes, fatigue, headaches, heart condition, hepatitis, cirrhosis, high blood pressure, seizures or mental disorders (see Son et al., Reference Son, Dyk, Bauer and Katras2011). Mothers who reported at least one problem were defined as suffering from a chronic health condition; level of education; information on work experience (number of months worked between interviews); and information on job stability (number of jobs replaced between interviews). We also indicated changes occurring from the beginning (2003) to the end (2007) of the follow-up period in the following events: whether the mother gave birth to an additional child; whether the number of eligible children declined; whether the mother got married; whether the mother acquired a high school diploma; and whether the mother participated in educational or training program. These events were introduced as five dummy variables indicating the occurrence of each (=1) during the follow-up. In this way we accounted for possible changes in individual characteristics. Finally, among the explanatory variables we considered the mothers’ hourly wage. We did so using a dummy variable indicating an hourly wage exceeding the minimum hourly wage by at least NIS 1 (=1).
Table 1 summarizes the study variables for the entire sample and according to welfare-use states as measured in 2007. The top panel of the table shows the prevalence of each welfare-use state, indicating that in 2007 two-thirds of the mothers who received ISB in 2003 were no longer on welfare. Overall, about 55 percent of the mothers earned at least NIS 1 above the minimum wage, indicating a high incidence of minimum-wage earners in this group. The mean and median hourly wage was slightly higher for welfare leavers (NIS 25 and NIS 22.61, respectively) than for welfare stayers (NIS 22 and NIS 20.34, respectively) – sums about equal to the minimum wage (NIS 20.0 at the time of the second interview). Regarding non-monetary measures, a low rate of 27.7 percent reported that they used their skills or occupation to some extent in their job, and most of the mothers (81.1%) were employed under standard employment contract. Only 17 percent worked irregular schedules (weekends and holidays).
TABLE 1. Summary statistics of the study variables by welfare exit
Note. **p<.01.
Regarding monetary aspects, the rate of mothers who received mandatory attributes (both annual vacation and sick leave) was 58.4 percent, although these benefits were obligatory and should have been given automatically; and only 32.9 percent received at least one non-mandatory attribute. The study groups showed statistically significant differences on all measures of job quality. Welfare leavers were more likely than employed welfare stayers to report that they used their skills or occupation to some extent in the work place (33.5% versus 16.7%, respectively), to be employed under a standard employment contract (86.2% versus 72.8%), received mandatory non-wage compensation attributes (63.1% versus 49%) and received non-mandatory attributes (39.7% versus 20%). An exception was working an irregular schedule: welfare leavers proved more likely to report working that way than welfare stayers (20.1% versus 10.8%, respectively). This association was not in the predicted direction (for full description and summary statistic of the control variables see Appendix, Table B.1).
Data analysis
We applied two logistic regression models to estimate the probability of being off welfare in the outcome period (2007). The first model contained the five components of job quality: use of own skills or occupation to some extent in the workplace; existence of standard employment contract; irregular work schedule; receiving both mandatory non-wage compensation attributes: annual paid vacation and paid sick leave; and receiving at least one non-mandatory non-wage compensation attribute: pension or training fund. To the second model we added the control variables listed in the Measurement section and the hourly wage as a monetary measure of job quality. Change in marital status and participation in education/training programs were excluded as the rate of mothers on both these variables was very low (see Appendix, Table B.1). Both models controlled for the probability of being employed in the outcome period (2007) (see Appendix, Table B.2 for results excluding correction for selection effect).Footnote 5
Findings
Multivariate results
Table 2 shows the results of two logistic regression models predicting welfare exit in the outcome period (2007). By comparing the two models we could examine whether the contribution of the job-quality measures remained significant, and whether, and to what extent, they changed, after controlling for wages and other factors closely related to welfare exit.
TABLE 2. Logistic regressions: Predication of welfare-exit
Note. *p<.05 ** P < .01.
(a) Reference group is welfare stayers (34%).
(b) Reference group consists of divorced, separated and widowed mothers.
(c) Reference group consists of mothers with below high school diploma.
Coefficients (standard errors in parenthesis) of the independent variables for the first model are presented in the first column; the second column shows the odds ratio; and the third column shows the average marginal effect (AME) of each of the explanatory variables. The fourth, fifth and sixth columns show these values for the second model, all relative to the welfare stayers’ reference group. The AME of any explanatory variable indicated change in the probability of occurrence of an event that defined the outcome variable (probability of exiting welfare) due to a one-unit change in the explanatory variable, other variables being constant. By multiplying the marginal effect by 100, we obtained the change in percentage points. The advantage of the marginal effects is that one can compare their absolute values among the explanatory variables and across different models.
Starting with the first model (Model 1), each of the job quality components was associated with welfare exit. Mothers who reported that they used their skills or occupation in their current job were more likely by 14.7 percentage points to leave welfare than those who did not use their skills; working under standard employment contract also increased this probability by 13.6 percentage points. Irregular work schedule, which indicates lower job quality in the case of welfare recipients, increased the probability of leaving welfare by 15.5 percentage points. Perhaps mothers who showed greater labor supply flexibility and were willing to take jobs that required working weekends and holidays (including an evening schedule in both) were more motivated to increase their working hours and more flexible regarding how to do so, hence were more likely to exit welfare. Regarding monetary aspects, while receiving mandatory attributes of compensation increased the probability of a mother leaving welfare by 6.3 percentage points, the effect of non-mandatory attributes was more pronounced and increased this probability by 15.1 percentage points.
The second model (Model 2) indicates that the five measures of job quality remained statistically significant for welfare exit, while their AME changed only slightly, after controlling for hourly wage and individual’s characteristics. As expected, wages were extremely important in determining welfare exit. Earning at least NIS 1 more than the hourly minimum wage increased the probability of leaving welfare by almost 10 percentage points. As for the individual’s characteristics, our study generally confirms findings from previous studies about the determinants of welfare use: among these are older age, heavier family burden, being immigrant, lack of recent work experience and suffering from chronic health problems (Achdut and Stier, Reference Achdut and Stier2018; Juon et al., Reference Juon, Green, Fothergill, Kasper, Thorpe and Ensminger2009; Seefeldt and Orzol, Reference Seefeldt and Orzol2005; Wood et al., Reference Wood, Moore and Rangarajan2008). Mothers with academic education were more likely to exit welfare, as expected. However, there was no significant difference among all other levels of education, controlling for job quality.Footnote 6 While educational level was related to job quality, formal education was found not as important in affecting welfare exit (Achdut, Reference Achdut2016; Heflin, Reference Heflin2003; Juon et al., Reference Juon, Green, Fothergill, Kasper, Thorpe and Ensminger2009; Nam, Reference Nam2005).
Overall, our findings indicate that holding wages constant, other monetary and non-monetary job quality aspects affected the odds of leaving the welfare system.
Discussion
Our findings show that monetary aspects in the form of non-wage attributes affect the probability of leaving welfare. These can be seen as measures that support employment – especially paid annual vacation and sick-leave pay – which afford mothers some degree of flexibility without losing income or jeopardizing their employment. This flexibility is needed especially in respect of single mothers, because they themselves bear the dual role of caregivers and providers, and because many welfare recipients also lack support from other family members or friends. Hence, they are more prone to frequent absenteeism from work (Holzer et al., Reference Holzer, Stoll and Wissoker2004).
Other monetary attributes, such as pension plan and training fund, were found strong predictors of welfare exit. These measures of job quality reduce the cost of employment versus welfare receipt as they introduce a form of saving that is heavily sponsored by the employer. In addition to the importance of these measures for welfare exit, they also enable mothers to have some sort of savings, particularly for old age. While the sum of money in these funds is indeed low and usually insufficient for a living, due to welfare recipients’ fragmentary employment and their relatively low earnings, it does provide some resources for the future.
Counter to our expectations, working on irregular schedule enhanced welfare exits. No formal child care was available at the times when these women worked, so they probably used informal childcare. In the US, for example, low-income single parents tend to rely on relatives’ help when working on nonstandard schedules (Enchautegui et al., Reference Enchautegui, Johnson and Gelatt2015). This informal care probably plays a key role in work-family balance here, as grandparents are an important source of support and assistance in Israel (Even-Zohar and Garby, Reference Even-Zohar and Garby2016; Okun, Reference Okun2016). Hence, these women use irregular hours as a strategy to balance work and family demands and vice versa.
There are several limitations to this study. One is a lack of information on factors linked to the productivity of workers that influence their likelihood of leaving welfare and getting a quality job. Among these are local labor market conditions and the type and quality of skills. These affect labor and welfare use outcomes; particularly well documented is the role of local economic conditions (e.g. Achdut and Stier, Reference Achdut and Stier2016, Reference Achdut and Stier2018), which might also affect the quality of jobs that can be obtained by welfare recipients. Additionally, there is no longitudinal information but rather two measurement points, so we were unable to control for the dynamics of welfare use, i.e. exits and re-entries during the follow-up period. We had information on the status of employment and welfare receipt at each of our measurement points. However, retrospective questions at the second observation period provided information on unemployment spells (translated into overall work experience) and job stability, two central factors both linked to the quality of the work and to welfare exit. We further accounted for some changes that might have occurred over these years, as we incorporated time-varying variables in our models indicating life events expected to affect the outcome under discussion.
The implications emerging from this study concern both practice and policy. Social workers, employment officers in the local employment services, and other advocacy and mediation agencies working with welfare recipients, helping them in their passage to self-sufficiency, should consider components of the jobs offered them other than wages and working hours. Accordingly, job quality needs to be a more prominent aspect of labor market policy, especially in policy measures aimed at welfare recipients with multiple employment disadvantages. More attention should be paid to workers’ perception of the quality of their work and how policies and practices at the level of the worker and of the social services agencies can contribute to achieving better results.
This perception runs counter to the assumptions underlying welfare reform in many countries, particularly in liberal market economies, namely that the best way to succeed in the labor market is to take any job. This “work-first” model assumes that consistent work will eventually lead to a living wage and economic self-sufficiency (Johnson, Reference Johnson2007). However, while the approach that emphasizes immediate job placement as the best road to self-sufficiency can work for recipients with higher skills and training, who can find better jobs, for most recipients the work-first notion presents few labor market prospects in terms of job quality. Accumulated empirical evidence indicates poor long-term labor market outcomes for many welfare recipients, in terms of employment and earnings growth (e.g. Achdut and Stier, Reference Achdut and Stier2016; Johnson, Reference Johnson2007; Wood et al., Reference Wood, Moore and Rangarajan2008; Wu et al., Reference Wu, Cancian and Meyer2008) and continuous and repeated welfare use (e.g. Achdut and Stier, Reference Achdut and Stier2018; Bäckman and Bergmark, Reference Bäckman and Bergmark2011; Königs, Reference Königs2018; Lightman et al., Reference Lightman, Mitchell and Herd2010) – let alone disconnectedness (e.g. Blank and Kovak, Reference Blank, Kovak, Heinrich and Scholz2009; Loprest and Nichols, Reference Loprest and Nichols2011). Some studies also argued that work-related ‘conditions of conduct’ (Davis, Reference Davis2019), namely immediate work requirements and sanctions which are the core of the “work-first” notion, may be ineffective or even harmful for certain groups of recipients, among them single mothers and those with multiple needs that are particularly distant from the labor market (Dean, Reference Dean2003; Reeves and Loopstra, Reference Reeves and Loopstra2016). For these groups, it has been suggested an alternative approach that prioritizes the right to seek appropriate work rather than an immediate return to formal employment (Dean, Reference Dean2003), as greater stringency makes these recipients further distance from paid work (Reeves and Loopstra, Reference Reeves and Loopstra2016). For example, Davis (Reference Davis2019) indicated in his US study that in states that impose harsher sanctions and stricter job search requirements, low-educated single mothers endure worse mental health.
This reality, along with an understanding of the significance of the quality of work, necessitates a shift in policy aimed at welfare recipients, in both pre-employment services and programs and job placement. Our findings contribute to the long-term debate on the best road to self-sufficiency (e.g. Kim, Reference Kim2010; Kluve, Reference Kluve2010; Ochel, Reference Ochel2005). They highlight the need to prioritize investment in program participants’ human capital development over the “work-first” approach; this will be by skill acquisition and vocational training to increase the odds of many low-skilled welfare recipients’ finding better quality jobs. This is because the type of job matters, and not only as regards welfare exit. However, as indicated by evaluation studies and several systematic reviews of welfare-to-work initiatives (e.g. Greenberg et al., Reference Greenberg, Michalopoulos and Robins2003; Kim, Reference Kim2012; Ochel, Reference Ochel2005), human capital intervention should focus on training rather than general education, and be accompanied by further job search assistance. It may even combine some employment with training. A mixture of both approaches proved key in achieving better results (Blank, Reference Blank2003). Furthermore, some aspects of job quality can also be improved by interventions directed at services or program operators, such as rewarding agencies and services for placement in jobs with particular characteristics – for example, that offer a standard employment contract or that match one’s skills or offer on-the-job training. As indicated by Foley and Schwartz (Reference Foley and Schwartz2003), this strategy would clearly require co-operation between social service agencies and groupings of employers or unions. Without doubt, such policies involve greater public resources and time, but eventually should lead to a more sustainable passage from welfare to employment.
Overall, most of the jobs available for welfare recipients are low-wage and constitute a narrow range of occupations and industries. Broadening these opportunities requires major public resources and long-term investment and effort, starting with the education systems. These are beyond the scope of the current discussion. However, some policy measures, as suggested, can be taken to improve the labor market experience of welfare recipients, to enhance work continuity and to promote decent material well-being through better jobs, which might lead to upward mobility. This holds for the design and implementation of welfare policies, including welfare-to-work programs, in other countries than Israel.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0047279419000977
