Abstract
This article disaggregates high- and low-status care work across eight liberal welfare regimes: Australia, Canada, Iceland, Ireland, Israel, Switzerland, the United Kingdom and the United States. Using Luxembourg Income Study data, descriptive and multivariate analyses provide support for a ‘migrant in the market’ model of employment, notwithstanding variation across countries. The data demonstrate a wage penalty in both high- and low-status care employment in several liberal welfare regimes, with the latter (service jobs in health, education and social work) more likely to be part-time and situated in the private sector. Migrant care workers are found to work disproportionately in low-status, low-wage types of care and, in some cases, to incur additional wage penalties compared to native-born care workers with equivalent human capital.
Introduction
The majority of cross-national studies of care work are qualitative, focus on a single country and/or provide comparisons across welfare regimes (e.g. Brennan et al., 2012; Budig and Misra, 2010; Lightman, 2017). Van Hooren (2012), for example, compares elderly care services across three welfare regimes: familialistic, social democratic and liberal. She posits that within market-oriented liberal welfare regimes, there is a ‘migrant in the market’ model of employment; this model disproportionately locates foreign-born care workers in the less regulated, more precarious private sector, where working conditions are poorer and wages overall lower.
This article builds on Van Hooren’s (2012) analysis in two specific ways. First, I approach the question quantitatively, examining the paid care economy in eight liberal welfare regimes. I measure how and if the employment and earnings of care workers are similar or different for individual countries within liberal welfare regimes, while disaggregating higher and lower status jobs in care. Second, I focus on the labour market outcomes of foreign-born care workers (the ‘migrants in the market’) and examine whether immigrants (also called ‘migrants’ without distinction) are more likely than native-born populations to perform low-status care work or incur a wage penalty relative to native-born care workers, regardless of their human capital and individual attributes.
Using the micro data files of the Luxembourg Income Study (LIS), I examine who is employed in what type of care work and examine the earnings penalty of high- and low-status (immigrant) care workers in Australia, Canada, Iceland, Ireland, Israel, Switzerland, the United Kingdom and the United States, in order to address the following key questions:
Does distinguishing between care work occupations with higher and lower status allow for a more nuanced analysis of any ‘care penalty’ within liberal welfare regimes?
How do immigrant care workers, in particular, fare in liberal welfare regimes?
Do commonalities and/or variation in the employment and earnings of immigrant care workers within liberal welfare regimes bolster Van Hooren’s (2012) suggestion of a ‘migrant in the market’ model of employment?
Ultimately, a significant wage penalty for high-status care work is found in the majority of countries examined, while a significant wage penalty in low-status care work is found in only three out of the eight liberal welfare regimes, when compared to professional or service and sales jobs in non-caring industries. In addition, the labour-force characteristics and demographic makeup of high- and low-status care work are found to be distinct, reinforcing the need for nuanced definitions of ‘care’. In liberal welfare regimes, migrant care workers are found disproportionately in low-status, low-wage jobs, and in certain countries, they incur additional wage penalties by virtue of their immigrant status compared to native-born care workers with equivalent human capital. Thus, overall, the data provide considerable support for a ‘migrant in the market’ model of employment. Yet, the substantial variation within liberal welfare regimes suggests a need for more in-depth analyses of different types of paid care at the comparative, national and subnational levels, tasks for future research.
Liberal welfare regimes and the commodification of care
Welfare regime theory rests on the assumption that countries can be grouped into clusters based on the quality of social rights, the extent of social stratification and the relation of state, market and family (Esping-Anderson, 1990). However, while welfare regime categorizations provide a convenient means to compare care economies, some scholars suggest they are outdated (and Eurocentric) and/or that differences within welfare regime categories are more important than differences between them (e.g. Brennan et al., 2012; Jensen and Lolle, 2013). Budig and Misra (2010), in their seminal study of care work, find significant wage variation across numerous welfare regimes. However, Williams (2012) suggests there has been considerable convergence in care provisioning across wealthy nations, both in the commodification of care services and in the employment of migrant women.
Liberal welfare regimes, in particular, are thought to assign ‘key roles to labour markets and families, with the state’s role largely limited to providing assistance targeted at those least well-off’ (Mahon et al., 2012: 421). They are characterized by a preference for market solutions to welfare problems, leading to relatively low levels of social spending, limited regulation of the labour market and high levels of overall inequality (Banting and Myles, 2013; Lightman and Lightman, 2017). Since the mid-20th century, liberal welfare regimes have experienced a consistent trend of converting the objectives of health and education from the delivery of a public good, to the sale of a market commodity tailored to specific (economically advantaged) groups. Ilcan (2009) terms this process ‘privatizing responsibility’, while Oesch (2015) notes that liberal welfare regimes lead to occupational and wage polarization within paid care work; by allowing earnings in interpersonal services to adjust to lower productivity growth, there has been an increase in low paid service jobs in elder and childcare and growing disparities in care work wages. Growth in the caring industry has been disproportionately located in low-status, precarious occupations (Oesch, 2015; Ruhs et al., 2010).
Thus, despite considerable variation at the state or provincial level, numerous similarities are identified across the eight liberal welfare regimes analysed in this article (Australia, Canada, Iceland, Ireland, Israel, Switzerland, the United Kingdom and the United States). Iceland is unique in that it offers universal, state-supported childcare services. However, in each of the other countries, the cost of childcare usually exceeds any government subsidies, except in the case of very poor families (Mahon et al., 2012). Long-term elder care is provided through a patchwork of arrangements in each country, with financing tied to socioeconomic status. In the majority of cases, national health insurance covers most medical and hospital-related costs for the elderly, as well as some (usually small) portion of the cost of medically-related care either in private homes or institutions. However, the cost of non-medical care (e.g. assistance with activities of daily living) is usually covered through subsidies at the subnational level, private insurance, personal savings and voluntary services (Boris and Klein, 2006; White, 2016). In addition, the quality of care accessed by those at higher and lower income levels is often widely divergent (Barron and West, 2013; Lightman and Lightman, 2017).
This article, providing a quantitative analysis of both high- and low-status paid care work within eight liberal welfare regimes, attempts to examine both the commonalities and the disparities within their care economies, with a focus on the employment and earnings of foreign-born workers.
Immigrant stratification in paid care
Internationally, migrants coming from poorer to richer nations for employment are disproportionately located in secondary sectors of the labour market, engaging in precarious work that offers limited protections and benefits, and allows workers minimal autonomy, recourse or control (Kalleberg, 2011; Vosko, 2009). Paid care employment is increasingly conceptualized within a transnational labour market (or ‘global care chain’), where disadvantaged or poor immigrant women (who are often racialized) provide care for pay in wealthier countries, typically in private sector and lower paying service jobs (Hochschild, 2012; Lightman, 2017).
Against a backdrop of rapidly aging populations and low domestic birth rates, migrant workers provide a market-based solution to mounting ‘care deficits’ in high- and middle-income countries, as native-born women increasingly (re-)enter the workforce (Budig and Misra, 2010; Van Hooren, 2012). Immigrant care workers often arrive with temporary work permits designed to discourage their broader integration or settlement and may encounter widespread workplace discrimination and abuse, often while negotiating their own intergenerational family separations, in a context of growing public hostility about wage undercutting and displacement for the native-born (Anderson, 2010; Parreñas, 2010).
In many liberal welfare regimes, migrant labour has become a defining feature of the care sector. Van Hooren (2012) finds that migrant employees work longer hours and do more night shifts than their native-born peers in elder care in liberal welfare regimes and that this polarization is especially acute for those employed in the private sector. Howes et al. (2012), for their part, document how recent shortages of nurses and teachers in the United States have been met by efforts to recruit from overseas, a strategy that makes it easier to restrict wage growth and to postpone investments in the state university systems that provide the bulk of training for these jobs. The United Kingdom, too, is increasingly dependent on nurses imported from Africa, while the Philippines, Indonesia and Vietnam export nurses as well as childcare workers to numerous wealthy countries, including Australia, Iceland and Ireland (Cangiano, 2014; Hugo, 2009). In Canada, the federal Caregiver Program, instituted in 1992, has been criticized both for being exploitative to migrants and for being based on false premises tied to a rapid ‘pathway to citizenship’ (Tungohan et al., 2015).
Thus, while the particular configurations of paid care employment vary depending on a country’s occupational structure and policies on importing immigrant labour, within liberal welfare regimes, care work relies heavily on the premises of immigrant stratification (Folbre, 2012; Hochschild, 2012). Yet depending on how care work is defined and operationalized, the outcomes of immigrant workers are highly divergent.
Operationalizing care: disaggregating high- and low-status occupations
Feminist scholars identify a particular economic disadvantage related to care work, or a ‘care penalty’, due to its association with ‘women’s work’ (Budig and Misra, 2010; England et al., 2002). Typically, care work is defined as employment that involves face-to-face interactions with children, the elderly or people with complex healthcare needs (England et al., 2002; Folbre, 2012). Oftentimes, care work is also identified as precarious (Standing, 2011; Vosko et al., 2003).
Nonetheless, there remains considerable disagreement as to how to best define and operationalize care work. Some scholars focus only on ‘direct’ care, which typically involves personal connections and emotional attachments to care recipients (e.g. Baughman and Smith, 2012; Ulrich et al., 2014). Yet other scholars include in their definition ‘indirect’ care, which is commonly conceptualized as providing support for direct care, and includes occupations such as janitorial work, food service preparation and domestic cleaning (Duffy et al., 2013; Lutz, 2012).
Narrower definitions of care work focus only on specific low-wage occupations such as home health aides or childcare minders (see Anderson and Hughes, 2010; Kaye et al., 2006). Broader definitions, such as those used in most cross-national research, encompass all employment in health, education and social services, often for reasons of limited data availability (e.g. Budig and Misra, 2010; Duffy et al., 2013). England et al. (2002), for example, combine individuals working in childcare, all levels of teaching (from preschool to university professors) and all types of healthcare providers (from nurses’ aides to registered nurses to doctors), and include individuals in the ‘helping professions’ (e.g. therapists, social workers and clergy).
Weeden (2002) applies a framework of ‘social closure’ (Weber, 1956) to disaggregate higher and lower status caring occupations in the United States. This framework identifies whether specific caring occupations control access to the profession and collectively negotiate employment conditions and benefits (i.e. the degree of social closure achieved). A high degree of social closure is exemplified in the case of professional occupations such as doctors and nurses, as well as many teachers. Low social closure is characteristic of occupations with high turnover and low entry barriers, such as many services and sales jobs in health and education, including home health aides and nannies (Barron and West, 2013; Lightman, 2017).
Applying Weeden’s (2002) framework in the context of a cross-national comparison, this article disaggregates higher and lower status care work. The following section outlines the care classification scheme developed and applied, which seeks to examine both commonalities and variation in the employment and earnings of foreign-born care workers in liberal welfare regimes.
Research design
To examine care work within liberal welfare regimes, this article uses the most recent micro data available from the LIS spanning the years 2010–2013. The eight countries analysed, Australia, Canada, Iceland, Ireland, Israel, Switzerland, the United Kingdom and the United States, are all most often designated ‘liberal’ welfare regimes (Ebbinghaus, 2012; Scruggs and Allan, 2008). 1 New Zealand and Japan, also typically categorized as liberal, are not included; in the former case, no data are available from the LIS, and in the latter case, there is no variable identifying whether respondents are immigrants.
The LIS gathers cross-sectional data from household-based national surveys and harmonizes the data to ensure comparability, providing among the best cross-national data available for comparing incomes. For this analysis, the sample is limited to employed individuals aged 18–64 who are not enrolled as full-time students. The unweighted sample size ranges widely across countries. Ireland and Iceland have the smallest samples (at 3219 and 3734 individuals, respectively) and the United States has by far the largest sample size (at 46,875 individuals).
Following existing cross-national care comparisons (e.g. Budig and Misra, 2010; Duffy et al., 2013; Lightman, 2017), I identify individuals working for pay in the direct care industries of education, health and social work, relying on the LIS standardized industry variable. In an effort to apply the framework developed by Weeden (2002) within the confines of the data, my classification scheme distinguishes between high- and low-status care work based on occupation (see Lightman, 2017). Using the International Standard Classification of Occupations (ISCO-08), I differentiate between professionals and services and sales workers within the selected caring industries, as a proxy for care work jobs with higher and lower status and greater or lesser social closure. An individual must be identified as in both a caring occupation and a caring industry to be coded as having a care work job.
Notably, in all countries of analysis besides Iceland, immigrant care workers are disproportionately located in the health and social work industries, as opposed to in education (ranging from 65.0% of immigrants in Israel to 70.5% in Australia), likely due in part to language and cultural barriers. However, following the reverse trend, in Iceland roughly two-thirds (63.8%) of immigrant care workers are employed in education industries; this finding takes on added interest given that Iceland is the only liberal care regime included that has a comprehensive national childcare scheme (Datta Gupta et al., 2006).
Examples of high-status care workers within my classification scheme include doctors, university professors, primary and secondary school teachers, nurses and social workers, all occupations that are thought to have high levels of social closure. Examples of low-status care workers (and those thought to have lower levels of social closure) include childcare workers and teachers’ aides, healthcare assistants, psychiatric aides and first aid attendants (Barron and West, 2013). Notably, the LIS data cannot account for country-specific factors such as regional variation, the split between private and public institutions or the particularities of national legal frameworks regulating the organization of education, health and social work. Appendix 1 provides expanded details on my care work classification scheme.
Following Weeden (2002), I hypothesize that across the case study countries, caring occupations that have achieved lower levels of social closure (low-status care work) will be more likely to comprise disproportionate numbers of women, immigrants and individuals in non-standard (precarious) employment (Anderson and Hughes, 2010; Potter et al., 2006; Vosko et al., 2003) and that this will result in overall lower earnings than equivalent non-caring jobs. For those individuals in professional, high-status care work, I also expect to see a care penalty relative to professionals in non-caring occupations, but anticipate an under-representation of immigrant workers, in part due to the higher level of social closure within these occupations (Barron and West, 2013; Weeden, 2002). Building on Van Hooren’s (2012) analysis, I anticipate that immigrant workers will be overrepresented in low-status care work. I also hypothesize that immigrants will receive a pay penalty overall (controlling for work in care) and that there will be variation across the countries of analysis, tied to the degree to which each country’s labour market is regulated, and providing evidence of disparities within liberal welfare regimes (Ebbinghaus, 2012; Jensen and Lolle, 2013).
Variables of interest
The dependent variables in the descriptive and multivariate analyses examine who engages in what type of care work and capture wage variations in high- and low-status care employment, focusing on the employment and earnings of immigrant workers. 2
The main independent variables compare high-status and low-status care workers to individuals working in other high- or low-status non-caring occupations, as well as comparing immigrants (defined here as people who were born outside of the country) to individuals born in the country. 3 A major limitation of the LIS is that it does not include consistent measures of respondents’ race/ethnicity across datasets. In addition, the immigration variable does not account for variation in residency status or entry class, and individuals in the country without formal legal status are either excluded or under-sampled.
In order to specify any particular care penalty, as well as capture any added disadvantage experienced by immigrants, numerous conceptually relevant control variables are included in the final models. To account for the highly feminized nature of care work, a control for gender is included. Variables for family structure and demographic characteristics include a control for age, one for being married or cohabitating and one for living with one’s child aged 0–5 years. No variable is available to capture female-headed households, despite this being among the most meaningful predictors of low-income and precarious work (Gornick and Boeri, 2016).
The potentially mediating effect of human capital is captured using educational attainment, relying on a categorical variable harmonized across countries. This variable has three categories: low (lower secondary education and less), medium (upper secondary education through to vocational post-secondary education) and high (university/college education and above). In addition, to capture the effects of job characteristics on earnings, five dichotomous variables are included in the final models, where available: a control for part-time work status, one for employment in the public/non-profit sector, one for being self-employed, one for being a multiple job holder and one for having non-permanent employment.
By including these control variables, the goal is both to ensure that any relationship between low earnings and care work, as well as between low earnings and immigrant status, is not attributable to these factors and to explore if and how each of these variables is influential and if the effects vary within liberal welfare regimes.
Descriptive results
Descriptive analyses allow for examination of the state of the labour force in each liberal welfare regime as well as a profile of the high- and low-status care workforce. Although not reported in the tables, the data show that in each country, females make up just under half of the workforce. However, there are major differences in terms of the percentage of immigrants in the labour markets; Iceland has by far the lowest (at 6.6% immigrants), while Australia (28.1%), Israel (29.8%) and Canada (30.5%) have the largest proportion. Notably, the latter case, in part, is a reflection of the Canadian sample only including immigrants in major urban centres, where the foreign-born are overrepresented.
Despite considerable variation across countries in terms of worker and labour-force characteristics, each abides by the general principles of a market-oriented labour market. Each liberal welfare regime included has a heavy reliance on the private sector (which comprises 75% of workers at a minimum, across countries where data are available). Reflecting wage disparities within care work, in each country of analysis, high-status care workers make, on average, at least 1.9 times as much in mean earnings as low-status care workers. Israel presents the extreme, as high-status care workers, on average, make 3.9 times as much in mean earnings as low-status care workers.
Table 1 provides a descriptive profile of high-status care workers in the eight liberal welfare regimes. In all countries except Canada, professionals in health, social work and education have lower mean earnings, on average, than professionals working in non-caring industries. This earnings disadvantage ranges from −22 percent in the United States to −5 percent in Ireland. Professional care workers have a small earnings advantage (2%) compared to other professional workers in Canada, on average. This, in part, is a function of the high levels of social closure among doctors, nurses and teachers within the country (Lightman and Lightman, 2017).
Professional (high-status) care work in health, social work and education (%).
AU: Australia; CA: Canada; ISL: Iceland; IRL: Ireland; ISR: Israel; CH: Switzerland; UK: the United Kingdom; US: the United States. Population is limited to individuals aged 18–64, who are employed, not enrolled as full-time students and have earnings >$0.
Table 1 also demonstrates that the proportion of the total workforce employed in high-status care work ranges from 10.4 percent in Iceland to 5.4 percent in Israel, with females consistently overrepresented in these jobs. Notably, immigrant workers are underrepresented in high-status care employment in the majority of liberal welfare regimes (five of the eight countries). The United Kingdom and Ireland stand as exceptions to the trend, however, as immigrants are overrepresented in high-status care. This is likely due, in part, to lower access barriers for immigrants to work in these fields for English-speaking members of the European Union (Aluttis et al., 2014). Across the countries of analysis, high-status care work predominates in the public sector and has low levels of self-employment and non-permanent employment.
Table 2 shifts the focus to low-status care workers. In all countries examined, low-status care work comprises a smaller percentage of the overall workforce than high-status care work, ranging from 6.8 percent in the United Kingdom to 2.3 percent in Canada. In each liberal welfare regime, there is a financial disadvantage to working in low-status care as compared to in other low-status service and sales jobs (a ‘care penalty’, as seen in the majority of countries with high-status care work), but there is substantial variation in terms of the degree. In Switzerland and the United Kingdom, the financial disadvantage is lesser in low-status care than it is in high-status care (at −1% and −6%, respectively). However, low-status care workers fare worst in Israel and the United States, where they make on average only 67 percent and 70 percent as much in earnings as other low-status workers. In the latter case, this can be partially explained by the large number of low-skilled immigrants who are without legal papers in the country.
Service and sales worker (low-status) care work in health, social work and education (%).
AU: Australia; CA: Canada; ISL: Iceland; IRL: Ireland; ISR: Israel; CH: Switzerland; UK: the United Kingdom; US: the United States. Population is limited to individuals aged 18–64, who are employed, not enrolled as full-time students and have earnings >$0.
In terms of the demographic composition of the low-status care workforce, in all countries analysed, the percentage of females is higher in low-status care than in high-status care. Israel has the highest proportion of females in low-status care work (at 93.2% female), as well as the largest financial disadvantage for low-status care work relative to other service and sales workers. Immigrants are overrepresented in low-status care in the vast majority (seven out of the eight) liberal welfare regimes, most prominently in Switzerland and Israel (where they are approximately 47% overrepresented as compared to the total workforce). Interestingly, immigrants are underrepresented in low-status care in the United Kingdom (by 19.5%), suggesting, again, that care work in this country is attracting and employing more professional foreign-born populations. In terms of job characteristics, low-status care work has a lower proportion of individuals employed in the public/non-profit sector than high-status care work. In all countries examined, low-status care work also has substantially higher levels of part-time employment than high-status care work.
Overall, the descriptive results support the hypothesis that there are substantial differences between and within the high- and low-status care workforce in liberal welfare regimes in terms of both demographics and job characteristics. Without controlling for any other factors, in the vast majority of the countries analysed, there is a care penalty for both high- and low-status care, supporting suggestions of the social devaluation of work in health, social work and education (Budig and Misra, 2010; England et al., 2002). In addition, the low-status care workforce is more feminized, overall disproportionately comprised of immigrants and more likely to be situated in the private sector than the high-status care workforce, reinforcing suggestions of less social closure within these occupations (Barron and West, 2013). Thus, the descriptive data support a ‘migrant in the market’ model of employment, as suggested by Van Hooren (2012), with foreign-born care workers disproportionately located in the less regulated private sector where average wages are comparatively low.
Probabilities of performing high-and low-status care work
To identify who is engaged in high- and low-status care work, binary logistic regressions are run. Logistic regression allows for examination of immigrant and gendered variation in care work, after statistically adjusting for family structure and demographic characteristics, as well as human capital. For these models, the dependent variables are a dichotomous measure of high- or low-status care, coded as 1 for employment in high/low-status care work, with 0 denoting all other occupations (including the opposite category of high/low-status care). For convenience in interpreting the results, the predicted probabilities are presented for each population group from the mean (their chances out of 100 of working in high or low-status care).
Table 3 demonstrates that in six out of the eight liberal welfare regimes, immigrants are less likely to engage in high-status care work than comparable native-born populations, controlling for family structure, demographic characteristics and human capital. These results bolster suggestions of high levels of social closure within professional occupations in health, social work and education (Barron and West, 2013; Weeden, 2002). Notably, however, the United Kingdom and Ireland stand as outliers, with immigrants having a higher probability of working in high-status care than equivalent native-born populations, further suggesting that open borders in the European Union may have facilitated the flow of English-speaking professionals from poorer member countries.
AU: Australia; CA: Canada; ISL: Iceland; IRL: Ireland; ISR: Israel; CH: Switzerland; UK: the United Kingdom; US: the United States. Population is limited to individuals aged 18–64, who are employed, not enrolled as full-time students and have earnings >$0.
Logistic regression results control for demographic characteristics (age, marital status/cohabitation and the presence of young children in the household).
Divided by 100, the product is a probability.
Overall, immigrants have the lowest probability of being high-status care workers in Israel (at 1.2 chances out of 100) and the highest probability of working in high-status care in Iceland (at 3.5 chances out of 100). Table 3 also demonstrates that women have at least twice as high probabilities of being in high-status care work as men within all liberal welfare regimes except the United Kingdom, where the reverse pattern is observed. In addition, individuals with a high level of education have by far the highest probability of working in high-status care, reinforcing suggestions of education as a partial means to counter social closure (Weeden, 2002).
Table 4 examines the probability of working in low-status care and generally demonstrates opposite trends for immigrants within liberal welfare regimes compared to Table 3 (high-status care), reinforcing Van Hooren’s (2012) ‘migrant in the market’ hypothesis. Here, the data demonstrate that in six out of the eight countries examined, immigrants have a higher probability of working in low-status care than comparable native-born populations, although the magnitude of difference is minimal in certain cases.
AU: Australia; CA: Canada; ISL: Iceland; IRL: Ireland; ISR: Israel; CH: Switzerland; UK: the United Kingdom; US: the United States. Population is limited to individuals aged 18–64, who are employed, not enrolled as full-time students and have earnings >$0.
Logistic regression results control for demographic characteristics (age, marital status/cohabitation and the presence of young children in the household).
Divided by 100, the product is a probability.
Overall, immigrants have the highest probability of working in low-status care in Iceland (at 4.0 chances out of 100), followed by Ireland (3.2 chances out of 100). Only in Canada and the United Kingdom do immigrants have a lower probability of working in low-status care than the native-born, controlling for demographic characteristics and human capital. In the case of Canada, this suggests that immigrants are overrepresented in low-status care at the descriptive level due to their lower levels of education. Table 4 also demonstrates that in all cases, women have a considerably higher probability of working in low-status care than men (most notably 12 times higher in Ireland) and that compared to high-status care work, individuals in low-status care work are more likely to have a low or medium level of education.
Altogether, the logistic regressions demonstrate a trend for immigrants in liberal welfare regimes of a higher probability of employment in low-status care and a lower probability of employment in high-status care than equivalent native-born populations, controlling for gender, demographic factors and human capital. The following section provides the final empirical analyses, examining if there are wage penalties for high- and low-status care work compared to equivalent professional and service and sales jobs, as well as capturing any specific wage penalty for immigrants in these occupations, using Ordinary Least Squares regressions.
Care and immigrant wage penalties
Table 5 presents the log earnings for high- and low-status care work, as well as for immigrants, across each liberal welfare regime examined, controlling for individual and job characteristics. In Model 1, high-status care workers are compared to equivalent high-status professional workers in non-caring industries, controlling for immigrant status. In Model 2, low-status care workers are compared to equivalent low-status service and sales workers in non-caring occupations, again controlling for immigrant status. Each regression is measured as a function of gender, age, marital/cohabitation status, the presence of young children in the household, education level and job characteristics, where data are available. The dependent variable is the natural log of annual earnings (including wages and self-employment income) with extreme earnings recoded to the 1 percent and 90 percent values of within-country earnings distributions. Logged earnings have the benefit of normalizing the earnings distribution as well as allowing the transformed regression coefficients to be interpreted as approximate percentage change in earnings for a one-unit change in the independent variable (Budig and Misra, 2010).
Percentage point differences in annual earnings for high- and low-status care workers (relative to other high- and low-status workers) and for immigrants (relative to native-born) by country. a
Population is limited to employed individuals aged 18–64, who are not enrolled as full-time students and have earnings greater than $0.
Ordinary Least Squares regression results control for gender, family structure and demographic characteristics (age, marital status/cohabitation and the presence of young children in the household); human capital (education level) and job characteristics (part-time employment, public sector employment, self-employment, being a multiple job holder and non-permanent employment).
The Australia dataset does not include variables for public sector or non-permanent employment.
The Canada dataset does not include a variable for being a multiple job holder.
The Iceland dataset does not include a variable for public sector employment.
The U.S. dataset does not include a variable for non-permanent employment.
The Ireland dataset does not include a variable for non-permanent employment.
The Israel dataset does not include variables for public sector employment, self-employment or non-permanent employment.
The Switzerland dataset does not include a variable for public sector employment.
Significant effects (p < 0.05) are bolded.
Model 1 in Table 5 demonstrates a significant care penalty for high-status care work in five out of the eight countries examined. This care penalty is highest in Canada (at −22%), followed by the United Kingdom and Israel (−16%), Iceland (−15%) and the United States (−13%). In all countries with a significant wage penalty except the United Kingdom, Table 4 previously demonstrated a lower probability for immigrants of working in high-status care than comparable native-born populations. However, Model 1 in Table 5 demonstrates that only in Iceland is there is an additional significant wage penalty for immigrants working in professional occupations. Differently put, wages for high-status care in liberal welfare regimes are typically not different for immigrants than for equivalent native-born populations, yet their probability of employment in these higher paying jobs is substantially lower. This suggests that access barriers, perhaps due to foreign-acquired education or language skills, may be a major concern. Thus, overall, evidence of a high-status care penalty within liberal welfare regimes is strong, yet there remains substantial variation across liberal welfare regimes.
Model 2 in Table 5 focuses on low-status service and sales work in care. Here, the data demonstrate significant care penalties in four countries, with the biggest substantive effect in Ireland (at −30%), followed by the United States (−22%), Israel (−13%), and Iceland (−11%). Work in lower status care is often physically demanding, repetitive, lacking in security (precarious) and typically without union or bureaucratic guidelines for pay and promotion (Anderson and Hughes, 2010; Folbre, 2012). Building on suggestions of dualized or segmented labour markets (Boyd and Cao, 2009; Piore, 1979), the LIS data demonstrate that wages are both lower and less likely to vary substantially between caring and non-caring occupations in service and sales jobs, where immigrants are overrepresented. Switzerland, however, stands as an outlier, with a 9 percent ‘care bonus’ for service and sales workers in caring industries, supporting prior findings that due to recent welfare reforms and increased social spending, Switzerland may no longer classify as a ‘liberal’ regime (Obinger, 2010).
Finally, immigrants within service and sales work incur a wage penalty compared to equivalent native-born populations in four out of the eight liberal welfare regimes, namely Canada (at −21%), Israel (−15%) and Iceland and the United States (both at −10%). This suggests that in some liberal welfare regimes, it is immigrant status, rather than employment status, that has a negative impact on service and sales worker ‘migrants in the market’. It does, however, remain a perplexing finding, in that the United States and Israel are perhaps the two most extreme liberal economies examined, while Canada stands towards the other end of the spectrum. To better understand these findings, it may be necessary to probe each country individually to see the unique circumstances in each that lead to these wage penalties; a generalized explanation applying uniformly to all four countries is unlikely to be viable.
Thus, even after controlling for gender, human capital and job characteristics, low-status care workers are significantly disadvantaged in terms of earnings within four of the eight liberal welfare regimes, and immigrants in four out of eight countries incur a wage penalty compared to equivalent native-born service and sales workers. This provides partial support for prior findings of a low social value accorded to low-status care (Anderson, 2010; Lightman, 2017). In the case of migrant workers, the data demonstrate that in certain countries, market mechanisms that disadvantage low-status service and sales jobs also lead to the devaluation of foreign-born populations. In other countries, however, within low-status occupations, the findings demonstrate that any earning penalty associated with immigrant status is eliminated after controlling for human capital variables and work status, suggesting the mediating effects of these factors on earnings within low-status jobs (Boyd and Cao, 2009; Dekker and Van der Veen, 2017), and again demonstrating substantial variation within liberal welfare regimes.
Conclusion
This article attempts to expand and apply Weeden’s (2002) distinction between caring jobs with higher or lower social closure within an international comparative framework, distinguishing between high- and low-status caring occupations, as a proxy for jobs with greater and lesser social closure. Rather than assuming that all care work is highly feminized, poorly paid and precarious, the classification scheme developed and applied with the LIS data empirically measures differences between and within care employment in eight liberal welfare regimes. Ultimately, the data support the assertion that high- and low-status care work jobs are qualitatively different, with the latter having substantially lower mean wages, as well as being more likely to be part-time and situated in the private sector, as suggested by Van Hooren (2012). Future analyses would benefit from datasets that allow more detailed examination of the structure of particular occupations in health, social work and education, to allow for more comprehensive analyses of the extent and impact of social closure within particular care work jobs and in some cases, for in-depth study of specific country labour markets.
Migrant care workers, the focal population of this study, are found to have unique employment outcomes within liberal welfare regimes, and consistent trends are identified in terms of the types of paid care work they engage in. In six out of the eight liberal welfare regimes, immigrants are less likely to be employed in high-status care work than comparable native-born populations and more likely to be employed in low-status care work, controlling for family structure, demographic characteristics and human capital. Thus, the data support research suggesting that as a result of their limited employment options and, in some cases, vulnerability to deportation, immigrants are disproportionately located in lower status sectors of the paid care economy, working in jobs unattractive to native-born workers due to their low prestige and unpleasant working conditions (Anderson, 2010; Duffy et al., 2013). In the case of high-status care, challenges related to accreditation and language and high levels of social closure in many health, education and social work professional occupations may account for immigrants’ lower probabilities of employment in many of the countries analysed.
The results on earnings in care work are more varied. In the majority of countries examined (Canada, the United Kingdom, Iceland, Israel and the United States), multivariate analyses demonstrate a significant wage penalty for professional care workers compared to equivalent professional workers in non-caring occupations. This presents a pattern of financial devaluation of high-status jobs in the highly feminized industries of health, education and social work across liberal welfare regimes. However, in the case of low-status care work, only four of the eight countries examined demonstrate a significant wage penalty compared to equivalent non-caring service and sales jobs. This suggests not only substantial variation across liberal welfare regimes but also that within segmented labour markets, it may be immigrant status, rather than work in care, that comparatively disadvantages low-status service and sales worker ‘migrants in the market’. Immigrants incur a significant wage penalty in professional jobs in Iceland, as well as a significant and substantive wage penalty in four of the eight liberal welfare regimes in service and sales jobs, controlling for work in care, human capital and individual and job characteristics. Supporting these findings, descriptive analyses demonstrate that without any relevant controls applied, there exists a care penalty in both high- and low-status care work jobs in the vast majority of liberal welfare regimes.
Limitations to the LIS data suggest that the current results may underestimate disparities experienced by marginalized immigrant groups working in care. Future research would benefit from the use of data – where such exist – which allow for an examination of interactions between immigrant status and racialized minority status, as well as the impact of variations in residency status or entry class for foreign-born care workers in liberal welfare regimes. Immigrants in irregular situations, for example, may be vulnerable to deportation and barred from numerous types of high-status care, and thus their employment outcomes can be expected to be worse than those for immigrants who are in the country legally. In addition, future analyses would benefit from comparisons across welfare regimes to examine the degree to which global trends towards precarious work and migration, and market-based solutions to social welfare, have led to either convergence or differentiation in the stratification of employment and earnings between migrant and native-born care workers in a variety of institutional contexts.
Overall, Van Hooren’s (2012) suggestion of a ‘migrant in the market’ model of employment applies for the majority of liberal welfare regimes examined, as migrants are disproportionately located in lower wage, lower status care jobs with poorer working conditions. However, due to variation in both the employment and earnings of migrant care workers in high- and low-status care, the data also reinforce the need for more nuanced typologies to classify care work, as well as to account for disparities in the national and regional structure of health, social work and education.
Ultimately, the findings from this article are meaningful in the current policy context. Growing government austerity and earnings polarization across liberal welfare regimes suggest that ‘migrants in the market’ are overall not faring well: they are disproportionately located in low-status, precarious and lower paid forms of care work, and in some cases, experience an additional pay penalty for being foreign-born net of their personal characteristics and human capital. Thus, the data reinforce the need to further specify what is meant by ‘liberal’ and by ‘care work’ and to further consider who is providing which type of care and what their employment outcomes are. The data suggest that resources, employment protections and other policy levers ought to be directed towards (illegalized) (female) immigrants working in lower status caring occupations, as well as considering how global shifts towards market-oriented care regimes disadvantage vulnerable workers and facilitate precarious working conditions in the realm of care and beyond.
Footnotes
Appendix 1
Details of care work classification scheme, Luxembourg Income Study, 2010–2013.
| Country, sample size, dataset | Care industries and occupations (derived from the International Standard Classification of Occupations (ISCO-08), and the LIS standardized industry variable) | |
|---|---|---|
| Australia (AU) (N = 16,770) 2010 Household Expenditure Survey and Survey of Income and Housing
Canada (CA) (N = 26,310) 2010 Survey of Labour and Income Dynamics Iceland (ISL) (N = 3734) 2010 Survey on Income and Living Conditions Ireland (IRL) (N = 3219) 2010 Survey of Income and Living Conditions Israel (ISR) (N = 10,416) 2012 Household Expenditure Survey Switzerland (CH) (N = 7193) 2013 Survey on Income and Living Conditions United Kingdom (UK) (N = 18,919) 2013 Family Resources Survey United States (US) (N = 46,875) 2013 Current Population Survey – ASEC (Annual Social and Economic Supplement) |
Professional (high-status) jobs in care work | |
| Education
University and higher education teachers; vocational education teachers; secondary education teachers; primary school and early childhood teachers; other teaching professionals (e.g. language teachers, special needs teachers, information technology trainers) |
Health and social work
Medical doctors – general and specialist; nursing and midwifery professionals; traditional and complementary medicine professionals; paramedical practitioners; veterinarians, other health professionals (e.g. dentists, pharmacists, physiotherapists, dieticians, speech therapists); social workers |
|
| Service and Sales (low-status) jobs in care work | ||
| Education
Childcare workers; babysitters; nannies; teachers’ aides |
Health and social work
Personal care workers in health services; healthcare assistants; birth assistants; psychiatric aides; home-based personal care workers; dental aides; hospital orderlies; pharmacy aides; first aid attendants |
|
Source: Adapted from Lightman (2017).
Population is limited to individuals aged 18–64, who are employed and not enrolled as full-time students.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
