4Skill Demands of Service Work

Publication Details

Introducing the session, moderator Peter Cappelli noted that, because researchers lack good measures of skills, it is difficult to assess changes in skill demands across the economy as a whole. He suggested that the approach taken in this session—examining skill in the context of a particular industry or occupation—would help to enhance understanding of “what skill means.” Finally, he suggested that the upcoming discussions would indicate, as some of the previous sessions already had, that “there are many choices as to how these jobs are organized.”


Mary Gatta (Rutgers University) opened her presentation by emphasizing that many service occupations cannot be moved offshore (Gatta, Boushey, and Appelbaum, 2007a; Blinder, 2007). She said “face-to-face in-person jobs … need to be performed here in direct sight,” and noted that the Bureau of Labor Statistics (BLS) projects that the service occupations cluster1 will grow by 19 percent between 2004 and 2014, much faster than the projected 13 percent growth rate for all occupations (Hecker, 2005). A total of 11 of the 20 occupations that BLS projects will grow most quickly to the year 2014 are in the service occupations job cluster, and all 11 have either very low or low annual earnings, less than $28,000 per year (Hecker, 2005; see Table 4-1).

TABLE 4-1 Occupations with the Largest Projected Job Growth, 2004-2014.


TABLE 4-1 Occupations with the Largest Projected Job Growth, 2004-2014.

Gatta explained that, because many service occupations require little formal education and pay low wages, economic studies often define them a priori as unskilled. In addition, she argued that the Dictionary of Occupational Titles “does not adequately measure or capture the skills in the service occupations” (Attewell, 1990; National Research Council, 1999). To avoid these weaknesses in national datasets, Gatta said, she and her coauthors turned to the case study and ethnographic literature to try to understand the skill demands of service occupations (Gatta, Boushey, and Appelbaum, 2007b). She cautioned that qualitative ethnographic studies provide rich data but use purposive rather than probability samples and investigate small samples of job holders (Sandelowski, 1995).

Gatta noted that a key feature of service work lies in its interaction between the service worker and the customer. She said that some researchers have argued that this interaction can be scripted and rationalized, so that “people are … in essence robots going through their work day” (e.g., Ritzer, 1996). Challenging this argument, Gatta described several studies that have reached quite different conclusions about the nature of service work. First, a seminal study of flight attendants (Hochschild, 1983) found that their jobs included not only serving the customer, but also smiling and conveying a bubbly personality—a form of “emotional labor” that is closely linked to the product being sold. Second, she noted that, although McDonald’s Corporation tries to script and control the interactions of both workers and customers, the company also employs large numbers of supervisors to help counter workers respond to unusual and unpredictable customer demands (Leidner, 1993). Third, Gatta described studies indicating that service workers perform “articulation work,” deploying a blend of emotional, cognitive, technical, and time management skills to quickly solve customer problems (Hampson and Junor, 2005).

In her own ethnographic study (Gatta, 2002), Gatta said, she found that restaurant servers were required to engage and build rapport with multiple customers seated at multiple tables, use micro computers to capture food orders and process credit card payments, and steer customers toward higher priced items without appearing to be manipulative. Another study (Newman, 1999) found that fast food workers were required to listen carefully to customers and communicate effectively with them, send out a rapid stream of instructions to coworkers, check the orders, receive money, and make change.

Gatta said that recent research also illuminates the need for aesthetic skills in many service occupations, including hospitality industry employees, who must have the right appearance and personality, and sales assistants, who must present a fashionable appearance. Luxury hotel workers must be familiar with local night life, theater, and restaurants at different price points and apply observational and active listening skills to discern a guest’s needs and to personalize and customize the guest’s experience.

In addition to other skills, service occupations often require caring skills, including nurturing, communicating, and making the client or customer happy, Gatta said. Because service occupations are often filled mostly by women and these caring skills are viewed as natural feminine qualities, the skills are neither recognized nor rewarded (England, Budig, and Folbe, 2002). At the same time, however, the jobs are divided by race, as white women interact with customers in the front of the restaurant, hotel, or other service location, and women of color work in less visible jobs as maids and dishwashers, in which wages are even lower.

Gatta said that the theory of skill-biased technological change suggests that the introduction of computers increases demand for (and wages of) highly skilled workers. However, although computers have been integrated into many types of service work, these workers’ wages have not risen. Referring to the finding by Autor, Levy, and Murnane (2003a) that computers complement workers in nonroutine cognitive tasks, Gatta suggested that the theory does not apply well to service workers because their nonroutine social skills and caring skills are poorly defined and are not seen as a complement to computer technology. In fact, she said, one study found that in jobs requiring high cognitive skills, social and interactive tasks were considered special skills and rewarded with higher pay, but in jobs demanding lower levels of cognitive skills, social and emotional tasks were seen as making the job more interesting and attractive to employees without the need for additional pay (Glomb, Kammeyer-Mueller, and Rotundo, 2004).

In closing, Gatta noted that the majority of those who will make up the U.S. workforce over the next 20 years are already at work (National Center on Education and the Economy, 2006) and called for addressing the workforce “literacy crisis.” She noted that workers displaced from manufacturing have not been trained in the social and interpersonal skills needed for the rapidly growing service occupations. Finally, she called for further research to more clearly define the social and interactive skills—both to guide retraining programs and so that these skills might be better rewarded in the labor market.


David Autor commented on the three main arguments Gatta presented (Autor, 2007b). The first argument was that service occupations are “here to stay.” Autor agreed with this argument, stating that service occupations are difficult to automate or move offshore. At the same time, he noted that these jobs do not require extensive formal education because they use “natural” human skills, such as visual recognition and fine motor coordination and spoken language. He presented an analysis of BLS occupational data showing that over the period 1980 to 2000, service occupations were the only low-skilled occupational group that grew in numbers.

The second argument was that service jobs require adaptive skills, emotional labor, and other skills that are not adequately rewarded. Autor agreed that service work is important and difficult, but, he argued, it is paid little because it is not scarce. He suggested that people might pay a high premium for the first hour of service in a restaurant or in house cleaning, but because many workers are willing to provide these services, the “equilibrium wage” is much lower.

Autor then addressed the third proposition in Gatta’s paper—that the gender and racial composition of people employed in service occupations contributes to their low wages. He presented his own analysis of BLS data indicating that there is not a clear relationship between the extent of female employment and wages in various service occupations. He argued that the service occupations cluster, with 51 percent female employment, is not as female dominated as three other occupational groups—technicians, sales, and administrative occupations—in which employment is 59 percent female yet wages are relatively high. However, Autor emphasized that, in comparison to other occupational clusters, the service occupations cluster employs more high school dropouts and more minorities.

While stating that he agreed with the paper authors that it is important to think about what can be done to improve the quality of service jobs, Autor also noted that policy options related to improve these jobs all have “downsides.” One option is to simply accept a world “of many McJobs,” with low wages and few benefits, he said. Another is to expand government subsidies for such important services as health care, child care, and education, but such subsidies would be vulnerable to changing political winds. Another possibility would be to regulate the employment of service workers through such approaches as government licensing of the occupations or by mandating higher wages or training levels, but these steps would be likely to reduce employment in these occupations and make it difficult to obtain services (Kleiner, 2006). Finally, he said it is a fallacy to believe that wages of service occupations can be set at a high level because the jobs are not easily moved offshore. He said that services will continue to compete for other uses of customers’ money, and if services become too expensive, customers will demand less of them.


Chris Wellin (Miami University of Ohio) opened with a quotation, explaining that, although written to describe children’s care for elderly parents, it was equally applicable to paid care work:

Care involves the constant tension between attachment and loss, pleasing and caring, seeking to preserve an older person’s dignity and exerting unaccustomed authority, overcoming resistance to care and fulfilling extravagant demands, reviving a relationship and transforming it (Abel, 1990, pp. 204-205).

Perhaps the best way to understand care work, Wellin said, is to recognize that these workers act as an extension of family members. He said that the quotation was important because it not only focuses on instrumental tasks, such as bathing and dressing, but also recognizes “the interpersonal and emotional relations in which the tasks are embedded”(Wellin, 2007a).

Wellin described his perspective on care work as ethnographic—one that recognizes the cultural and organizational context of the work (Wellin, 2007b). Historically, he said, these jobs “were defined by their place at the bottom of a status hierarchy” in hospitals. Although they now work in nursing homes and in the community as home health workers, where they no longer play a marginal role on the caregiving team, he said that these earlier views of their work still “cling to the workers.”

Briefly describing the demographic trends driving demand for low-wage caregiving work, Wellin noted that, by the year 2025, more than 20 percent of the U.S. population will be over the age of 60, with the fastest growth among those over 80 “for whom disability rates are really high.” Given the natural limits on the possibilities for medical intervention to improve the health of older, disabled people, care focuses on making them comfortable, both physically and mentally. As a result of these trends, demand is growing for caregivers. Two groups of caregiving occupations—nursing aides, orderlies and attendants; and personal and home care aides—are among the ten occupational groups with the largest projected job growth over the decade from 2004 to 2014 (Hecker, 2005).

In 2004, nearly 3 million low-wage care workers were employed in the United States, including 1.45 million nursing aides, orderlies and attendants; 701,000 personal and home care aides; and 624,000 home health aides (Hecker, 2005). More than 90 percent of care workers are women (Eaton, 1999). Over one-third are African-American, 67 percent have a high school diploma or less education, 25 percent have at least one child aged 5 or younger, and one-third have family incomes below 150 percent of the poverty level (Potter, Churilla, and Smith, 2006). Hourly wages are low, ranging from less than $7.00 per hour to over $8.00 per hour for a certified nursing assistant working in a hospital, and job turnover is high.

Wellin explained that care jobs are growing much faster in community-based settings (home care or assisted living) than in hospitals or nursing homes, both because of efforts to contain rising public expenditures for care and because federal law requires provision of care for the disabled in the least restrictive setting possible. In these settings, caregivers work for “clients,” rather than caring for “patients” in hospitals or “residents” in nursing homes. Because clients demand more personalized, tailored care, the worker has greater freedom and an opportunity to provide “total care.” At the same time, however, this is where wages are the lowest and the worker lacks any support from colleagues.

The work is complex because it requires intimate contact—“encroachment on adult privacy and dignity”—with individuals who vary widely in their preferences, abilities, and disabilities, Wellin said. He argued that, even more than the other types of interactive service work that Gatta described, low-wage care work is viewed as natural, instinctual work done by women and their families. In fact, a worker who behaves in a more “professional” manner, such as taking too much time to attend to an individual patient’s needs or trying to communicate about the patient with an attending physician, is likely to be penalized or fired.

In institutional settings, Wellin said, low-wage care workers provide over 80 percent of direct care and “nearly all of the social contact and support.” Typically, both caregivers and care recipients describe their relationship as if they are family members. For care recipients, sharing their life story is important to maintain a sense of identity, and the caregiver plays an important role in this.

Retaining low-wage care workers is often difficult, Wellin said, citing a survey that found that only half of certified nursing assistants were still in that occupation two years after certification. He noted that research has identified several factors associated with higher retention, including a lower number of beds, a lower ratio of staff to residents, nonprofit status, a higher quality of care, and involvement of staff in training and care planning. These factors lower turnover, even when wages remain low, he said. In addition, cultural and ethnic similarity enhances recruitment and retention by providing a supportive community, Wellin said, citing as an example the fact that many Filipinos work as home health aides in the San Francisco area.

Training of Low-Wage Care Workers

Wellin explained that current training of low-wage care workers stresses instrumental tasks, basic medical knowledge, resident safety and rules, with little attention to “psychosocial care skills.” Training and certification requirements vary widely across the states. Some care workers receive inservice training in basic nutrition, infection control, and safe transfer techniques to avoid occupational injuries that are among the highest of any job in the national labor market. Less often, they are trained in resident rights and ethical concerns, dementia care, family stress issues, communication, and spiritual counseling (Wellin, 2007b).

Training is important because many workers view their jobs as a “longer term career choice,” Wellin said. He noted that the occupation is an “anomaly” combining “high commitment, a high sense of mission, with low occupational stability and rewards.” He reminded the audience that the shift to community-based care increases not only the technical-medical, but also the psychosocial skill demands of the work, and that the growing population of elderly people with more serious disabilities, including dementias, also increases the complexity of the work.

Wellin argued that training to meet these growing skill demands must be “intertwined,” incorporating both low-level medical training and training in psychosocial skills. Workers should be trained in basic human development and adult aging and in communication skills, recognizing the care worker’s key role in extending the preferences of the individual in a “very vulnerable situation.” Finally, he strongly supported including low-wage care workers in ongoing assessment and modification of plans for the care of those they work with as “perhaps the most powerful way to apply their knowledge.”

The delivery of this training should be integrated with the particular needs of the work setting, Wellin suggested. For example, as nursing homes and facilities are increasingly licensed to meet special needs, such as dementia, workers could be trained and certified in skills related to these needs. Portable certification would be valuable for care workers moving from one job to another. In closing, Wellin argued that the fates of the higher wage knowledge workers discussed in the earlier session and the predominantly female care workers he described are “intertwined.” The quality of life of the professional or knowledge worker who is working “intense hours and long hours” may be defined by the continuity and quality of the child care and elder care provided to family members (Wellin, 2007b).


Peter Kemper (Pennsylvania State University) opened by explaining that he has a rather different perspective, as an economist interested in long-term care policy and currently engaged in an evaluation of a program designed to improve the quality of care and the jobs of care workers. First, he said he agrees with Wellin that direct care requires many skills. Kemper said that Wellin’s argument that caregivers’ tasks “require complex, empathic and embodied practices” (Wellin, 2007b, p. 17) is “right on target” and consistent with the knowledge and adaptive skills discussed in the knowledge worker session, “making this a high-skill job in some important respects.”

Second, Kemper agreed that the training of care workers should recognize their important role not only in health care but also in enhancing quality of life by providing psychosocial care. In addition, he said, Wellin’s argument that care work should be personalized should also inform education and training policies.

Third, Kemper commented on the need for “a living wage for such workers, and to attach employment to basic fringe benefits” (Wellin, 2007b, p. 41). Although the work is valuable, he said, raising wages would be difficult because the demand is “derived.” Although the growing elderly population is a large force driving demand, Medicaid is the primary payer for care and policy keeps Medicaid reimbursement low. Kemper said that, as the private pay market grows, the “two-tier market … will get worse as time goes on” and that wages will rise “at some level.” He suggested thinking about wages and quality of care together, citing the example of the Better Jobs Better Care demonstration program (Institute for the Future of Aging Services, 2007), which is based on the idea that improving the jobs of direct care workers will improve quality of care and care recipients’ quality of life.

Kemper offered several observations about education and training. First, improving the training of care workers and their supervisors could play a role in improving job quality, which would help attract workers to care jobs and help them advance their careers. Second, he suggested that society as a whole needs “a broad education on disability,” as many people lack awareness about what it means to live with disability. This would be similar to education about awareness of, and respect for, other cultures and education about diversity. Kemper suggested that public education about disability should include a component emphasizing the “rewards and possibilities” in caregiving careers.

Kemper said that many people view care work as a dead-end career, because many direct care workers are not interested in moving up to nursing jobs. However, there are specialized positions, such as those of medication aides and dementia aides, and government policies can support certification of these specialties. In addition, there are opportunities to take on some supervisory responsibilities as a peer mentor or team leader. He said that the Better Jobs Better Care demonstration is providing training in these areas and in cultural competence and person-centered care. The program provides modest amounts of training that are compatible with full-time work.

Kemper said that training of managers and supervisors is critical, as care workers “don’t leave their job; they leave their supervisors.” These supervisors tend to be registered nurses, whose education does not develop supervisory skills. Finally, Kemper concluded that, because few caregiving skills are employer-specific and job turnover is high, employers lack incentive to train,2 so public policies are needed to support training.


In response to Cappelli’s question about whether skill demands of service jobs are rising, Wellin reiterated that care workers have moved out of hospitals—where they were part of a large team including nurses, doctors, and perhaps a social worker or member of the clergy—and into nursing homes, assisted living or home health care, where they are the primary caregivers. This shift greatly expanded their skill demands, he said.

Eileen Appelbaum said that the skill demands of service jobs will depend on “what we want to say about the quality of services.” For example, the New Jersey Supreme Court ruled that children ages 3 to 6 are entitled to a free quality education, and that providing this level of quality would require more highly skilled early childhood teachers. To implement this ruling, Appelbaum said, the state has established a new teacher certification program for lead teachers in early childhood classrooms, and assistant teachers are obtaining associate degrees through an apprenticeship program that includes on-the-job training with classroom training. Speaking to the audience, Appelbaum asked whether, if they had a young child or grandchild of 2 or 3, they would want that child to interact daily primarily with a person who cannot speak English. “If we are talking about having quality of services,” she said, “then we need to be thinking about the quality of training.” In addition, she said, as more organizations reduce layers of supervision, employees more often are assigned work through an e-mail message or other written instructions. Arguing that “most of these jobs that we consider low-skill now require literacy,” Appelbaum expressed grave concern about current national literacy levels (Kutner, Greenberg, and Baer, 2006). Because many adults with “basic” or “below” basic literacy levels are at work and have family responsibilities, she suggested thinking about how to use technology “to make literacy training flexible in time and space.”

To summarize Appelbaum’s argument, Cappelli suggested that service work could be performed at very different levels and that, if society decided that the services should be provided at a higher level of quality, then the skill demands would rise as well. Appelbaum responded: “absolutely.” Gatta then mentioned that the hospitality industry in New Jersey is partnering with community colleges to develop a system of skills certification, training, and job ladders for casino workers, with the goal of reducing turnover and increasing the skills and quality of these jobs.

Paul Osterman commented on Autor’s statement, in his response to Mary Gatta, that every option for improving the quality of service job has a “downside.” Osterman asserted that Autor views any option that might interfere with labor supply and demand as an “imperfection” in the labor market. Osterman pointed out that both he and Autor enjoy tenured positions, and this kind of imperfection “seems like the natural state of affairs,” yet, when workers organize a union to protect their interests, economists tend to view this only as an imperfection. Osterman reminded Autor of his earlier study (Autor, Levy, and Murnane, 2003a) about a bank that introduced technology in two different departments. One department introduced technology in a way that created very routine, fast-paced jobs, and the other department integrated tasks, creating jobs that required higher levels of abstract reasoning and were more productive, and raised the salaries accordingly. Osterman said the study revealed that employer choices about how to organize work are influenced not only by constraints (or “imperfections”) in the labor market, but also by values. He suggested being open to the possibility of organizing work in ways that provide gains for both employees and employers yet do not interfere with supply and demand forces (e.g., Osterman, 2000).

Autor responded that simply “exhorting people to see” that low-wage service work involves many unrecognized, valuable skills would not affect their wages. He argued that he does not view all alternatives for improving the quality of service jobs as imperfections. For example, he said, he views training not as a “distortion” in the labor market, but rather as an investment. Autor went on to say he thinks that the U.S. labor market is “operating relatively well, given the availability of skills and supply and demand,” but “that doesn’t mean there aren’t ways to push it in the direction that all of us would be happy about.” Cappelli responded that there were debates in the early 1990s about how to encourage firms to organize work in ways that enhanced job quality, skills, and wages (e.g., Commission on the Skills of the American Workforce, 1990), but there were never clear answers.

Appelbaum responded by reminding the group that Eric Wanner had talked earlier in the day about the important role labor market institutions play in constraining employers’ choices. She said that, if the federal government required higher minimum wages, employers might choose to invest more in training to increase the productivity of these more expensive employees. She again suggested focusing on the quality of the service, arguing that, if society decides that children, old people, disabled people, and other vulnerable groups must receive high-quality service, then the service workers would require higher skills. Wellin added that the current backlash against managed care and baby boomers’ more aggressive demands on the health care system might raise public awareness of the value of direct care work and workers. Sam Leiken asked what mechanisms are in place—such as consumer satisfaction surveys—to judge the quality of elder care services. Kemper responded that measures of the quality of physical and medical care are available, but it is much more difficult to measure the quality of psychosocial care. He said that some early efforts are under way to assess the extent to which different nursing homes have moved toward person-centered care.

Autor interjected that immigration policy is an “incredibly central” aspect of the current structure of service occupations. He said one reason these jobs are “low wage and low skill” is that there is an abundant supply of people with little formal education “who are willing to take those jobs at those wages.” Although low-wage personal service jobs cannot be moved to India, he said, it is possible to bring in labor from the rest of the world to do these jobs.

Mary McCain, (Techvision 21 Consulting Group), said that the “exciting” discussion of future skills and competencies should be grounded in the realities of a public and private education and training system that is “at best possibly dysfunctional.” Based on her experience working with companies, she noted that employers find it extremely difficult to find and pay for training, as well as to find employees with enough basic education and language ability to learn. She echoed Appelbaum’s concern about the large fraction of the national workforce with basic or below basic literacy levels.

Gatta commented that some employers in the health care and hospitality industries have moved toward integrating basic literacy with occupational training, working with public education institutions to create the curriculum. Gatta said that it is a particular challenge to create training, career ladders, and certification systems that can help low-wage, entry-level workers advance their skills and careers.

Suzanne Ripley, an expert on disability (Academy for Educational Development), warned that it is important to distinguish between elder care provided in the home and care provided in institutional settings. She said that, in Virginia, an individual can hire his or her own caregiver, with only a minimal background check, and that the caregiver’s work is not supervised or monitored. Ripley expressed concern that the panelists had not mentioned the problem that direct care workers’ lack benefits, describing this as “an overwhelming difficulty in recruiting or retaining staff,” especially because so many women with children work in these jobs. In addition, Ripley said, Medicaid, which pays for these workers, is a form of welfare, and such jobs “don’t pay well.” She asked how to direct public policy and public attention to the need to provide benefits and a living wage to direct care workers.

Cappelli concluded the session by noting that its value “to help us understand the various choices that are made that shape skill requirements and skill demands.” He specifically noted the importance of choices about the level of service that is provided and employer choices about how to organize and reward work.



See Appendix B for a list of the groups of jobs included in the service-occupations cluster.


Kemper was referring to human capital theory (Becker, 1964), which suggests that employers will invest in training for skills specific to the firm, but they will not pay for general training to develop skills that could be used in another firm.