View the Data Supplement to this report.
NO SHORTAGE OF “NONSTANDARD” JOBS
Nearly 30% of workers employed in part-time, temping, and other alternative arrangements
by Ken Hudson
Even in the midst of a booming economy, with wages rising and unemployment at historic lows, reliance by the U.S. economy on nonstandard jobs – part-time work, independent contracting, temping, on-call work, day labor, and self-employment – remains as strong as ever. Employers argue that these jobs (also called contingent work) provide the flexibility needed to be competitive. But these perceived advantages notwithstanding, as of 1997, most nonstandard workers, on average, were paid less, were less likely to receive health insurance or a pension, and had less job security than workers in regular full-time jobs. The disparities between nonstandard and regular full-time jobs persist even when comparing workers with similar personal, educational, and job characteristics.
Some types of nonstandard work indeed pay high wages, but even these arrangements are usually deficient with respect to fringe benefits and job security. Compared to regular employment, these jobs are also characterized by a higher degree of wage variability among workers. In other words, while some nonstandard workers may be highly paid, other nonstandard workers in the same type of arrangement receive significantly lower wages, and these pay differentials are greater than the differentials for regular full-time workers. But the most common types of nonstandard work arrangements – such as temping and part-time work – are, on average, inferior in all respects to regular full-time jobs. Moreover, most nonstandard workers are employed in the worst kinds of nonstandard jobs.
This report examines the Contingent Work Supplement to the Current Population Survey (CPS), compiled by the Bureau of Labor Statistics in 1997. This survey was first fielded in 1995 and provided the first nationally representative examination of contingent and nonstandard work. Those data were examined in the Economic Policy Institute studies Nonstandard Work, Substandard Jobs (1997) and Managing Work and Family (1997). The CPS collected this data again in 1997 and 1999. This report examines the 1997 data and updates much of the analyses presented in Nonstandard Work, Substandard Jobs. (Additional information and updates to the tables from Nonstandard Work, Substandard Jobs not explicitly referenced here can be found at the Economic Policy Institute Web site at www.epinet.org.)
Between 1995 and 1997, the prevalence and the quality of nonstandard work changed little. Moreover, in the areas in which the quality of nonstandard jobs has improved, these improvements have been far less evident for women than for men. Women and minorities continue to be disproportionately employed in nonstandard jobs with low wages and no benefits, while white men are over-represented in nonstandard work arrangements with the best wages and benefits.
Among the specific findings from the CPS data are the following:
- In 1997, 28.7% of American workers – about one out of three women (33.7%) and one out of four men (24.3%) – worked in nonstandard jobs, down slightly from 29.4% in 1995.
- Workers in all nonstandard work arrangements are more likely than workers in regular jobs to receive low- and poverty-level wages.
- Most types of nonstandard workers average lower wages than do full-time workers with similar personal characteristics, including education.
- Among nonstandard workers, just 13.6% of women and 11.6% of men receive health insurance coverage from their employers. Only 15.7% of female workers and 9.3% of male workers in nonstandard arrangements receive pension coverage.
- Compared to workers with similar personal characteristics in regular full-time jobs, the hourly wage penalties for women in nonstandard jobs are 20% for regular part-time workers, 18% for temps, 26% for the self-employed, 20% for on-call workers, and 7% for independent contractors. For men, the wage penalties amount to 27% for regular part-time workers, 15% for temps, 11% for the self-employed, 10% for on-call workers, and 1% for independent contractors.
While it is true that many workers prefer the flexibility provided by some kinds of nonstandard jobs, large numbers of workers feel compelled to accept these arrangements for economic and personal reasons beyond their control. Unfortunately, given current labor market policies, nonstandard employment has the potential to become a mechanism for providing substandard wages and benefits. Responsible public policies should endeavor to ensure that workers are not penalized in terms of pay and benefits because of their work arrangements.
Prevalence of nonstandard work
Nonstandard work refers to any work arrangement other than employment in a full-time wage and salary job with a standard employer-employee relationship. (For a more rigorous definition, see the appendix.) Seven types of nonstandard workers are described here: part-time workers, employees of temporary help agencies (temps), contract company workers, independent contractors (including both those who are self-employed and those who report they are wage and salary workers),1 the self-employed, on-call workers, and day laborers.2 Part-time workers are defined as wage and salary workers who are employed fewer than 35 hours per week and who are not employed in another type of nonstandard work arrangement.3 Regular full-time (or standard) workers are employed in wage and salary jobs of 35 hours or more per week.
In 1997, 28.7% of American workers were employed in nonstandard jobs ( see Table 1). This is a slight decline since 1995, when the share was 29.4%. In 1997, about one out of three women (33.7%) and one out of four men (24.3%) worked in nonstandard jobs. The largest change over this period in the distribution of workers across work arrangements occurred among the self-employed; the percentage of workers in this group fell from 5.5% in 1995 to 4.8% in 1997. Overall, there was almost no change between 1995 and 1997 in the distribution of workers across work arrangements.
Job quality is reflected in a number of job characteristics, including wages, fringe benefits, and job security, all of which seem to be strongly associated; jobs that have low wages are also more likely to lack health insurance, retirement benefits, and employment security (Hudson 1998). In 1997, as in 1995, nonstandard jobs were more likely than regular full-time jobs to pay low wages and less likely to provide either health or pension benefits. Nonstandard jobs were also more likely than regular jobs to be of limited duration. The quality of nonstandard jobs did not improve substantially between 1995 and 1997.
Average hourly wages, by work arrangement and sex, are shown in Table 2. The averages are unadjusted for differences in education, experience, and other factors that affect wages. Part-time workers, temps, on-call workers, and women who are self-employed have lo
wer average wages than regular full-time workers of the same sex; all other nonstandard jobs pay higher average wages than regular jobs. For example, the average wages of men who work part-time are nearly $6 per hour less than men working in regular full-time jobs. For women, the pay difference is over $3 per hour – less than the difference for men but still substantial. A similar wage difference exists for female temps – their earnings, on average, are also more than $3 per hour less than those of regular full-time female workers.
Table 3 shows that workers in all nonstandard work arrangements are more likely than regular workers to receive low- and poverty-level wages. Among both men and women, nonstandard workers were two to three times as likely to have low hourly wages (defined as a wage in the lower 20% of the wage scale for each sex) in comparison to regular full-time workers. Nonstandard workers were also more than twice as likely as regular full-time workers to receive poverty-level hourly wages (a wage too low to lift a family of four above poverty, working full time for a full year). The percentage of jobs in each nonstandard work arrangement receiving high wages (defined to be a wage in the top 20% of the wage distribution for that particular sex) reveals the same pattern that is evident in the average wage data in Table 2. Taken together, these data reveal some important characteristics about the distribution of wages in nonstandard work arrangements. Compared to regular full-time workers, temps, on-call, and part-time workers of either gender, as well as self-employed women, have a wage distribution that is extremely skewed toward the low end of the wage scale. Other nonstandard workers, however, are, when compared to regular workers, over-represented at both the low and high end of the wage scale. This indicates that, compared to regular full-time workers, nonstandard workers are more concentrated at both the top and the bottom of the wage scale, with fewer in the middle. In other words, these wage distributions have a high degree of wage inequality when compared to regular full-time jobs.
Another important consideration is the effect of work arrangement on wages for workers with similar characteristics. Table 4A shows that most types of nonstandard workers receive lower wages than full-time workers with similar personal characteristics, including education. (Data for 1995 are included in Table 4B for comparison.) For example, regular part-time employees are paid substantially less per hour-reductions average 27% for men and 20% for women-than are regular full-time workers with similar characteristics. Temps average 15-18% less per hour, and on-call workers get 10-20% less. These pay penalties have changed little since 1995. Contract company workers and male independent contractors are paid more, on average, than their regular full-time counterparts with similar personal characteristics.
When the wages of nonstandard workers are compared to those of standard workers with the same job characteristics (industry, occupation, union representation, and receipt of health insurance and/or a pension) and the same personal characteristics, then the pay penalties shrink and the pay premiums grow (see the bottom half of Tables 4A and 4B). In other words, some nonstandard workers’ wages are comparable with (or even exceed) the wages of standard workers when the comparison is limited to workers with similar job and personal characteristics. This seeming contradiction occurs because nonstandard jobs are concentrated in low-paying occupations without fringe benefits or union representation.
When we control for personal and job characteristics, independent contractors make 13% (women) to 20% (men) more per hour than standard workers; self-employed men make 19% more. Between 1995 and 1997, there was a notable increase in wages for men who were either self-employed or working as independent contractors. In total, 37.2% of nonstandard workers (62.4% of men and 14.0% of women) are in nonstandard work arrangements with average wages higher than those of workers in standard jobs with similar personal and job characteristics. Unfortunately, the pay penalties experienced by workers in most of the low-wage nonstandard work arrangements (part-time workers, temps, on-call workers, and self-employed women) persist even when controlling for job and personal characteristics. Some 61.2% of all nonstandard workers are employed in the types of arrangements that, on average, pay less than regular full-time jobs occupied by workers with similar characteristics.
The share of nonstandard workers receiving fringe benefits from their employers follows the pattern found for wages. All types of nonstandard workers are less likely than regular full-time workers to have either employer-provided health insurance or pension benefits (see Table 5 and Table 6). The disadvantages of nonstandard work arrangements with respect to fringe benefits persist even when controlling for personal and job characteristics.4 Although some temporary help agencies claim to offer health insurance to their employees, the overwhelming majority of temporary workers of both sexes have jobs with no health care or pensions.5
Nonstandard workers are more likely than regular full-time workers to have jobs of limited or uncertain duration. Table 7 shows the percentage of wage and salary workers who reported that their jobs were temporary, or who could not work in their current jobs as long as they wished (even if their work performance was satisfactory).6 It also reports the percentage of the self-employed and self-employed independent contractors who expect their jobs to last less than one year.7 Among the various nonstandard work arrangements, temps were the most likely to report having a job of limited duration. The gender differences among the self-employed are striking: women who are self-employed and who are self-employed independent contractors have rates of limited job duration almost twice that of their male counterparts. The insecure nature of nonstandard jobs persists even when controlling for personal and job characteristics.8
Who works in nonstandard jobs?
Nonstandard work arrangements vary significantly in their gender and racial composition. Women are generally over-represented among nonstandard workers: 33.7% of women work in nonstandard jobs compared to 24.3% of men. Women are also concentrated in the nonstandard work arrangements that have the poorest-quality jobs. For example, regular part-time employees and temps – work arrangements with the largest pay penalties and the lowest likelihood of providing health insurance and pensions – are disproportionately female ( Table 8). In contrast, independent contractors – the best-paid type of nonstandard work – are disproportionately male.
Workers are also sorted into types of nonstandard work by race and ethnicity. Temps are disproportionately black and Hispanic ( Table 9
>). Whites are over-represented among independent contractors and the self-employed. Table 10 and Table 11 show work arrangements by race and sex. The best types of nonstandard work (i.e., independent contracting and self-employment for men) are disproportionately filled by white men, while the worst types of nonstandard jobs (i.e., regular part-time and temp work) are disproportionately filled by women and minority men. Despite this tendency, blacks and Hispanics of both genders are under-represented among nonstandard workers and over-represented among regular full-time workers.
Employment preferences and reasons for working nonstandard jobs
The extent to which nonstandard workers prefer their current work arrangement over a regular full-time job is highly correlated with the quality of their jobs. Table 12 shows the percentage of workers in selected nonstandard work arrangements that would prefer standard employment. The majority of temps and on-call workers would rather have a regular full-time job. In contrast, independent contractors and self-employed workers overwhelmingly prefer their nonstandard work arrangements. This is true even for self-employed women who earn less than their counterparts in standard jobs. However, in 1997, over 75% of self-employed women were white and married. This suggests their preference for self-employment may have been influenced by their ability to rely on the earning power of their husbands to maintain their standard of living. A preference for nonstandard work can reflect constraints faced by workers with conflicting obligations. For example, 74.6% of the men and 30.6% of the women who prefer part-time work cite school attendance as the reason for their preference. Another 44% of the women who “prefer” part-time work cite “family or personal obligations” as the reason.
Workers were also asked about their reasons for working in nonstandard work arrangements. These reasons typically fall into three categories: (1) economic reasons (i.e., they chose nonstandard jobs for economic reasons beyond their immediate control);9 (2) flexibility and family (they chose nonstandard jobs to have more flexible schedules or to allow them to care for their children or other family members); and (3) personal reasons. This last category encompasses a broad range of responses, such as the desire to seek training or education and a preference for short-term work.
Workers’ reasons for working in nonstandard work arrangements mirror their preferences ( see Table 13). Workers who prefer nonstandard work are more likely to work in those arrangements for reasons other than economic necessity. They are also likely to be employed in nonstandard work arrangements that have relatively better wages and benefits. Women working on-call are somewhat of an exception, however. Although these women, like most workers in low-quality nonstandard jobs, prefer traditional employment, 41.6% report they work on-call for flexibility and family reasons. This suggests that when some women “choose” to work in poor-quality nonstandard work arrangements, possibly because of family considerations, they would actually prefer to have regular full-time jobs.
To what extent are nonstandard workers also attending school and planning to get better jobs once they finish? Table 14 shows that only a small percentage of nonstandard workers aged 18 to 24 are students. While it is true that a large percentage of part-time workers in this age group are students, most are not.
The growth in nonstandard work arrangements over the past 10 to 20 years is difficult to assess. While the data used in this analysis were gathered in a survey designed specifically to assess participation in nonstandard work, this survey was first fielded in 1995. Thus, it does not provide information on long-term trends in nonstandard work arrangements. Other surveys (especially the basic component of the CPS) can be used to examine the time trends in a few types of nonstandard jobs: part-time employment, employment in the personnel-supply-service industry, and self-employment. However, the basic CPS uses somewhat different definitions of work arrangements than does the Contingent Work Supplement. While the basic CPS data are consistent over time, in any given year they will be somewhat different from data collected in the Contingent Worker Supplement.
Part-time employment varies with the fluctuations in the business cycle, rising during recessions and periods of slow economic growth and falling during periods of strong growth (Kalleberg, Wenger, and Hudson 1999). The total proportion of workers in part-time jobs has increased since 1973 ( Table 15). The share of workers employed part-time involuntarily (i.e., workers who want full-time work but cannot find it) has declined over the 1990s with improvements in the economy.10
While employment in the personnel-supply-service industry11 makes up a very small portion of the nonstandard workforce, two facts about these jobs stand out. First, they are among the worst jobs in the American economy. Second, these jobs are also one of the fastest-growing sectors of the labor market. Since 1982, the proportion of workers in temp jobs has increased more than fourfold, from 0.5% to 2.3%.
The basic CPS shows the level of self-employment over the last two decades has been fairly stable, with some slight decline since the end of the last recession in the early 1990s. In comparing the data in Table 15 with the rest of the data on self-employment in this report, it should be noted that there are two important differences. First, the data in Table 15 include both the self-employed as well as independent contractors who are self-employed. Second, the data on self-employed workers in Table 15 do not include self-employed workers who are incorporated.
Between 1995 and 1997 there was little change in the share or quality of nonstandard jobs in comparison to standard employment. This is somewhat surprising – the economy experienced good rates of growth during this period (2.1% and 3.6% in 1995 and 1996, respectively) and the minimum wage increased in 1996. The minimum wage increase alone should have improved the wages of the lowest earners and helped to narrow wage differentials between the lowest-wage earners and higher-paid workers. Unemployment, however, changed little in the period examined, hovering around 5.4% for most of the two years. These moderate improvements in both the overall economy and the minimum wage could be expected to shrink wage differentials and improve the quality of nonstandard jobs, but these changes seem to have had little effect on the share or the quality of nonstandard work.
Nonstandard work arrangements continue to be troubling because so many of them are of low quality. While these jobs may afford certain workers and their employ
ers a measure of flexibility, they pay less and provide fewer fringe benefits than regular full-time jobs, even when the workers have similar education and experience. Policy makers and working people are rightly concerned that nonstandard employment has the potential to become a mechanism for paying substandard wages and benefits. Public policies are needed to ensure that workers are not penalized with respect to their pay and benefits simply because of their work arrangement.
December 1999 EPI would like to thank the Ford Foundation for its support of this project. Appendix
(Click here for supplemental tables 1-9)
(Click here for supplemental tables 9-18)
(Click here for supplemental tables 19-27)
Distinguishing between regular full-time and nonstandard work
Nonstandard arrangements differ from regular full-time jobs in at least one of the following ways:
- 1. the absence of an employer, as in self-employment and independent contracting;
- 2. a distinction between the organization that employs the worker and the one for whom the person works, as in contract and temp work; or
- 3. the temporal instability of the job, characteristic of temporary, day labor, on-call, and some forms of contract work.
Defining the types of nonstandard work
Workers in this group were respondents who reported they were wage and salary workers and that they worked less than 35 hours each week, and they were not classified in any of the other nonstandard work arrangements (NSWAs) listed herein.
Temporary help agency (or temps)
Workers in this group were respondents who reported being a wage and salary worker and answered “yes” to the following question: “Are you paid by a temporary help agency? (A temporary help agency supplies workers to other companies on an as-needed basis or supplies workers to other companies primarily for short-term assignments.)” On-call
Workers in this group were respondents who reported being a wage and salary workers and answered “yes” to the following question: “Some people are in a pool of workers who are only called to work as needed, although they can be scheduled to work for several days or weeks in a row, for example substitute teachers, and construction workers supplied by a union hiring hall. These people are sometimes referred to as ‘on-call’ workers. Were you an on-call worker last week?” Day labor
Workers in this group were respondents who reported being self-employed and answered “yes” to the following question: “Some people get work by waiting at a place where employers pick up people to work for a day. These people are sometimes called day laborers. Were you a day laborer last week?” Self-employed
Workers in this group were respondents who reported being self-employed and answered “yes” to the following question: “Are you self-employed,” for example “as a shop or restaurant owner?” Independent contracting-wage and salary
Workers in this group were respondents who reported being a wage and salary worker and answered “yes” to the following question: “Last week, were you working as an independent contractor, an independent consultant, or a freelance worker? That is, someone who obtains customers on their own to provide a product or service. Independent contractors, independent consultants, and freelance workers can have other employees working for them.” Independent contracting-self-employment
Workers in this group were respondents who answered “yes” to the following question: “Last week, were you working as an independent contractor, an independent consultant, or a freelance worker? That is, someone who obtains customers on their own to provide a product or service. Independent contractors, independent consultants, and freelance workers can have other employees working for them” and answered “yes” to the question “Are you self-employed as an independent contractor, independent consultant, or freelance worker?” Contract company
Workers in this group were respondents who reported being a wage and salary worker and answered “yes” to the following question: “Some companies provide employees or their services to others under contract. A few examples of services that can be contracted out include security, landscaping, or computer programming. Did you work for a company that contracts out you or your services last week?”
We classified as “contract workers” all persons who did contract work, regardless of whether they work at the employers’ work site, the work site of a single contractee, or the work site of more than one contractee. This conception of contract work differs from that used by the BLS, which does not classify as contract workers persons who did not work at the contractee’s work site. BLS requires a respondent to answer “no” to the question, “Are you usually assigned to more than one customer?” and “yes” to “Do you usually work at the customer’s work site?” We do not require any particular answer to those questions.
Workers in this group were respondents who reported that they were wage and salary workers, worked 35 hours or more each week, and were not classified in any of the nonstandard work arrangements listed above.
In both the 1995 and 1997 Contingent Work Supplements, the Bureau of Labor Statistics has made a distinction between independent contractors who report they are wage and salary employees and those who report they are self-employed. The reasons for this response are unclear, but the data from the surveys reveal important distinctions between these two groups of nonstandard workers. Wage and salary and self-employed independent contractors often differ on the basis of their occupational characteristics as well as the quality of the jobs and the personal characteristics of those who fill them. The most frequently occurring occupations of wage and salary and self-employed independent contractors are listed respectively in Tables S20 and S21 in the online Data Supplement (www.epinet.org). Two distinctions are readily apparent. First, many wage and salary independent contractors are employed in occupations where workers are paid a base salary in addition to some type of commission or additional pay for each product sold or customer served. Second, self-employed independent contractors are more likely to work in higher status and “professional” occupations than their wage and salary counterparts.
On-call workersUnlike the February 1995 survey, the February 1997 survey makes a distinction between on-call workers who “only work when called” and those who “work regular hours but must be available” when called. Among all on-call workers, 43.2% indicated that they had a regular schedule in addition to being on call, while 52.3% indicated they only worked when called. Another 4.1% responded “other.” This indicates that the on-call group, which made up 94.1% of the on-call/day laborer group in the 1995 survey and 98.5% in 1997, includes a substantial portion of workers who have jobs with regular schedules, although they also work when called.
There are significant job-quality differences between these two types of on-call workers. The jobs of on-call workers with regular schedules are much more likely to resemble the jobs of regular, full-time wage and salary workers. However, as Tables S22, S23, and S24 in the online Data Supplement (www.epinet.org) show, there are large pay penalties for the on-call workers who work only when called. Table S25 in the Data Supplement shows the share of these workers with health insurance and a pension.
In addition to workers who work part time in what are otherwise regular jobs, a substantial portion of other nonstandard workers also work part time. Table S26 in the online Data Supplement (www.epinet.org) shows that nonstandard workers, especially women, are much more likely than regular wage and salary workers to work part time.
Multiple Job Holding
In 1997, nonstandard workers were more likely than regular full-time workers to hold more than one job. Table S27 in the online Data Supplement (www.epinet.org) shows that the highest rates of multiple-job holding are among wage and salary independent contractors and among female contract company workers. (Note: in Table S27, as in all of the tables, workers are categorized according to their “primary” job.) Among men, on-call workers with regular hours were more likely than other on-call workers to have multiple jobs. The opposite was true for women on-call workers.
Endnotes 1. See “Independent Contractors” in the Appendix to this report. 2. Survey respondents were categorized into work arrangements based on their responses to a series of questions. See “Defining types of nonstandard work” in the Appendix to this paper. 3. See “Regular Part-Time” in the Appendix to this report. 4. See Table S11 in the Data Supplement to this paper, available on the EPI web site (www.epinet.org). 5. While many nonstandard work arrangements purport to offer health insurance or pensions, these fringe benefits are frequently offered only on terms that are either prohibitively expensive or impossible to meet. The negative effects of nonstandard work on receipt of health and pensions benefits persist even when considering employers’ offers of benefits and not just the actual provision of the benefits (see Kalleberg, Reskin, and Hudson, forthcoming). 6. All on-call/day laborers were regarded as having jobs of limited duration 7. It is possible that the lower percentages for the self-employed and self-employed independent contractors compared to the wage and salary workers is due in part to the difference in the way “limited duration” is measured for the two groups. 8. See Table S12 in the Data Supplement to this paper, available on the EPI web site (www.epinet.org). 9. The “economic reasons” category includes the following: (1) “laid off and hired back as a” nonstandard worker, (2) “only type of work could find,” (3) “hope job leads to permanent employment,” and (4) “other economic.” 10. The decline in involuntary part-time employement also is due to a change in the questions asked by the basic Current
Poplulation Survey. This change was implemented in January 1994 (see Nardone, Tom, 1995, “Part-Time Employment:
Reasons, Demographics, and Trends,” Journal of Labor Research, Vol. 26, pp. 274-92). 11. Employment in the personnel-supply-service industry (BLS detailed industry code 731) is the best longitudinal measure available for estimating the relative portion of the workforce employed by temporary help agencies. However, it should be noted that, while the two overlap considerably, they are not completely contiguous.References Hudson, Ken. 1998. “Duality and dual labor markets.” Paper presented at the meetings of the American Sociological Association, San Francisco, Calif., August 1998.
Kalleberg, Arne L., Edith Rasell, Naomi Cassirer, Barbara F. Reskin, Ken Hudson, David Webster, Eileen Appelbaum, and Roberta M. Spalter-Roth. 1997. Nonstandard Work, Substandard Jobs: Flexible Work Arrangements in the U.S. Washington, D.C.: Economic Policy Institute and Women’s Research & Education Institute.
Kalleberg, Arne L., Barbara F. Reskin, and Ken Hudson. Forthcoming. “Bad jobs in America: standard and nonstandard employment relations and job quality in the United States.” American Sociological Review. Revision of paper presented at the 92nd Annual Meetings of the American Sociological Association, Toronto, Canada.
Kalleberg, Arne, Jeffrey Wenger, and Ken Hudson. 1999. “Trends in nonstandard work arrangements.” Work in progress.
Spalter-Roth, Roberta M., Arne L. Kalleberg, Edith Rasell, Naomi Cassirer, Barbara F. Reskin, Ken Hudson, David Webster, Eileen Appelbaum, and Betty L. Dooley. 1997. Managing Work and Family: Nonstandard Work Arrangements Among Managers and Professionals. Washington, D.C.: Economic Policy Institute and Women’s Research & Education Institute.
(Click here for supplemental tables 1-9)
(Click here for supplemental tables 9-18)