Computer Chips and Paper Clips: Technology and Women's Employment, Volume I (1986)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Effects of Technological Change: Employment Levels and Occupational ShiRs What are the implications of innovations in telecommunications and infor- mation processing, and associated changes in the organization of work, for the availability of jobs to women, particularly over the next 10 years? The previous chapter suggests that these innovations have varying effects on the quantity of jobs in different occupations and industries. Much may depend on how they are implemented and on underlying economic and social conditions. If the new information technologies do adversely affect clerical employment, women will be disproportionately affected. Over the next 10 years, will there be significant displacement of present clerical workers? Will there be sufficient clerical jobs for all those who seek them? Any attempt to look at future effects is, of course, limited; at best, prediction is uncertain, an informed guess. This chapter seeks to reduce that uncertainty. The context of this inquiry is one of substantial controversy in the existing research literature and substantial concern expressed in the media. Some studies predict large relative job losses for clerical workers. For exam- ple, in a major study based on an input-output matrix for the United States with moderate economic growth and with two scenarios of diffusion of computer- based innovations, Leontief and Duchin (1984) predict that employment growth in clerical occupations will be very small relative to growth in the total labor force. With the most rapid diffusion, they projected total employment to grow from 89.2 million person-years in 1978 to 124.1 million in 1990 and to 156.6 million in 2000; clerical employment, however, was projected to grow from 15.9 million in 1978 to only 16.7 million in 1990 and 17.9 million in 2000. Hence, under their assumptions, clerical workers would decline from 17.8 percent of the labor force in 1978 to 13.5 percent in 1990 and to 11.4 percent in 2000. Furthermore, Leontief and Duchin project that if workers 62

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 63 continued to train for the mix of skills required in 1978, it is possible that 5 million clerical workers would be unemployed by 1990. Of course, job growth would occur elsewhere: in professional, service, craft, and operative occupa- tions. ~ [though a number of criticisms have been directed at this study (see Hunt and Hunt, 1985a,b; Kraft, 1985), it systematically illustrates one end of the range of predictions. Several economists have noted that the job-creating aspects of new technolo- gies are more difficult to predict than the job-displacing aspects are to observe. In testimony before the House Committee on Science and Technology, Nathan Rosenberg (1983) pointed out that although technological change virtually al- ways involves occupational shifts, it need not involve overall job loss; he sug- gested that current U.S. employment problems are more the result of macroeco- nomic policy than of technological change. Similarly, Eli Ginzberg (1982), introducing the special technology issue of Scientific American, emphasized the job-creating nature of technological change. Without technological ad- vance, economic growth slows and employment declines can occur, especially in industries that fail to innovate. A case in point with respect to clerical work is provided by the finance indus- try, a heavy user of automated information processing and a classic example of technological change that did not result in an employment decline. Between 1972 and 1982, clerical workers as a proportion of all workers employed in the finance industry decreased from 46.1 to 43.9 percent (Hunt and Hunt, 1985a), indicating that it became possible to produce any given output with relatively fewer clerical workers than previously. Office automation may have contrib- uted to this relative reduction in clerical workers. During this same period, however, total employment in the industry expanded substantially from 3.9 million to 5.3 million, an increase of 35.9 percent, about 1.5 times the rate of employment increase in all nonagricultural industries (Bureau of Labor Statis- tics, 1985b) in part because the new technologies facilitated the provision of new and better services to consumers at lower cost. Thus, although clerical employees comprised a smaller proportion of the work force in the finance industry in 1982 than in 1972, the total number of clerical workers employed in the industry rose considerably. Indeed, the increase in clerical jobs in the fi- nance industry over the decade was 37 percent, considerably higher than the 28.8 percent increase in clerical employment for the economy as a whole (Hunt and Hunt, 1985a). PROBLEMS IN EMPLOYMENT PROJECTIONS UNDERLYING FACTORS Of course, no one can be certain whether the pessimistic view, the optimistic view, or something in between will prove most accurate for clerical employ-

64 COMPUTER CHIPS AND PAPER CLIPS ment over the next 10 years. Four major underlying factors affect this future, some of which can be better predicted than others: (1) the general performance of the economy overall and the associated general employment picture; (2) specific policies that affect women's employment opportunities; (3) changes in the supply of labor; and (4) developments in the available technology. These factors are considered briefly here, in reverse order. The technology that is likely to influence employment in the next decade either has already been developed or its likely parameters are fairly well speci- fied. As described in Chapter 1, the ease of processing and transmitting infor- mation is increasing rapidly as the cost falls. Some uncertainty exists about the rate at which innovations in telematics will diffuse throughout the economy. In the panel's judgment, diffusion will not accelerate over the next 10 years: delib- erate rather than headlong speed seems likely. The keyboard will remain the main mode of data entry; the new integrated workstation will remain unavail- able to all but a few; networking between systems will remain a problem. Of course, unforeseen applications of known technology can alter a situation dra- matically. These technologies do have the potential to contribute very greatly to productivity increases, but they remain relatively untapped. The productivity gains that are possible can be applied in several ways: to develop new products and services, to improve quality, or to cut costs. Only the third way is likely to have substantially negative effects on employment levels, and those will occur only if the lower costs (and prices) do not induce increased demand. The rate of growth of the labor force i.e., changes in the supply of labor—is obviously an important factor in its adjustment to technological change. If, for example, labor force growth slows considerably while the rate of economic growth does not change, the economy can more readily absorb workers who may be displaced by technology. If, however, labor force growth remains high, workers displaced by technology will be more likely to remain unemployed along with new entrants and reentrants to the labor force. The people who will enter the labor force during the next decade have already been born, and recent trends in labor force behavior can provide a guide to future behavior. Again, however, unforeseen change is possible. Throughout most of the postwar pe- riod, for example, forecasters consistently underestimated women's labor force participation. Public and private policies in the areas of child care, transporta- tion, training, flex-time, and so on can change basic structural factors that influence people's labor force participation. It is important to note, too, that the education and training of current workers and new entrants to the labor market affect the ease with which they can respond to changing labor market condi- tions. (Future labor supply is examined in greater depth below.) Specific policies that affect women's employment opportunities directly, such as enforcement of equal employment opportunity laws, can have a major impact on the availability of jobs for women. Even if the number of jobs avail-

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 65 able in occupations and industries that have been dominated by women fall or experience slower growth in the next decade, if women's opportunities increase elsewhere in the labor market, the effect of the shrinkage of traditional employ- ment for women would be mitigated. Currently there is some uncertainty about the equal employment opportunity enforcement effort. Women made substan- tial progress in integrating many occupations, particularly in the last decade (see Blau and Ferber, 1986; Reskin and Hartmann, 1986), and it seems likely that some progress will continue. The general performance of the U.S. economy is perhaps the most important factor in assessing future job opportunities for women (or men). If economic growth is rapid and demand for labor is high, all those whose jobs become outmoded by technological change will almost certainly be able to find new ones. But if overall economic performance is sluggish, replacement jobs will be hard to find, and the costs of adjusting to job loss that is due to technological change will be high in both economic and human terms. Future economic performance is difficult to predict because it depends so heavily not only on national economic factors but also on the outcome of politi- cal processes and international economic developments, such as the volume and nature of trade, the strength of the dollar, and the success of OPEC in controlling the price of oil. Among important public policies that affect the general health of the economy and the overall demand for labor are fiscal and monetary policies that influence interest rates, the deficit, and inflation and tax policies, tariffs, and other statutory incentives and disincentives that influence the development and use of innovations and so alter capital/labor ratios. All these policies affect the development of new technologies and the rate at which they are used. And government policies, in turn, are influenced by alterations in political and social commitment to providing full employment. The panel made no attempt to predict fixture economic growth, but the dis- cussion in the rest of this chapter implicitly assumes modest growth rates over the next 10 years. In the panel's judgment, this assumption is a reasonable basis on which to gauge the effect of technological change on women's employment. But it is possible that the economic future could be vastly different, and technol- ogy could play a different role in that future. If economic performance deterio- rates and economic pressures on employers intensify, the productivity gains made possible by the new technologies are likely to be applied to cost cutting, and substantial technological displacement might occur, along with the cyclical unemployment that would result from poor overall economic performance. Whatever the impact of technological change and economic growth on the levels of employment, their impact on occupational shifts is less ambiguous. Technological change and economic growth invariably generate shifts in de- mand for workers in various occupations. Clearly, a healthy economy will con- tribute to the ease of adjustment to these shifts.

66 COMPUTER CHIPS AND PAPER CLIPS DATA PROBLEMS Another source of uncertainty in discussing the employment effects of tech- nological change is the inadequacy of the available data for studying the con- nection between technology and employment. The data are inadequate for as- sessing the impact of technological change on both the quantity and quality of employment. To assess changes in employment opportunities or outcomes with any degree of confidence in the generalizability of the results requires data from a represen- tative sample of jobs in a local or national labor market, with jobs grouped into occupational categories on the basis of some standard occupational classifica- tion scheme. But none of the existing occupational classification schemes e.g., those of the Bureau of the Census, the Bureau of Labor Statistics (BLS), the Dictionary of Occupational Titles, or the Standard Occupational Classifica- tion (SOC)—was designed with the assessment of technological change as an important consideration. These schemes are also inconsistent with one another. Moreover, they are updated periodically (the decennial census classification, for example, is revised for each new census) without systematic attention to the way technological changes have redefined the task content of jobs. A specific example illustrates the difficulties. Suppose one wanted to deter- mine how widespread the use of word processing was by 1980. In principle, this is the sort of question it should be possible to answer from 1980 census data. It is impossible to do so, however, because the 1980 census classification has no separate category for "word processors." Formerly they were included in the category "typists." For the 1980 census, because word processors work with computer technologies, the Census Bureau decided to reclassify them with keypunch operators, as does the SOC. This change was also implemented in the monthly Current Population Survey (CPS), but not until 1983. The change in the SOC is a dramatic one, affecting major occupational groups as well as more detailed categories such as typist or keypunch operator; 1980 data are now not generally comparable with earlier censuses. The Bureau of Labor Statistics' periodic reports on the labor force would be of no greater help, because they are generally based on the Current Population Survey (which has too small a sam- ple size for reliable data on the numbers of workers who use word processors) and because the classification also does not identify "word processors" sepa- rately. Furthermore, none of the classifications in use would identify workers who use word processing as an auxiliary part of their jobs for example, writ- ers, editors, managers. Thus, there is no direct means of tracing the adoption and use of this new technology. Of course, industry statistics on the sale of stand-alone word processors and on the sale of word-processing programs for general-purpose microcomputers would allow some basis for an estimate, but such statistics can reveal nothing about which categories of workers use word

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 67 processors, how many workers share each workstation, and so on, or how the organization of work has changed in response to the introduction of such equip- ment. The fundamental problem is that information on technology is not linked to information on workers in currently available data (see Hunt and Hunt, 1985b). The inconsistencies across time, data sets, and agencies cause difficulty even in analyzing categories such as secretaries or clerical workers, let alone word processors. Although BLS uses "census categories" in the household data col- lected in the Current Population Survey, it uses other classification systems in other data collection efforts. The area wage surveys, for example, rely on the categories used by the employers being surveyed. (Some of these may even identify word processors separately.) A third data collection effort, the Occupa- tional Employment Statistics (OES) program, the biennial surveys of establish- ments that form the basis for the occupational projections, uses yet another classification scheme. Consequently, the numbers that are developed from the data sets sometimes differ widely for the same occupational category. For example, according to OES data, there were 2,797,000 secretaries in 1984 (Silvestri and Lukasiewicz, 1985~; according to the annual averages from the CPS monthly data published in Employment and Earnings (January 1985), there were 3,935,000. Although the OES classification system is different from the CPS, the difference probably does not stem from differences in the defini- tion of "secretary" between the two systems but from differences between self- reporting (CPS) and employer reporting (OES). For another example, OES data show a total of 18,716,000 administrative support workers in 1984 (17.5 percent of the civilian employed labor force of 106.8 million), while the CPS total is 16,722,000 (15.9 percent of the total of 105.0 million). In the 1980 census, the count was 16,851,000 (or 18.5 percent of the labor force). To address some of these data problems, the panel commissioned a study on recent and future trends in clerical occupations from H. Allan Hunt and Timo- thy L. Hunt (1985a) of the Upjohn Institute. They developed a reasonably consistent set of occupational categories for clerical workers across the 1950- 1980 censuses. To look at more recent trends, they use CPS data for the 1972- 1982 period, and they sometimes report 1983-1984 CPS data separately be- cause they are not compatible with earlier data. In looking at projections, they use the BLS projections based on OES data. Their estimates form the basis for the panel's discussion later in this chapter. Some of the same concerns apply to the assessment of employment quality, which is discussed in Chapter 4. There are several systems of inquiring about the quality of work life: the Dictionary of Occupational Titles, the Quality of Employment Survey, and numerous disparate surveys on job satisfaction. None has emerged as the clearly dominant method for measuring quality, and none is likely to do so. Moreover, there has been no systematic attempt to link employ-

68 COMPUTER CHIPS AND PAPER CLIPS ment quality to specific technologies used on the job, although occasional pri- vate attempts have been made, e.g., the survey of women who do and do not use video display terminals that was taken of the readership of women's maga- zines (9-to-5, 1984a). There is at present no way of knowing how many jobs are affected by technological changes, much less what the effects are. A statistical system that does not distinguish between typing on a typewriter and data entry on a word processor or keypunch machine is inadequate as a basis for determining the extent of technological change in the office. The point is not limited to office jobs, of course, but pertains equally to many categories of jobs affected by technological changes. The lack of an adequate data base requires that one maintain a healthy skepticism regarding research on both the quantity and quality of the employment effects of technological change. This chapter next turns to further examination of the supply of labor, particu- larly women's labor. It then reviews recent trends in clerical employment as a base from which to understand future possibilities. The following section re- views several aggregate projections of demand for clerical labor, illustrating a "most plausible worst case scenario" that the panel believes places a lower limit on the likely demand for clerical employment in 1995. Employment pro- jections for specific subfields of clerical work and likely occupational shifts and changes in skill requirements are also discussed. The conclusion attempts to link information about supply and demand to assess the impact of technological change and other factors on employment levels in clerical jobs, currently the dominant source of jobs for women. THE SUPPLY OF WOMEN WORKERS LABOR FORCE PARTICIPATION RATES The future supply of women workers is a function of the number of women and the rates at which different groups of women participate in the labor force. The number of women is determined by the fertility that prevailed 16 or more years earlier, by women's current mortality rates, and by the net balance of women's immigration and emigration. Of these factors, of course, the first is generally dominant. Women (and men) who will be at work in 1995 have al- ready been born. Barring unforeseen developments in mortality or immigra- tion, we can project fairly confidently the number of women of working age in 1995 (see, for example, Bureau of the Census, n.d.~. In addition, knowledge of fertility, mortality, and migration permits one to project current and future age distributions, which is important because women's labor force participation rates vary by age. Current projections by the Census Bureau indicate that the working-age pop- ulation will continue to grow, although at a slower rate than in the recent past.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 69 The slowdown is primarily a result of the fact that the baby boom generation has come of age and already entered the labor force. The population cohorts reach- ing working age over the next 10 years will not be as large. If labor force participation rates remained constant by age, the projected average age of the labor force would also nse. The rates at which women will participate in the labor force are, however, much more difficult to project. Two factors contribute to the difficulty: the variations in participation rates among groups of women and the rapid increases in participation rates among almost all women. The rates at which women participate in the labor force have historically varied according to women's ages, their potential wage rates, marital status, husband's income if mamed, level of education, and presence and age of children. The rates have also dif- fered by racial and ethnic group. Labor force participation rates have histon- cally been highest for women aged 20-24; for unmamed or divorced women; for well-educated women; and for women without children, especially young children, at home (Bureau of the Census, 1985:Tables 671, 673, 675~. 'The rates for black women have historically been higher than those for white women, and white women's rates have been somewhat higher than those for most women of Spanish origin (who may be of any race). In 1984 the labor force participation rate for black women was 55.2; for white women it was 53.3; and for women of Spanish origin it was 49.6 (Bureau of the Census, 1985 : Tables 660 and 670~. But there can be marked differences within a group: for example, in 1984 the Cuban women's participation rate was 55.1, higher than that for any other group of Spanish origin and higher than the rate for all white women. (For a description of racial and ethnic differences in women's rates, see Sullivan, 1978.) And although differences in rates among groups of women have persisted, there has been a pronounced, and largely unanticipated, increase in the labor force participation rates among all groups of women. While no single reason for this increase can be identified, many recent eco- nomic, social, and cultural changes are conducive to it. Among demographic changes, lower levels of fertility and the greater use of child care have made it easier for mothers of young children to hold jobs; at Between 1985 and 1990, the population over the age of 16 will increase by about4.8 percent for both men and women. The rate of growth will slow between 1990 and 1995 to around 3.8 percent, with the male population growing at a slightly higher rate of 3.9 percent (calculated from Bureau of the Census, n.d. :43, 53, 63, 73; middle series of projections). During the five years from 1995 to 2000, as the children of the baby boom enter the labor force, a growth of 4.4 percent in the female working-age population and 4.6 percent in the male working-age population is projected. In 1985 the working-age population is estimated at 95.3 million women and 87.7 million men; by 2000 the projected population will be 108.3 million women and 99.9 million men. Principally as a result of recent fertility trends, the median age of the population will rise from 32.7 for women and 30.2 for males in 1985 to 34.9 for men and 37.7 for women in 2000.

70 COMPUTER CHIPS AND PAPER CLIPS least one study suggests that further improvements in child care would encour- age even more such mothers to work (Presser and Baldwin, 19801. Later ages at marriage and at first childbearing have enabled many women to gain greaterjob seniority and so perhaps increased their commitment to work outside the home. Among factors related to financial need, inflation and husbands' unemploy- ment or the erosion of husbands' real wages may have encouraged some wives to enter the labor force or to remain in it. Financial pressures are also a factor in decisions by divorcees or widows to reenter the labor force. Moreover, in- creases in marital instability in the past several decades may have encouraged some women to maintain their work roles as a source of financial indepen- dence. Institutional changes, including legal action against sex discrimination and sexual harassment in education and employment, greater equity in pensions and other fringe benefits, and greater availability of maternity leave, may also have played a role in the higher participation rates. Cultural change, coincident with the rise of the feminist movement and reflected in the changed aspirations and expectations of women, may be especially significant. As more women work, employment becomes the norm; there are changes in attitudes toward women's work and the role of work in adults' lives. These changes have probably encour- aged more women to remain in the labor force. Finally, some analysts have argued that shifts in labor demand have in- creased opportunities for "women's jobs" (Oppenheimer, 1970~. Although it is difficult to demonstrate that women entered or stayed in the labor force because of the availability of jobs, it is significant that very large numbers of the same kinds of workers who had traditionally held certain jobs arrived in the labor market just at the time of highest demand for those jobs. PROJECTIONS OF LABOR FORCE PARTICIPATION RATES Given the number of influences that might be affecting women's labor force participation rates, it is very difficult to make assumptions about trends in the rates. Hence, projections of future labor supply do not have a firm basis. While the number of women of working age is well known and relatively easy to project, past projections underestimated the rapid increases in the rates and therefore in women's labor supply. Because no one can predict whether the participation rates will continue to increase or at what rate they will change, projections of women's labor supply are uncertain. By contrast, projections of men's labor supply have been relatively accurate because a basic assumption has been reliable: after completing their education, very high proportions more than 75 percent of all working-age males will remain in the labor force until retirement age. This assumption holds almost without regard to men's marital status, education, or any other characteristic except health status, al-

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS TABLE 3-1 Actual and Projected Labor Force Participation Rates for Women Aged 16 and Older by Race, 1970-1995 Year All Women Whites Blacksa 1970 43.3 42.6 49.5 1975 46.3 45.9 48.9 1980 51.5 51.2 53.2 1984 53.6 53.3 55.2 990b 56.6 n.a. ma. 1995b 58.9 58.4 62.7 aIn 1970, refers to "blacks and other races." bProjected rates. SOURCES: Data for 1970, 1975, 1980, and 1984 from Bureau of the Cen- sus (1985:Table 660); data for 1990 and 1995 from Fullerton (1985:Tables 2 and 3). 71 though black men have lower labor force participation rates (by about 5 to 10 percent) than white men at all ages. The best available current projections are those of BLS. Table 3- 1 presents women's labor force participation rates between 1970 and 1984 and projected rates for 1990 and 1995. These projections are probably the most responsible professional estimates of labor force participation rates, but it should be noted that rates rose so quickly after 1970 that labor force projections for 1985 were surpassed during the 1970s. The estimated size of the female labor force in 1990 would be 55.5 million and in 1995 it would be 59.9 million; in compar~- son, in 1984 the size was approximately 49.7 million. Table 3-2 provides the same information for men for the same time period. TABLE 3-2 Actual and Projected Labor Force Participation Rates of Men Aged 16 and Older by Race, 1970-1995 Year All Men Whites Blacksa 1970 79.7 80.0 76.5 1975 77.9 78.7 71.0 1980 77.4 78.2 70.6 1984 76.4 77.1 70.8 1990b 75.8 ma. n.a. 1995b 75.3 75.8 69.5 aIn 1970, refers to "blacks and other races." bProjected rates. SOURCES: Data for 1970, 1975, 1980, and 1984 from Bureau of the Cen- sus (1985:Table 660); data for 1990 and 1995 from Fullerton (1985:Tables 3 and 4).

72 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-3 Size of Civilian Labor Force, Proportion Female, and Actual and Projected Growth Rates, 1970-1995 . Labor Force Growtha (average annual rate of change) Civilian Proportion Civilian Labor Force Female Labor Male Female Year (millions) (percent) Force Workers Workers _ . 1970 82.8 38.2 - - — 1975 93.8 40.0 2.5 1.9 3.5 1980 106.9 42.5 2.7 1.8 4.0 1984 1 13.5 43.8 1.5 1.0 2.2 0b 122.7 45.3 I.3 0.8 I.9 1995b 129.2 46.4 1.0 0.6 1.5 aAnnual rate of growth from the preceding date. bProjected data. SOURCE: Fullerton (1985:Table 1). Although the overall male labor force participation rate is projected to be higher than the overall female rate in every year, the male rate is projected to decline slightly, largely because of earlier retirement. Steeper declines occurred among both black and white males between 1970 and 1984 than are expected to occur from 1985 to 1995. Tables 3-1 and 3-2, taken together, strongly suggest that men will account for a declining proportion ofthe labor force. Summary projec- tions, shown in Table 3-3, show an increase in the proportion of the labor force that is female and consistently higher growth rates in the female labor force. However, growth of both the male and female labor force is expected to decline after 1985. For the 1970-1984 period, the number of women in the civilian labor force grew at an annual rate nearly twice as large as that anticipated between 1985 and 1995. PROJECTIONS OF AGE-SPECIFIC RATES Along with a general increase in women's rates of participation at most ages, some changes in the patterns of age-specific labor force participation are being projected for the next 15 years. For most of the postwar period, women's labor force participation rates peaked at ages 20-24, dropped at ages 25-34, and rose again at ages 35-54. The drop in rates at ages 25-34 was attributed to with- drawal from the labor force for childbearing. In 1970, 57.7 percent of women aged 20-24 were in the labor force, the rate dropped almost 13 points, to 45.0 percent, for women aged 25-34 and then rose 6 points, to 51.1 percent, for women aged 35-44. In every year since 1970, the percentage of women in the labor force at ages

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 73 20-24 has risen, and the drop in rates between ages 25 and 34 has declined. In 1975 the rate for the latter group dropped 9 points, from 64.1 to 54.9 percent; by 1980 the decline was only 3 points, from 68.9 to 65.5 percent. By 1984, the decline was less than 1 point, from 70.4 to 69.8 percent. Projections for 1990 no longer include a decline at ages 25-34 (Bureau of the Census, 1985:Table 6601. In addition to this change, the peak age for women's labor force participation is expected to shift. By 1990, labor force participation for women is expected to peak at 78.6 percent for ages 35-44, rising in 1995 to 82.8 percent for that age group. Although these cross-section rates do not indicate changes in behavior ofthe same women over theirlifetimes, the patterns they suggest are essentially borne out in data about successive cohorts. Table 3-4 shows both cross-section rates and cohort effects (shown in the steps going down the table). Each cohort was less likely than the one before to experience dramatic declines in labor force participation rates during the childbearing years, so that the most recent cohort experienced no decline at all. The members of each cohort, as they age, increase their participation in the labor force, and each successive cohort has higher participation than the one before. The projected pattern would resemble the male pattern of labor force partici- pation much more closely than have earlier female patterns. The 35-44 age group for men has typically had the highest labor force participation, with rates well over 95 percent. Among males, a striking change has been projected in age pattern increased withdrawal of older workers from the labor force. In 1970, 83 percent of men were in the labor force at ages 55-64; this proportion is projected to drop to 64.5 percent by 1995. Participation among males over the age of 65 is projected to be about ore-half of what it was in 1970, with a decline from 26.8 to 13.3 percent. By contrast, very little change is projected in the participation rates for older women workers: in 1984, 41.7 percent of women aged 55-64 were in the labor force, along with 7.5 percent of those over 65. OTHER FEATURES OF WOMEN S LABOR FORCE PARTICIPATION The increased similarity of the labor force participation of women and men over the life cycle corresponds to an increase in female attachment to the labor force. Women are working more and more consistently over the life cycle. One measure of attachment is the turnover rate, the extent to which workers in the labor market at any one time leave during the year and are replaced by others. If the same workers stayed in the labor market all year long, the turnover rate would be zero. For women workers, the turnover rate has fallen rather steadily from 32.1 percent in 1957 to 12.7 percent in 1983; for men, the turnover rate fluctuated modestly around 8.0 percent between 1957 and 1977 and after 1977 fell to 4.5 percent (Blau and Ferber, 1986:Table 4.31. Although there is still

74 U. a' - V: Cal Ct sly an Ct At o U. sly Ct to : 3 Cat o U: in: o ._ Ct ._ .= Ct o o r oo Ed ~ C) of At: o lo to ax In Via UP to -1 Via O at l Cal O do 0 mo ~ 00 ~ ~ 00 ~ ~ . . . . . . . ~ ~ ~ ~ ~ t— _ . . . . . . . ~ ~ rid ~ ~ ~ ~ ~ ~ ~ _ ~ ~ ~ 0 ~ 0 ~ ~ ~ ~ . . . . . . . ~ _ ~ ~ Cat \0 ~ ~ 0 . . . . . . . Us ~ ~ cr. ~ Go 0 00 ~ 00 Cat 00 00 . . . . . . . _ 00 0 ~ c~ t— O u~ u~ ~ ~ ~D O ~ _ . . . . . . O _ U~ V) ~ ~ V~ _ oo ~ ~ oo . . . . . oo ~ ~D V) V) oo ~ ~ . . . . 0N 0 ~ cr . . . ~ ~ oo . . U~ _ o U~ - 1 - - ~ O V) \0 ~ 0N _ _ oo C~ . . _ 1— _ 0N . . ~ ~ _ '= ~ ~ ~ _ - , . . ~ —~ O ~V~ =- . . oo U~ O . . \0 0 O oo 1— ~) V) oo _ . . U~ ~ O V) O ~ O l~ l~ C~ _ _ _ _ _ 1 1 1 1 1 —~ _ ~ ~ ~ ~ ~ In c~ a~ c~ _ _ _ _ _ =. ~ oo . . ~ — - 1 - cr - ~ O U-) O ~ \0 ~ ~ oo oo C~ ~ ~ ~ Ch _ _ _ _ _ D - 0 3 m 00 _ O . ~ E~ Ct Ct _ Ct :: _ 00 _ _ U: C) ._ C~ ._ Ct O - 0 e~ m 00 . ~ o^ _ _ ca C) ._ C~ ._ - Ct O m E: ^ Ct _~ _ 1 ce V) ~: _ ~ ~ .. =0 .. V) — C: .o ~ ._ _ ~ 0 ~

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIMS 75 some difference in average turnover for women and men, women's labor force behavior is approaching that of men. Moreover, when the differences in earn- ings or occupations between women and men are taken into account, turnover differs very little between them (Reskin and Hartmann, 19861. Job tenure, a measure of attachment to a particular job rather than to the labor force as a whole, also shows convergence between women and men (Blau and Ferber, 1986:Table 7.81. Another indicator of women's increased labor force participation and de- creased turnover is their work-life expectancy. In 1979-1980 a 20-year-old woman could expect to work 27.2 years, compared with 14.5 years in 1950; a 20-year-old man could expect to work 36.8 years in 1979-1980 (S. Smith, 1985~. The percentage of women working full time, year round has also in- creased steadily, although it is still less than that of men (Cain, 1985:Table 7~. In 1983, 48 percent of women held full-time, year-round jobs, representing an increase of about 10 percentage points over a 20-year period. In 1983 the rate for men had not changed much and stood at about 64 percent (Blau and Ferber, 1986:75~. EDUCATIONAL ATTAINMENT OF THE LABOR FORCE Education is an important factor influencing women's labor force participa- tion and their earnings. Both the amount and type of education of workers are important for labor market outcomes. Educational preparation is especially important to workers' ability to adapt to change. Throughout most of the history of public education in the United States, women have been about as well educated as men. Women have been more likely to complete high school than men, but men were introduced earlier and in larger numbers to college education. Thus, although the average years of schooling attained by women and men-have been fairly equal, women and men have been distributed across levels of education differently. Men have had higher proportions of both high school dropouts and college degrees; women have had higher proportions of high school graduates. These trends are shown in Tables 3-5 and 3-6. For the population as a whole, more men than women have completed col- lege. In recent years women have earned almost as many college degrees as men (49.8 percent of four-year degrees in 1981; Blau and Ferber, 1986~. Table 3-7 shows the changes in the population over age 25 who completed college between 1967 and 1981; it shows the effects of recent increases in college completion for women. It also shows increases in the proportion of black women and men who have completed college. The educational attainment of men and women in the labor force has also been fairly equal. In 1962, median years of school completed were 12.3 for

76 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-5 Mean Schooling Levels by Birth Cohort Years of Schooling Birth Cohort Males Females 1951-1954 12.57 12.65 1946-1950 12.62 12.39 1941-1945 12.21 12.05 1936-1940 11.85 11.70 1931-1935 11.50 11.39 1926-1930 11.17 11.16 1921-1925 10.89 10.97 1916-1920 10.46 10.56 1911-1915 9.83 10.09 1906-1910 9.41 9.75 1901-1905 8.85 9.15 1896-1900 8.44 8.71 1891-1895 7.92 8.19 1886-1890 7.51 7.89 1881-1885 7.31 . 7.70 1876-1880 7.20 7.65 1871-1875 6.92 7.27 1866-1870 6.79 7.16 SOURCE: Smith and Ward (1984:Table 17). Reprinted with permission. TABLE 3-6 Schooling Distnbutions of Selected Birth Cohorts Years of Schooling Less Less Less More Birth Than Than Than Than College Cohort 0 5 9 8 12 12 12 16 Degree Men 1866-1870 7.5 21.1 77.8 35.5 85.8 9.0 5.2 1.6 1.9 1886-1890 5.7 17.4 66.3 28.4 79.3 12.0 8.8 2.3 3.3 1906-1910 1.4 6.6 41.6 20.3 61.5 23.1 15.4 4.6 6.3 1926-1930 0.7 2.8 17.4 8.2 38.6 42.3 19.1 5.7 7.7 1946-1950 0.4 1.5 6.6 2.3 19.1 45.6 35.3 12.9 15.9 Women 1866-1870 8.5 26.0 82.1 33.6 88.1 5.7 6.2 2.2 3.3 1886-1890 6.4 21.6 71.9 28.2 82.1 8.6 9.0 2.8 4.6 1906-1910 1.8 8.8 47.5 22.3 67.0 18.0 15.0 4.7 7.8 1926-1930 1.0 3.9 22.8 10.5 43.5 31.2 25.3 8.7 14.2 1946-1950 0.5 1.2 7.8 3.0 19.4 37.6 43.1 17.6 20.4 SOURCE: Smith and Ward (1984:Table 18). Reprinted with permission.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS TABLE 3-7 Population 25 Years or Older with Four or More Years of College, 1967-1981, by Gender, White, Black, and Spanish Origin 77 Spanish Year Men Women White Black Ongina Percentage 1967 12.8 7.6 10.6 3.9 n.a. 1968 13.3 8.0 11.1 4.3 n.a. 1969 13.5 8.2 11.2 4.6 n.a. 1970 14.1 8.2 11.8 4.5 n.a. 1971 14.6 8.5 12.0 4.5 n.a. 1972 15.4 9.1 12.6 5.1 n.a. 1973 15.9 9.6 13.1 6.0 n.a. 1974 16.9 10.1 14.0 5.6 n.a. 1975 17.6 10.6 14.6 6.5 8.4 1976 18.6 11.3 15.4 6.6 6.1 1977 19.2 12.1 16.2 7.3 6.2 1978 19.7 12.2 16.4 7.2 7.1 1979 20.4 12.9 17.2 7.7 6.6 1980 20.9 13.6 17.9 8.0 7.6 1981 21.1 13.4 17.8 8.2 7.7 Number (thousands 1967 6,369 4,188 9,991 377 n.a. 1975 9,679 6,565 15,063 713 301 1981 13,208 9,466 20,775 1,084 485 aPersons of Spanish origin may be of any race. SOURCES: Data for 1967 through 1974 from Bureau of the Census (1980:Table 6/15); data for 1975 through 1981 from Bureau of the Census, Educational Attainment in the U. S., venous years, all Table 1. women and 12.1 for men; in 1983 they were 12.8 for both. Over those two decades the proportion of the labor force who were college graduates nearly doubled, from 9.7 to 18.4 percent for women, and from 11.9 to 23.1 percent for men. Table 3-8 shows the educational attainment of the labor force in 1983 by sex, race, and Spanish origin. Fewer white women had completed college than white men, and their median attainment was 0.1 year less than for men. Both black and Hispanic women had higher attainments than the men in their groups, and both were more likely to have completed high school or college or to have had some college education. Overall, however, the college attainment of mi- nority workers in 1983 was similar to that for whites in the 1960s. Despite the equality in years of schooling for men and women in the labor market, there are substantial differences in courses of study. In recent years there has been some convergence. Table 3-9 shows the change in the percentage of bachelor's degrees awarded to women in selected fields between 1966 and 1981. The percentage increased in every field except home economics, and at

78 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-8 Educational Attainment of the Labor Force by Sex, Race, and Hispanic Origin, 1983a Years of School Completed Fewer than 4 years of high school ( % ) Whites Blacks Hispanics Males Females Males Females Males Females 4 years of high school only ( %) 1 to 3 years of college (%) 4 or more years of college (%) 21.1 16.9 37.6 44.9 17.9 20.0 23.4 18.2 34.0 39.8 15.9 10.3 25.3 43.5 19.5 11.7 48.1 41.0 30.4 35.9 13.5 14.5 8.0 8.6 Median school years completed 12.8 12.7 12.4 12.6 12.1 12.3 aData are for workers aged 16 and over. SOURCE: Blau and Ferber (1986:Table 7.2). Reprinted with permission. least doubled in six fields: in agriculture 10 times, in architecture 4 times, in business 4 times, in computer and information science 2 times, in engineering 25 times, and in economics 3 times. Nevertheless, women are still substantially less likely than men to earn degrees in the physical sciences, engineering, or architecture. Although their distribution across fields has become more similar to men's, it still differs significantly. The index of segregation across 22 fields TABLE 3-9 Percentage of Bachelor's Degrees Awarded to Women by Discipline, 1966 and 1981, Selected Fields Discipline 1966 1981 Agriculture 2.7 30.8 Architecture 4.0 18.3 Biological sciences 28.2 44.1 Computer and information sciences 13.0a 32.5 Education 75.3 75.0 Engineering 0.4 10.3 English and English Literature 66.2 66.5 Foreign languages 70.7 75.6 Health 76.9 83.5 Home economics 97.5 95.0 Mathematics 33.3 42.8 Physical sciences 13.6 24.6 Psychology 41.0 65.0 Social sciences 35.0 44.2 Economics 9.8 30.5 History 34.6 37.9 Sociology 59.6 69.6 All fields 39.9b 49.8 aData are for 1969, the earliest year available. bIncludes first professional degrees. SOURCE: Blau and Ferber (1986:Tables 7.3 and 7.4). Reprinted with . . permission.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 79 was 35.8 in 1981 (calculated from data in Women's Bureau, 1983:Table IV- 13~. This index can be interpreted as meaning that 35.8 percent of women or men would have to change fields for women and men to be identically distrib- uted across fields. The courses of study taken in high school also differ by sex. Among students enrolled in vocational courses, for example, girls are much more likely to take business courses (e.g., typing), and boys are much more likely to take craft and industrial arts courses (National Center for Education Statistics, 1984: Table 3.6~. Equal proportions of high school boys and girls take first-year algebra, but girls are somewhat less likely to take higher-level math courses: for example, 10 percent of boys but only 6 percent of girls take precalculus or calculus (Na- tional Science Foundation, n.d.:Table II-9. And at two-year colleges (in 1978), male students outnumbered female students 10 to 1 in engineering, 3 to 1 in agriculture and natural resources, and almost 2 to 1 in physical sciences; women outnumbered men 12 to 11 in the biological sciences (National Science Foundation, n.d.:Table II-131. The proportion of advanced degrees awarded to women has increased sub- stantially since 1971. The proportion of first professional degrees awarded to women doubled between 1971 and 1975 and again to 1981, when women earned 27 percent of the degrees awarded (14.4 percent in dentistry, 24.7 in medicine, 42.6 in pharmacy, 35.2 in veterinary medicine, 32.4 in law, and 14.0 in the theological professions). With respect to other advanced degrees, women in 1981 earned 50.3 percent of the master's degrees awarded and 31.1 percent of the doctor's degrees. In sum, although women and men have equal years of education on average, men are more likely to take mathematics, engineering, and physical science courses. Men are more likely to have advanced education. These differences are declining, however, as women's participation in college and graduate work has increased and their fields of study have changed. Overall, minorities have less education than whites, but this gap, too, has been declining. THE POTENTIAL EFFECTS OF TECHNOLOGICAL CHANGE THE INFLUENCE OF LABOR SUPPLY The data reviewed above suggest that the labor force will continue to grow, although at slower rates than in the past. First, the population base from which the labor force will be drawn will experience slower growth. Most of the baby boom has entered the labor force, and the new replacement cohorts are smaller. Second, labor force participation rates for men are expected to continue their slow decline. Third, labor force participation rates for women, although they will continue to increase, are projected to do so at a slower rate. Although the rate of entry is expected to slow for women, women may be more likely to work

80 COMPUTER CHIPS AND PAPER CLIPS full time. There will also be some shift in the composition of the available labor supply, with women becoming relatively more numerous in the labor force, and young people and older men becoming less numerous. Minorities are also ex- pected to become a larger proportion of the labor force, since their birth rates are generally higher, and their current populations are younger. In addition, immigration is likely to continue to contribute disproportionately to the minor- ity population and to result in the entry of predominantly working-age people. These anticipated changes in the supply of different groups of workers to the labor market are likely to have effects on their average wage rates, because they have not in the past been seen as substitutes for one another. As noted in Chapter 1, there is substantial sex segregation in the labor market: many jobs are pre- dominantly female, others predominantly male. There is also segregation by race and by age. Changes in the supply of women workers tend to affect wom- en's wages relative to those of men because such changes tend to affect the supply to some occupations more than others. As the availability of the usual labor supply for a job changes, however, employers may change their hiring practices or make other adjustments in their production process. The projected decreased rate of growth of female labor supply could be ex- pected to contribute to a higher female-to-male wage ratio. In addition to the demographic reasons for a decreased supply of women workers noted above, changing attitudes and increased organization among women workers are likely to be important. Fewer women may be willing to work for low wages; they may- demand full-time work, equitable wages, and seniority increases for their longer job tenure. Higher wages will certainly benefit women, but it is impor- tant to keep in mind that, in the long run, a higher price for female labor could in turn lead to substitution; younger people of both sexes, male immigrants or minorities, or automation might become feasible substitutes for adult, native- born, white women workers. Or, if the products that women workers produce become more expensive because of higher wages, consumers of those products might substitute cheaper products. Whether or how such substitution for women workers would occur is uncertain, but it is certainly one possible effect of supply changes that tend to raise the wages of women workers. If economic growth remains at roughly current levels and widespread tech- nological displacement does not occur, the employment prospects for adult women appear good. Some establishments that hire young workers are already complaining of shortages of labor. The shift in age and sex composition of the labor force suggests that additional shortages may be perceived before 1995, especially if economic growth increases. If economic growth slows or if tech- nological change causes significant displacement of women workers from their jobs, however, serious unemployment for women may occur. Because of the variety of interrelated factors influencing women's labor force participation, women are unlikely to withdraw from the labor force simply because their jobs

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 81 have been eliminated or new jobs fail to appear. Rather, a substantial number of women, like men, will most probably remain in the labor force as unemployed workers. Although "discouragement" leaving the labor force because of lack of jobs, rather than lack of desire to work is more pronounced among women, it is not likely to be large enough to alter this picture. For a smooth adjustment to large technological displacements, one would have to look to the job vacancies being created by the shrinking youth labor force, by the trend toward early retirement of older male workers, and by the increased availability of "men's" jobs in response to equal employment opportunity programs. Or one must hope that the change will occur slowly enough that it can be relatively easily accom- modated. A well-educated work force is clearly better able to respond to occupational shifts in the demand for labor. People who have completed a solid curriculum in high school will be able to learn new skills on the job and benefit from retraining (see Committee on Science, Engineering, and Public Policy, 1984~. Those with scientific and technical training will also be able to enter rapidly expanding fields with such jobs as computer systems analysts, programmers, and opera- tors; electrical engineers; and electrical and electronics technicians. These jobs, although not among the largest occupations, are among the top 40 occupa- tions in the number of new jobs they are expected to create by 1995 (Committee on Science, Engineering, and Public Policy, 1984~. To the extent that workers lack basic educational preparation and technical and scientific background, they are likely to be less able to respond to forthcoming changes. As noted above, in some aspects women and minorities lag behind majority men in their educational preparation. THE DEMAND FOR WORKERS As noted in Chapter 1, technological change can affect the demand for work- ers in several ways. First, new technology may cause an increase in labor productivity, making it possible to produce the same output with a smaller number of workers. As a consequence, fewer workers may be demanded than previously. As noted in Chapter 1, technological change permitted the nation's food supply to be pro- duced by an ever-shrinking proportion of the work force. Consumers and the economy as a whole benefited. A large proportion of the work force left food production, while employment in manufacturing and service jobs increased. Economic output grew and per capita income increased. For individuals caught up in this transition, however, the costs of adjustment were often high. A decline in employment because of increased labor productivity is not inev- itable, however. By facilitating such increases, new technologies decrease the cost of producing the product and thus lower its price. The lower price in turn

82 COMPUTER CHIPS AND PAPER CLIPS may encourage consumers to demand more of the product. In addition, techno- logical change may allow for the provision of a higher-quality product or it may stimulate the introduction of new products, increasing consumer demand for the output of the industry. In the case of food production, the increase in labor productivity was very large relative to the possible increase in demand for food by consumers. Furthermore, the new products that resulted- canned, frozen, processed, and packaged foods were not produced by farm workers. Thus, an enormous decline in the proportion (as well as the absolute number) of farm workers occurred. In other industries, aggregate demand and newly created products and services are likely to be more favorable to employment levels. As an example, the adoption of word-processing equipment may indeed in- crease labor productivity, permitting the existing work to be accomplished by fewer workers. But its introduction is also likely to create new work. The capa- bilities of microprocessors are ideally suited to many revisions, more personal- ized form letters, updated statistical reports, and more charts and graphs, all of which can be produced more cheaply than previously. As a consequence, the production of a firm becomes more intensive in information content, a trend that has been going on for many years, and the decline in employment, to the extent it occurs, is considerably less than if the new technology were used simply to produce the previous output. Thus, while a major impetus behind the adoption of new technologies is the increase in labor productivity that they permit, their impact on employment levels is difficult to forecast. A second potential impact of the application of new technologies is that they can facilitate the reallocation of functions among occupations, resulting in ei- thera reduction or en increase in the demand for a particular type of worker. For example, the greater ease of data entry and retrieval using the new computer technologies may encourage employers to shift some of these tasks from cleri- cal workers to other workers or even to consumers (as in banking). The many capabilities and ease of use of the latest word-processing equipment and soft- ware may also result in professional and managerial workers doing more of their own "typing." Such developments would reduce the demand for clerical workers. In the opposite direction, however, the application of computers can make routine some decisions that formerly required broad-based knowledge and judgment, so functions previously performed by managerial or profes- sional workers would be transferred to clerical workers, increasing the demand for their services. For example, in the insurance industry, rating and underwrit- ing are now often done by computers using standardized decision algorithms: a clerical worker can enter the data, describing the client and his or her needs, and, if there are no special conditions, issue a policy. A third impact of the introduction of new technologies is its tendency to alter the skills needed to produce the former (or new) output. It may render specific skills obsolete, or nearly so, and greatly increase the demand for others. Such

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 83 effects are evident when looking at specific occupational categories, such as stenographers and computer operators. Between 1972 and 1982, when clerical employment overall rose by 28.8 percent, employment of stenographers de- clined by 47.2 percent (Hunt and Hunt, 1985a). This decrease was caused by the miniaturization of dictation equipment and improvements in magnetic tape technology combined with growing consumer acceptance of new methods. When dictation equipment is used, the material is entered by the clerical work- ers directly into the typewriter or word processor, and the shorthand skills of the stenographer are no longer needed. The decline in the number of stenographers is not a reflection of a decline in the amount of dictation being done indeed, the- amount of dictation appears to be increasing but the particular skills of stenographers are no longer needed to accomplish this task. Over the same 10- year period, the employment of computer and peripheral equipment operators increased 195.5 percent. This increase was clearly fueled by the diffusion of the new computer technologies. It should be noted, however, that changes in skills needed to accomplish specific tasks do not always translate into changes in occupational employment, because single tasks do not often predominate in jobs. Thus, rather than reducing employment of secretaries, automated word processing may simply provide them with more time for other work. In sum, the introduction of new technologies may result in increases in the productivity of affected workers, reallocation of functions among occupations, and changes in the skill requirements of jobs. For the reasons outlined above, it is difficult to predict the consequences of these processes on employment levels of workers in any given occupational category. Moreover, the factors may oper- ate together to make such forecasts even more uncertain. So, for example, with word-processing technology, it may be possible to generate frequent statistical reports by simply entering a few updated numbers. While some new skills will need to be learned, others will not be needed, and, overall, the greater ease of production may create a demand for additional statistical reports. But it is not certain that those reports will be generated by clerical workers. It is useful to keep in mind that these effects of technological change on the demand for workers in different occupations may affect the wage rates of work- ers as well as their employment levels. In the long term, changes in wage rates can influence employment levels. At issue here is the fact that different effects on different occupations are likely to affect women and men differently, chang- ing the female-to-male wage ratio. UN EM PEOYME NT Of the three major types of unemployment "frictional" (the expected un- employment that would normally result from workers' job changes and new entrants), "structural" (the unemployment that remains at the peak of the econ-

84 COMPUTER CHIPS AND PAPER CLIPS omy's upswings, and is not frictional, being caused by changes in the structure of demand), and "cyclical" (which comes and goes with the ebb and flow of the business cycle) structural unemployment in particular has troubled observers since the late 1960s. Sources of structural unemployment, caused by imbalance between the types and locations of available employment on the one hand and the qualifications and locations of workers on the other hand, include skill mismatch, geographic mobility and immobility, discrimination, and institutional barriers to employ- ment. When structural factors exist, both unemployment and shortages occur, despite a balance between supply and demand in the economy as a whole. Technological change probably contributes most to the first factor, but may contribute as well to the importance of the second if it increases the mobility of capital. It is not, in any case, the sole cause of structural unemployment. "Displaced workers" are those who are laid off more or less permanently; theirjobs may have disappeared because of technological change, competition, or decreased demand for the products they make. Displaced workers and new entrants to the labor market may or may not be considered structurally unem- ployed, depending on how quickly and easily they can find new employment. Because the adoption of new technologies almost invariably results in shifts in demands for different types of workers, some displacement of workers is likely to occur. Even if, due to counterbalancing effects, broad categories are relatively unaffected, specific occupational groups are certain to be affected. The consequences of such shifts in occupational demand for unemployment among workers in the affected groups depend primarily on the magnitude of the change, the speed with which it occurs, and the flexibility and skills of the workers involved. Relatively small changes in demand or those that take place slowly are likely to cause little unemployment. At the level of firms, normal attrition may be sufficient to reduce the size of the work force in the affected occupations. At the level of the labor force, attrition may also operate as older workers retire from declining occupations and are not replaced by new entrants. Relatively large changes, however, especially when they occur over a rela- tively short period of time, can cause more serious problems. Workers may lose their jobs and find themselves without the skills necessary to obtain new jobs. Thus, the adoption of new technologies has the potential to create groups of workers who are structurally unemployed. The magnitude of this problem de- pends not only on the extent and timing of shifts in demand but also on the resources devoted by firms, as well as the public sector, to retraining and plac- ing those workers. Workers can be encouraged to move where there are jobs; they can be reeducated, retrained, or retired; capital movement can be con- trolled; discrimination can be reduced; and institutions can be modified. Increased structural unemployment, should it occur, can make it difficult for the economy to achieve desired low rates of unemployment at acceptable rates

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 85 of inflation: if there is a growing pool of workers who lack the necessary skills for the available jobs, increases in aggregate demand will rapidly generate shortages of qualified workers. As the wages of those workers are bid up, labor costs and thus prices rise. This phenomenon may be an important factor in the rising trend of unemployment over the past two decades. The policy priority of the government has been on reducing inflation, which, given larger structural shifts, causes unemployment to increase. If structural unemployment in- creases, a further upward revision in the government's target rate of unemploy- ment is likely to occur. Several observers have noted that, since 1969, the economy has experienced increased unemployment (e.g., Podgursky, 19841. At the peak and the trough of each of the four lousiness cycles between 1969 and 1982, unemployment has been greater than before. At the peak of the first cycle (in the third quarter of 1969), the unemployment rate was 3.6 percent; at the peak of the fourth cycle (in the third quarter of 1981), it was 7.4 percent. Similarly, at the trough of the first cycle (in the fourth quarter of 1970), unemployment was 5.8 percent; by the trough of the fourth cycle (in the fourth quarter of 1982), it was 10.6 per- cent. In his analysis, Podgursky (1984) finds that the large number of young people and women who entered the labor market in the early years (early 1970s) contributed to the increase, and that in later years (late 1970s, early 1980s), male unemployment in blue-collar jobs contributed disproportionately to the increased unemployment. Of special interest is that in the most recent trough the share of unemployment attributable to clerical occupations grew. Most im- portantly, Podgursky finds that the secular increase in total unemployment al- most certainly points to rising structural unemployment and a probable diff~- culty in absorbing the employment effects of technological change. Although he notes that the evidence also suggests there has been some slackening in the strength of aggregate demand over time, the increased structural unemploy- ment suggests a need for specifically targeted employment programs; changes in monetary and fiscal policies that affect aggregate demand are not likely to reach all the structurally unemployed. There is nothing immutable about the residual unemployment left at the peak of recovery or about the increasingly high levels of unemployment at the trough of the cycle. Both are susceptible to structural solutions, and when aggregate demand is strong enough (as in war- time), structural solutions become worthwhile (for example, child care pro- grams in World War II; training programs during the Vietnam War). Important current structural factors include skill mismatches between the available workers and the available jobs, geographic imbalances (workers seeking jobs do not live where jobs are available), differences in expectations between workers and employers (unemployed workers do not want the avail- able jobs because of their working conditions; employers do not want to hire available workers because they believe them to be unqualified or because they

86 COMPUTER CHIPS AND PAPER CLIPS are prejudiced), and other barriers to entry or movement. From a structural viewpoint, the displacement of women may be especially serious because women workers have not often been viewed by employers as interchangeable with men and because they tend to be less mobile geographically and occupa- tionally. Although it has not been established that the new computer-based technolo- gies are associated with this worsening problem of structural unemployment, the increased contribution of clerical work to unemployment in the last cycle raises the possibility that they might be involved. Moreover, information tech- nologies can be expected to have greater effect in the future as their use be- comes more widespread. As noted earlier, the predominantly female clerical occupations are likely to be the most affected. RECENT TRENDS IN CLERICAL EMPLOYMENT OVERALL GROWTH While forecasting inevitably holds many uncertainties, the best starting point for trying to predict future clerical employment is to review recent trends. This section and the next one on future employment are based largely on work done for the panel by H. Allan Hunt and Timothy L. Hunt (1985a).2 The tremendous historical growth in clerical employment is illustrated in Figure 3-1, which shows that the proportion of clerical workers to total employment has doubled in the last 40 years. In 1940 just under 1 employee in 10 was a clerical worker; by 1980 this proportion had risen to nearly 1 in 5; the number of clerical work- ers increased from 5 million to 20 million (Hunt and Hunt, 1985a). Will this trend continue? Both the increasing unemployment evident in the past four lousiness cycles and the rapid pace of the technical developments in the office described in Chapters 1 and 2 cast doubt on future growth. The first "computer revolution" in the 1960s was also expected to decrease the need for clerical workers, but clerical employment continued to increase. Nevertheless, many observers are convinced that the new technologies will substantially re- duce clerical employment. Reductions in the cost of computing hardware, com- bined with the reductions in size made possible by microprocessor technology, facilitate the widespread use of the new equipment in many settings. These changes may actually constitute a revolutionary development. As Hunt and Hunt (1985a) note, some support for the position that the new computer revolution will stem the rate of growth in clerical employment is 2The data developed by Hunt and Hunt have been adjusted rather extensively for consistency (see Hunt and Hunt, 1985a, for a detailed explanation of the adjustments). Thus, the figures re- ported here do not correspond exactly with published Census Bureau figures.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 0.20 0.18 0.16 O 0.14 o o a: J G UJ J 0.12 0.10 0.08 0.06 0.04 0.02 o 87 1940 1950 1960 1970 1980 YEAR FIGURE 3-1 Decadal growth in clerical employment as a proportion of total employment, 1940-1980. SOURCE: Hunt and Hunt(1985a:Figure 1-A); teased on 1940-1980 censusdata. found in the apparent reduction in the rate of increase of clerical workers as a proportion of the total labor force. While the clerical proportion rose almost linearly from 1940 to 1970, there is a slight reduction in the rate of increase between 1970 and 1980 (see Figure 3-11. Is this the beginning of the end of the growth of clerical employment? The question cannot yet be answered. In addi- tion to the relatively small number of observations for the post-1980 period, the data for those years are influenced by the major recession of 1981 to 1982, the deepest in terms of unemployment since the 1930s, and the recovery from it, making it hard to separate short-run cyclical influences from long-run trends. In past recessions the proportion of clerical employment tended to increase since employers usually cut back less on clerical workers than on other categories of employees. Significantly, this effect did not hold in the 1981-1982 recession. Overall, as may be seen in Figure 3-2, clerical employment as a proportion of total employment actually declined slightly in the early 1980s, following the above-noted reduction in the rate of increase in the proportion of clerical em- ployment during the 1970s. While the available evidence is consistent with the hypothesis of slowed growth in clerical employment, however, there is little in past trends that would support a projection of catastrophic decline in this occu- pational group in the near future.

88 0.19 0.18 o - O 0.17 o CC J ~ 0.16 CC J c: 0.15 _ 0.14 COMPUTER CHIPS AND PAPER CLIPS A: . l l 1 I 1 1 1 1 1 1980 1 984 1960 1964 1968 1972 1976 YEAR FIGURE 3-2 Annual changes in clerical employment as a proportion of total employment, 1958-1984. SOURCE: Hunt and Hunt (1985a:Figure 1-C); based on 1958-1984 CPS data. OCCUPATIONAL SHIFTS WITHIN CLERICAL WORK Within the clerical category there has been considerable variation over the past 30 years in how specific occupations have fared. Table 3- 10 ranks clerical occupations by the rate of change in employment levels from 1950 to 1980. The rate of increase in this period was by far the largest for computer and peripheral equipment operators, although this occupation grew from a vein small base, 868 people in l9SO. By 1980 there were nearly 400,000 computer and periph- eral equipment operators, making it the thirteenth largest clerical occupation. Teachers' aides were the second fastest growing clerical occupation over the 30-year period; their numbers also grew from a fairly small base, 6,000 in 1950, to more than 200,000 by 1980. The third fastest growing clerical occupa- tion was typists, although there was actually a significant decline from 1970 to 1980; the rapid growth of the occupation from l9SO to 1970 offset the recent decline. Following in order of rate of growth were Libras attendants, clerical supervisors, bank tellers, receptionists, and cashiers. With the exception of the computer operator category, the jobs that grew rapidly do not appear to be related to the use of new technologies. It is also important to recognize that very large categories such as secretary provided a high proportion of new jobs over the period, even at lower overall rates of growth.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIMS 89 A few clerical occupations showed declines between 1950 and 1980. The most rapid declines were among stenographers and telegraph operators; both of these categories were affected by technology. Demand for telegraphic services has fallen off as cheaper, more convenient long distance communication has become available. As discussed above, the use of dictation equipment has con- tributed to the decline of stenographers. Yet the functions of long distance communication and of dictation and transcription are, if anything, increasing, assisted by new equipment and different kinds of workers. Tabulating machine operators and weighers also declined substantially. As Hunt and Hunt (1985a) note, tabulating machine operators illustrate how a technology-specific occupation can experience a spurt of rapid growth and then an equally sudden decline. Tabulating machines were developed in the 1950s for analyzing data on punched cards, and their use grew rapidly. Between 1950 and 1960, the number of tabulating machine operators nearly tripled. But the adoption of new technology with more advanced capabilities reduced the use of tabulating machines, and the number of employees fell by nearly 90 percent after 1960. Other declining occupations between 1950 and 1980 were those of messengers and office helpers, calculating machine operators, and telephone operators. These declines appear to be related to the availability and use of new technologies. DEMOGRAPHIC TRENDS IN CLERICAL EMPLOYMENT The differential rates of change among the different clerical occupations have affected women and men and minority and majority women somewhat differently because of their different employment patterns in those occupations. The fastest-growing occupations between 1950 and 1980 (see Table 3-10) are generally among the most predominantly female of the clerical occupations: typists, receptionists, teachers' aides, bank tellers, and keypunch operators. One of the rapidly growing occupations over the period insurance adjusters, examiners, and investigators although still more male (44 percent in 1980) than clerical occupations as a whole (22 percent), has experienced rapid in- creases in the percent female: overall employment in the occupation grew at a rate of 5.4 percent per year, while female employment grew 16.1 percent per year (Hunt and Hunt 1985a:Tables 2.9, 2.15~. In contrast to these rapidly grow- ing occupations, many of the clerical occupations in which males are relatively overrepresented have experienced average or below-average growth: utility meter readers; messengers and office helpers; shipping and receiving clerks; ticket, station, and express agents; and postal office mail carriers. Indeed, over the period as a whole, females in all clerical occupations increased from 61 to 78 percent. Minority representation in clerical occupations also increased over the pe- riod: CPS data from 1972 to 1982 show an increase from 8.7 to 11.8 percent

9o to of oN o to ~4 As .> Id ce of Do - ~o to - c~ o c) o 'e .c - - o O Ed ~ lo c) ~ o ox E o E ~ ~ V) o ~ o Go ~ V) ~ ~ ~ ~ oo Do Do ~ ~ ~ ~ _ o . . . . . . . . . . . . . . . . . . . . . C~ ~ ~ ~ ~ ~ ~D ~ — cr~ _ — oo ~ ~ - , — oo x O ~ ~ O ~ ~ ~ X ~ ~ (~N C~ ~ oo O~ ~ CrY _ C~ _ _ _ ~ ~ o o ~ ~ ~ o o ~ ~ ~ ~ C~ ~ ~ oo ~ t— — - ] d. ~ `'0 oo - , — - V~ oo o oo — ~) ON C'-] oo ~ oo ~ ~ — ~ oo ~D ~ ~ ~ ~ ' — ~ oo _ _ _ ~ O ~ _ _ _ ~ ~ 0 ~ — ~ ~ ~ — ~ <5~ C~) [~ ~— O~ ~S} ~ \o ~) =\ ~} ~ oo C~ O — oo Ch ~ ~ ~ — ~ t— ~ ~ ~ t— t— —) —) — oo k.~ ~ ~ oo {~\ oo — ~) ~ ~ — 0 ~ ~o ~ oo 0 oo ~ oo oo ~ ~t =~ — —) C~ ~ ~ ~ X 0 ~ ~o ~ oo ~ ~ ~ ~ ~ ~ _ —) d —) — ~ ~ oo oo ~ O O ~ c~ ~ oo — C~ — — O — — ~ ~ oo ~ ~ — ~} C~ oo cM — — — _ r~ ~ 0 ~ 0 oo O oo oo ~ C~ — d ') _ _ _ 0 ~ ~ O Q ~ os o ~ ~ ~ ~ ~o ~ X oo ~ _ _ - , _ oo ~) ~t ~) oo q ~, ~ oo — — ~ ~ x ~o 0 ~ ~ ~ ~ ~ ~ ~ a~ ~ ~ — oo oo — ', ~ ~ ~ ~ ~ ~ O O O ~ ~ ~ ~ ~ ~ U~ ~D ~ ~ ~ r~ ~ ~ t~ ~ ~ o~ — ~ _ ~t c~ Q _ c~ _ O U~ ~ O O ON O C~ ~ oo cr — ~ u~ ~ ~ ~ ~ ~— ~ — `~o — — ', - oo cr ~ ~ ~ ~ O ~ ~ v~ ~ ~ ~ ~ 0N ~ ~ ~ ~ ~D u~ ~ ~ - ~ ~ u~ o - ~ - ~ 'c} o - - o o 04 ~ . - c) ;> o o c c c ~ c) ; ° a u | · __ O ~ ~ ~ ~ ~ ~ c · E = c, ~ ~ ~ ~ ' ~ ~ ~ ~ ' ~ ~ E .' ~ ' °

91 00 ~ ~ ~ ~ 0 ~ ~ ~ u~ ~ ~ ~ _ _ _ ~ ~ 0 ~ ~ a~ 0 . . . . . . . . . . . . . . . . — — . . . . . c~) (~~ c~} c—~ tr) —] ~ ~ ~ — — — — — o o ~ ~ ~ ~ ~ ~ ~ 1 1 1 1 1 1 1 oN ON — ~ ~ — — ~ ~ ~ ~ — o ~ ~ — ~ ~ ~ ~ - - ~ 00 - } — ~ C~ t— ~ C~ t— O v0 00 — O — O o0 ~ ~ — ~ O CN ON c~ ~ ~ ~ c~ oo ~ oN - O' - - ~ ~ - ~ oo ~'o c~ t— c - ] ~ ') ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ^ ^ ^ ~ ^ ^ ^ ~ ~ ~ ~ ~ ~o ~ r~ ~ ~ ~ 0 0 oo ~ u~ ~ oo — ~ ~ ~ ON ~ [_ _ _ ~ oo O ~ ~ oo oo ~ oO — ~ oo d — — oo C'I O~ ~ ~ ~ oo - ~ oo ~ ~ ~ ~ - - oo - oo t— ~ o ~ o oN ~ ~ o ~ - ~ ~ ~ ~ o o ~ ~ O ~N ~ 00 ~ ~ ~ — ~ ~D ~ ax ~ v~ ~ I,q _ oo ', t_ ~ ~ _ ~ ~t ~ o~ ~ ~ ~o oo ~ ~ ~ — — r~ 0 ~ ~o 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ^ ^ ^ ^ ^ ~ ~ ~ ~ ~ u~ ~ ~ ~ ~ ~ ~ ~ ~ oo o - ~ oo v~ ~ ~ - - oo ~ oO — ~ - } O ~ oo ~ ~D O ~ ~D ~ ~ ~ ~ ~ ~ _ M - - ) ~ - ON ~ O — ~ ~ ~ — ~ O ~ ~ ~ ~ ~ C~ ~ U~ c~ ~ ~ ax 0 ~ c~ r~ — ~ — 0 ~ — ~ — 0 ax c', — — ~ ~ ~ — — 0 — ~ oo ~ ~ ~ a~ - } C~ — ON ~ ~ ~ ~ 0 ~ ~ ~ ~ oo ~ oo ~1- Ir) tr) d- t— ~— — oo — 0 ~ ~ ~ _ _ _ — =\ —) ~.D ~ ~ ~ — - oo oo ~ ~ 0 ~ — ~ , ~ ~ ~ oo ~,o — 1— ~ ~ ~ oo ~ ~ — In ~ O ~ ~ ~ ~ oo u~ ~ — ~ o~ ~ — ~ ~ ~ 0 0 oo 0 ~ oo ~o — _ ~ ~ ~ O _ ~ _ CN 0N ~ ~ ~ O oo ~ O ~ oo ~ — ~ O ~D — ~ v~ ~ ~ oo ~ ', a~ oo c~ ~ ~ cr~ c~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ _ 0 ~ ~ ON O — ~ ~ ~ ~ — ~ oO ~ ~ — ~ oo ~ — ~D — oo - ) c~ _ _ _ r_ ~ _ _ - } C~ ~ _ - C~ ~ ~ & ~ ~ ~ 3 ; ~ ~ _ . e I c 0 ~ _ X .' O C) O O O Ce & U) O ~ ~ ~ & ~ ~ _ ~ O ~ c~ ~ ~ ~ m E~ ~ ~ Z ~ cq ~ m ~ ~ ~ ~ ~ E~ ~ v' c~ u, ._ c°, o c' ~' ce ct c~ 3 - ._ C) ~: o U' Ct D . ~ U~ ~ C~ Ct :e ·r~ C) oo C) — — 3 ~ ~ _ c' ~ 0 ce ~ _ ~ .. E~ zO O

92 COMPUTER CHIPS AND PAPER CLIPS (Hunt and Hunt, 1985a:Table 2.171. Some of those occupations with relatively greater minority representation grew relatively rapidly: for example, teachers' aides (20.1 percent minority in 1982), typists (17.4 percent), keypunch opera- tors (20.3 percent), and computer and peripheral equipment operators (15.5 percent). Other occupations, however, experienced very slow growth: stock clerks and storekeepers (16.1 percent minority), telephone operators (17.3 per- cent), messengers and office helpers (19.1 percent), and postal clerks (26.9 percent). Table 3-11 displays the numbers and percentages of several minority groups (blacks, Hispanics, and Asians) among female clerical workers in 1980 for detailed census occupational categories. Asian women, for example, are 1.6 percent of the female clerical work force, and they vary from 3.1 percent of transport, ticket, and reservation agents to 0.5 percent of female meter readers. Hispanic women, 4.6 percent of female clerical employment overall, range from 11.4 percent of female teachers' aides to 2.6 percent of female proof- readers. Black women, who are 9.2 percent of female clerical employment, are 32.9 percent of female postal clerks, but only 4.0 percent of bookkeepers and accounting clerks. Clerical employment growth has also varied by region and geographic area. Although a thorough analysis of such differences is beyond the scope of this report, several recent studies highlight geographic changes that are significant for women and minorities. For example, a recent study by 9-to-5 (1985) found that, like blue-collar employment, clerical employment in the Midwest has declined since 1980, associated with the recessions in 1980 and 1982. Unlike the nation, this region experienced actual declines, not only lack of growth. Surprisingly, clerical employment fell even more rapidly than total employ- ment between 1980 and 1983 in the four states (Illinois, Indiana, Michigan, and Ohio) that constitute the most industrial portion of the region. This study also found that the decline in clerical employment was especially large in center cities, such as Detroit and Chicago, while employment in the suburban areas of those cities either did not decline as much or expanded. The 9-to-5 report traces the declines in clerical employment in the Midwest generally to declines in government employment. A recent study by the New York regional office of the Bureau of Labor Statistics found that finance job growth had fallen off steeply between 1983 and 1984 in Manhattan: in 1983 the finance industries had added 10,900 jobs (and 16,700 in 1982), but in 1984 only 2,300 new finance jobs were created (New York Times, December 15, 1985:591. Technological change and the movement of jobs to the suburbs were cited as causes for the slower growth. In research on insurance companies in the San Francisco Bay area, Nelson (1983) found similar suburbanization of jobs occurring as back offices were relocated from center cities to suburbs. She concluded that land cost was only one factor in the decision to relocate; man- agement also sought a different labor force than that available in the cities. The

93 e O .= ~ ~ ~ Z A: C C ~ 'SIDE ~ =- O ~ 9C O C t: ~ ~ so V) At> AD .~- 3 c Cal ._ ._ ~ = c o .o o . U. o V~ ~ ~ _ ~ ~ o ~ ~ _ ~ ~ _ C~ ~ ~ U) ~ ~ oo ~ o ~ ~o ~ ~ ~ ~ ~ ~ oo ~ ~ oo ~ — 0 ~ a~ oo ~ oo o _ ~ ~ ~ _ ~ ~ ~ ~ ~ ~ ~ oo ~ ~ oo ~ oo ~ ', _ _ U~ ~ ~ C~ _ ~ o _ _ ~ o ~ ~ _ _ _ C~ ~ _ ~ - ) d- — . . . . . . _ oo U~ o oo ~ a~ c~ O — ~ ~ ~ U~ U~ o ~ ~ _ ~ — ~— oo oo ~ ~ d- ~ ~ ~ ~ \0 0 0N ~ °° °o 0N ~ ~ ~ — ~ — ~ ~ (5\ t— {~\ ~ t — ~1 ~1 o ~ — ~ ~ ~o ~ c~ u~ U~ ~ 0 C~ oo ~ 0 C~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ oo ~ o~ Q — oo — ~ ~ ~ oo ~0 — ~ ~ t— ~ ~ O ~ ~ t~ D ~1 O ~ ~ ~ ~ ~ — O ~ — ~ ~ ~ _ - _ u~ 0 ~ oo _ ~ ~ ~ t~ a~ ~ r~ ~ u~ ~ ~ ~ ~o O ~ — oo x . . . . . . . . . . . . . . . . . . . . . . . . ~ ~ ~ ~D O ~ ~ ~ ~ ~ ~ — ~ 0N ~ ~ O ~ oo ~ ~ O ~ ~ oo oo oo oo ~ oo oo oo oo oo oo ~ oo ~ oo oo ~ oo oo oo oo ~ oo oo \~0 ~ t— ~ ~ — ~ ~ oo ~ O ~ — ~ ~ ~ ~D — CN O oo ~ ~ C~ ~ ~ ~ C~ ~ — ~ 00 ~ d oO — — C7~ ~— —) ~.C~ O ~ oo ~ oo ~ ~ ~ u~ ~ ~ ~ ~ ~ ~ oo ~ ~ 0 ox O ~ O O Ch ~ ~ ~ ~ ~ ~ ~ O ~ ~ ~ ~ oo ~ ~ O — ~ — c~ O oo — — O \0 ~ ~D oo ~ ~ ~ ~ _ _ c~ ~ ~ ~ _ _ ~ ~ ~ ~ ~ _ _ oo 0 ~ ~ oo ~ ~ — ~ ~ ~ O ~ ~ - ) — oo - o ~ . - ~ (e ~o u, ~ ~ ~ o ~ ~ c~ ~ o o ~ .= '53='e'~et~'!~=La.2au.'~n E ~ ~ ~ ° ~ ~ ~ ~ c ~3 ~ ~ o °~ ~ ~ ° E ~ ~ °-, c ~ = E E ~ ~ ~ ~ c E ~ o ~ ~ ° == V~ V) CQ C) V~ o c) ~ ~ V) (Q ~ ~ S ~ E~ P:: S o ~ ~ O ~Q ~ V) ~ Z Ir) d- — oO oo — Ct~ — c~ ~D — O ~ ~ — _ _ ~ ~ ~ o ~ ~ ~ ~ ~ oo ~ ~ o t— — —} oo ON ~ oo oo ~ ~ ~ — oo ~ —, ~, o — ~o ~ C~ ~ ~ oo ~ ~ — ~ ~ ~ ~ o —, ~ ~ _ [~ ~ ~ o ~ U~ ~ ~ ~ — — o ~ ~ U) oo C) 04 e ~: o . - ce c) c) C: ~ . - ) — ~ o ~ ~: ~ u, . ~ . - ~ ~ ~ ~ r s se

94 O AN ~ \0 ~ ~ ~ 00 0 ~ Cal ~ ~ EN ~ ~ ~ 00 00 r - , ~ A, _ _ _ _ _ ~ ~ ~ ~ _ _ o o ~ o _ — o — — vat 0 — ~ ~ ~ ~ Do Go ~ ~ ~ us ~ vat cry Do ~ ~ Go 0 vat — Q ~ ~— Q ~o ~ ~ ~ ~ o ~ v~ ~ — ~ ~ ~ t~ o ~ ~ o oo ~ ~ t~ ~ ~ ~o ~ ~ 1~ ~1 ~ ~1 d. ~ ~ ~ ~ — — — ~ ~ ~ ~ _ oo ~ o ~ ~ o ~ ~ ~ ~ ~ ~ ~ ~ ~o — . . . . . . . . . . . . . . . . U~ ~ ~ V) ~ ~ ~ ~ ~ ~ U~ ~ V~ ~ ~ ~ ~. O ~ ~ ~ — ~ — O — ~ O — ~ ~ ~ ~ V~ ~ ~o ~ oo o C}~ ~ — ~ U~] ~ ~ 0N oo ~ t— ~ d O ~ oo oo — ~ ON ~ ~ q o ~ o ~ V) — V~ ~ ~ ~ _ ~ _ o ~ o o C~ ~ ~ ~ ~ ~ ~ C~ ~ =~ ~ ~ — _ ~ ', _ _ — O ~ ~ ~ ~ — ~D ~ ~ ~ ~ ~ ~ ~ t— ~ ~ ~ oo ~ ox ~ ~ O ~) t— ~ ~) C-1 43$ — — t— — ~ ~D O oo u~ ~ ~D — _ - , — Ch~ ~ — — O ~ C~ ~ ~ ~ ~ oo ~ — — oo ~ U~ ~ t— U) ~ ~ ~ . U) oo oo O ~ ~ — ~ ~ O 0N - , oo oo _ _ ~ C~ U~ - , ~ ~ \0 0 — — ~ O ~ — ~ ~ O ~ ~ ~ ~ ~ ~ U~ ~ _ ~ . . . . . . . . . . . . . . . . . . . . . . . 00 ~ ~ ~ 00 00 ~ 00 ~ ~ ~ — ~ ~ ~ ~N ~ 00 00 ~ 00 ~ ~ 00 00 00 00 ~ ~ 00 ~ 00 00 00 ~N ~ ~ 00 ~ 00 ~ ~ 0 oo ~ 0 0 ~ ~ v~ a~ oo ~ ~ 0 0 — 0 ~ ~ 0 r~ 0 ~ ~ — c~ ~ a~ oo ~ ~ c~ oo ~ ~ ~ ~ C~ O ~ ~ ~ ~ ~ ~ ~ ~ oo t_ _ _ ~ ~ ~ ~ _ ~ o co ~ ~ oo ~ ~ \0 ax ~ ~ — 0 ~ ~ ~ — ~ — ~ — oo ~ _ _ ~ ~ _ _ _ _ oo oo ~ _ ~ — O \0 ~ oo O oo ~ ~ . ~ ~ ~ _ 0N ~ O ~ 0N ~ ~ O — ~ ~ ~ O oo oo oo — ~ \0 — C~ ~ — ~ ~ ', ~ _ ~ ~ ~ —, — ~ oo ~ oo ~t <~ ~ =, 0\ — ~— C~ °O 1— ~l ~ oo ~ ~l ~ 0 ~— 0 ~ ~ ~ ~ — ~ t~ ~ 0 ~n — ~ ox ~ ~ _ ~ _ o ~ ~ oo ~ 0c — ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ O `~C~ — ~ O 0N ~ ~ — _ ~ _ ON ~ _ _ _ _ -° ~° ~ °, _~u ~ ~ ~ - ~ ~ E ~ y ._ ~ ~ C) o ~ C) o ~ 3 .= . ~ ~ ~: . O ~ ~ C) ce Ct cq ~ _ · _ C~ C.~ · ~ · ~ ~ =

9s Cot — O ON ~ 00 00 . . ~ . . . . . - ~ ~ - - ~ - — ~ or at — c - ~ ~ — — ~ ~ ~ oo ~— ~ — ~ ~t ~ _ ~ t_ Do ~ ~ ~ ~ Do ~ at 0 ~ rot Do - ) ~ ~' ~ ~ ~ ~ He ~ x ~ ~ en - ~ ~ - Do - to ~ ~ - ~ - ~ ~ ~ ~ ~ ~ ~ ~ ~ oo - - ~ ~ ~ ~ oo ~ - Do ~ - en ') ~ ~ - to ~ ~ to . . . . . . . . . ~ An ~ ~ to v) ~ ~ ~ ~ ~ ~ ~ oo ~ oo ~ ~ ~ ~ ~ ~ ~ ~ N ~ He ~ ~ en 00 ON ~ 00 ~ cr oo No Do ~ ~ oo Do ~ ~ ~ ~ ~ ~ ~ O ~ of ~~ ~ {3~ Rio ~ ~ Do ~ O~ ~ — t~ ~ — ID ~ — ~ Do ID t - - ~ ~ oo - ~ - ~ - oo '] ~ =\ O~ ~ oo - ~ ~ ~ t 0 —-= ~ ~ _ ~ ~ v~ ~ r~ — ~ ~ ~t ~ — ~t _ - ] - ~ t— - ~ oo O O — ~ — 0 — u~ ~ — ~ — 00 c~ 0 ~ \0 ~ ax ox ~c7 ~ 0 ~ ~ ~ u~ ~ _ ~ _ 00 c—~ ~ ~ ~) ox a~ (~} C'I C'I =\ ~o ~d 0 cr~ — — 0 ~ r~ — ~ 0 — ~ ~ oo ~ ~ ~ — ~ — 0 V) ~ ~t ~ — C~ —) —) ~ ~ CN —] =~ c~ ~ ~ 1— C~ — t— ~ — ~\ vo — — ~ ~ _ _ ~ ~ ~ ~ _ - ) _ _ oo C~ _ ~ o~ 0 _ ~ oo _ . . . . . . . . . . . . . . . . . . . . . . . ~ ~— ~ ') d - . — x In 0 ~ ~ ~ ~ ~ ~ ~ oo O r~ ~ ~ ~ oo oo oo oo oo oo oo oo oo oo oo oo oo ~ oo oo oo oo ox ~ oo ~ oo oo ~ ~ _ — ~ — ~ oo ~ 0 C~ oo ~ ~ ~ ~ ~ ~ oo ~ ~ — 0N — c~ ~) oo ~t q- ~ ~ ~ ~) oo ~ ~ ~ — ~ ~ d CJ~ — °o oo — ~ ~ ~ ~ ~ ~ 0 0 oo oo ~ ~ U~ ~ _ oo ~ ~ ~ 0 _ t_ ~ ~ - } _ _ ox ~ ~ ~ ~ ~ ~ CS~ ~ — Ox =, ~ 1~ ( 0 c%] ~d C~ \~} oo ~ — ~ ~ ~ 0 _ ~o oo _ c~, _ c-, _ ~ c~, _ c ~ — 8 ~ ~ ~ °° ~ ~ °° ~ ~ — ~ ~ ~ — ~ ° ~ ~ ~ — ~ ~ ~ ~ ~ ~ ~ ~ o ~ o U) ~ ~ ~ ~ ~ _ ~ ~t ~ _ c~ ~ ', ', ~ ~ ~ _ ~ ~ ~ ~ oN ~— _ v, ~ ~t ~D ~ ~ ~ oo — ON ~ ~ oo ~ ~ — r— oo — oo — =\ — — m} ~ — (> ~t ~ a~ t_ ~ _ _ O, ~ U~ C~ ~ ~ ~ _ ~ ~ ~ ~ ~ ~ o C~ _ _ _ ~ _ t_ - , ~ ~ _ _ _ C~ Dc .= C i~ ~ C ·= ,S C) Ct ~ U ~ cy ' - ~ ~e ~ ~ ~ ' ' 1 ~ ~ ~ ~ ~ .~ ~u, . Ct C~ 3 U: :: C) _ Ct C) O oo a, _ O C) C~ Ct .- '_ · tV ~ _ ·— E~ C~ . . C¢ ~ ~ 00 U: ~ O C¢ ¢: . ~ . C) . o o Z ~

96 COMPUTER CHIPS AND PAPER CLIPS suburban labor force is more likely to be white, college-educated, and married, and more willing to work part time. Shifts such as these in San Francisco, New York, and the Midwest clearly have negative implications for minority women in clerical jobs. SOURCES OF CHANGE IN C LERICAL WORK Sources of change in the size of occupations can be understood as resulting from three factors: overall economic growth; the differential growth rates of various industries, each of which uses occupations in different proportions (for example, finance uses relatively more clerical workers than manufacturing, while manufacturing uses more assemblers than finance); and the changes in the intensity with which the various industries use each occupation (for exam- ple, the insurance industry may alter its use of sales personnel relative to other workers overtime). This last "intensity" factor can be measured by an occupa- tional staffing ratio, which indicates the importance of a single occupation rela- tive to total employment in the industry (see Hunt and Hunt, 1985a:Chapter4, for a more complete discussion). Industries vary in their relative importance in the economy, and, as noted, different industries use different mixes of occupations to produce their final output. Table 3-12 summarizes some of this basic information for 11 broad industrial categories in 1982, with particular emphasis on clerical employment. As shown, clerical staffing ratios vary from only 2.4 percent in agriculture to 43.9 percent in finance. Since the number of clerical jobs in a given industry is the product of the total employment level in the industry and the staffing ratio for clerical workers in that industry, an industry may employ a large number of clerical workers even though it has a relatively low staffing ratio for clerical workers if its total employment is large. For example, as Table 3-12 shows, the service industry group employs about 5.5 million clerical workers, more than any other industry, although its clerical staffing ratio is not especially high. Slightly fewer than 3 million clerical jobs are located in each of the next two largest employers of clerical workers retail trade and finance. Although clerical workers are dispersed broadly throughout the economy, these three industries combined—services, retail trade, and fi- nance account for more than 11 million clerical jobs, almost 60 percent of total clerical employment. In teens of detailed industries, the 20 largest employers of clerical workers in 1982 are shown in Table 3-13, ranked by the total clerical employment in the industry.3 Thus, the first industry listed, state and local government and educa- 3The total number of 105 industries in this analysis represents the entire economy. Hunt and Hunt (1985a) found they could adjust the industrial categories of both the CPS data and the Bureau of Labor Statistics OES data into 105 compatible categories.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS TABLE 3- 12 Employment by Industry, 1982 97 Total Percent of Percent of Industry Percent of Workers in Clencal All Clencal Employment Total Industry in Employment Workers Employed (thousands) Employment Clencal Jobsa (thousands) in Industry Agriculture 3,401 3.4 2.4 83 0.4 Mining 1,028 1.0 12.5 128 0.7 Construction 5,756 5.8 7.8 451 2.4 Durables 11,968 12.0 12.6 1,513 8.2 Nondurables 8,318 8.4 12.9 1,074 5.8 Utilities 6,552 6.6 22.3 1,463 7.9 Wholesale trade 4,120 4.1 20.5 844 4.6 Retail trade 16,638 16.7 17.1 2,840 15.4 Finance 6,270 6.3 43.9 2,750 14.9 Services 30,259 30.4 18.1 5,473 29.7 Public administration 5,218 5.2 35.0 1,827 9.9 Total 99,528 100.0 18.5 18,466 100.0 aThis percentage is also known as the clerical staffing ratio. SOURCE: Calculated from Hunt and Hunt (1985a). tional services, has the largest number of clerical employees, 2.5 million, and, as the fourth column shows, accounts for more than 13 percent of all clerical employment. The table also shows clerical staffing ratios and total industry employment. In addition to state and local government and education, the larg- est employers of clerical workers are the federal government, banking, insur- ance, communications and transportation industries, and a variety of service sector industries from business services to personal services. The top 10 em- ployers of clerical workers account for about two-thirds of all clerical employ- ment, while the top 20 industries account for more than 80 percent. Since industry employment is crucial to assessing occupational employment trends, the trends over the last 27 years in total industry employment are pre- sented in Figure 3-3, which aggregates the employment in the top 10 clerical- employing industries and shows it relative to employment in the 105 industries that constitute the total economy in this analysis. The numbers are reported in index number form to make it easier to compare the growth trends in the top 10 with average growth in all 105 industries. As can be seen in Figure 3-3, the top 10 clerical-employing industries show much less cyclical variation in employ- ment than the economy as a whole. Employment in these 10 industries taken together remained positive through two of the three recessions during the pe- riod; only in 1982 did the composite growth rate of these 10 industries become negative, and then only barely so. Also important is that the average growth rate of these 10 industries clearly outdistanced the all-industry average for the entire 27-year period. As Hunt and Hunt (1985a) note, their higher growth stems

98 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-13 Clencal Employment by Industry, 1982 Industry Total Clerical Cumulative Industry Clerical Staffing Percent of Percentage of Employment Employment Ratio TotalClerical TotalClerical (thousands) (thousands) (percent) Employment Employment State and local government and educational services 13,068 2,512 19.2 13.4 13.4 Miscellaneous retail trade 10,476 2,496 23.8 13.3 26.8 Wholesale trade 5,294 1,531 28.9 8.2 34.9 Banking 1,650 1,180 71.5 6.3 41.2 Federal government 2,739 1,138 41.5 6.1 47.3 Insurance 1,700 911 53.6 4.9 52.2 Miscellaneous business services 3,139 896 28.5 4.8 57.0 Hospitals 4,166 666 16.0 3.6 60.5 Social services, museums, and membership organizations 2,755 587 21.3 3.1 63.7 Credit agencies, security and commodity brokers 1,015 577 56.9 3.1 66.8 Legal and miscellaneous services 1,628 560 34.4 3.0 69.7 Telephone and other communication 1,174 529 45.1 2.8 72.6 Physician and dental offices 1,309 394 30.1 2.1 74.7 Construction 3,913 324 8.3 1.7 76.4 Eating and drinking places 4,781 224 4.7 1.2 77.6 Electric services and gas distribution 792 207 26.2 1.1 78.7 Trucking and warehousing 1,206 199 16.5 1.1 79.8 Miscellaneous printing and publishing 846 192 22.8 1.0 80.8 Real estate 986 188 19.1 1.0 81.8 Miscellaneous personal services 1,219 186 15.3 1.0 82.8 SOURCE: Hunt and Hunt (1985a:Table 4.5); based on BLS data.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 220 o 200 To 11 oo 0, 1 80 X UJ Z us ~ 140 o J 120 100 99 — ~ ~ Top 10 Industries 0~0 All 105 Industries _ 1~ 1 1 1 r Up 1960 1965 1970 1975 1980 YEAR FIGURE 3-3 Total employment in the top 10 clencal-employing industnes, 1958-1984. SOURCE: Hunt and Hunt (1985a:Figure 4-A); based on 1958-1984 BLS data. largely from the fact that employment in these industries does not ordinarily retreat during recessionary periods but continues to expand. When the employment trends in each of the top 10 clerical-employing indus- tries are considered separately, the diversity among them emerges. Table 3-14 presents the 10 industries ranked by their growth in total employment over the 27-year period. In the fastest-growing industry, miscellaneous business serv- ices, the growth rate in employment is particularly striking, more than double the growth rate of the second-ranked industry. Its employment increased almost six times between 1958 and 1982, compared with about 67 percent for all indus- tries. This industry provides a wide variety of services to business firms, in- cluding accounting, customized computer software, consulting advice, and temporary personnel placement. Several other industries banking, credit agencies, security and commodity brokers, and, to a lesser extent, insurance- contributed significantly to clerical job growth during these years. Each of these industries has a staffing ratio for clerical workers in excess of 50 percent, the highest of all industries (see Table 3-1 3~. Hospitals were another important source of clerical employment growth (although the staffing ratio is low, it is a large sector), with rapid growth in total employment over the entire period;

100 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-14 Growth in Employment in the Top 10 Clerical-Employing Industries, 1958-1984 Total Employment (thousands) 1958 1980 1981 1982 1983 1984 Industry Miscellaneous business services Credit agencies and commodity brokers Percentage Growth, 1958-1984 617 3,112 3,334 3,400 3,577 Hospitals Banking State and local government Social services and museums Miscellaneous retail trade Insurance Federal government 4,033 554 340 912 964 991 1,077 908 2,750 2,904 3,014 3,034 617 1,571 1,628 1,656 1,663 5,647 13,375 13,259 13,098 13,099 1,113 2,099 2,177 2,112 2,117 2,980 5,275 5,358 5,278 5,259 6,259 10,452 10,488 10,397 10,592 1,021 1,676 1,702 1,714 1,721 2,191 2,866 2,772 2,739 2,752 1,166 243 2,988 229 1,684 173 13,185 133 2,179 96 5,526 85 11,106 77 1,757 72 2,783 27 Top 10 industries 23,222 48,714 49,275 49,230 49,898 51,619 122 All 105 industries 60,346 96,653 97,392 95,855 96,540 100,505 67 SOURCE: Calculated from unpublished data provided by H. Allan Hunt and Timothy L. Hunt, Upjohn Institute. unlike most of the other top 10 industries, however, hospitals experienced ac- tual employment declines between 1983 and 1984. Employment in the federal government grew very slowly over the entire period, and it has not grown at all since 1967. Alone among the top 10 clerical employers, it experienced slower growth than the national average. Employ- ment growth in state and local government and education, the largest single employer of clerical workers among all industries, was generally above aver- age, but employment has actually declined since 1980. The foregoing analysis of total employment growth in the top 10 clerical- employing industries and Figure 3-3 suggest that clerical employment has gen- erally increased faster than average employment because it has been concen- trated in industries that have grown more than the average. Growth in clerical employment can be more fully analyzed by decomposing it into each of the three factors that contribute to it: (1) overall economic growth, (2) differences in the rates of growth of industries (as noted), and (3) changes in the staffing ratios within industries. Changes in the latter two factors cause an occupation to increase or decrease in relative importance in the occupational structure. Changing staffing ratios, which are probably the most visible manifestation of the specific effects of technological change on occupational employment, have also affected clerical employment growth. As will be shown below, some in- dustries have increased their use of clerical workers per dollar of output, while

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 10 others have reduced their use. The clerical staffing ratios for computer-related occupations, for example, have risen over time in many industries due to the dramatic increases in the use of computers, while the staffing ratios for stenog- raphers have been falling for some time.4 An analysis of the clerical employment change for the 1972-1982 period, performed by Hunt and Hunt (1985a), shows that growth of clerical employ- ment at 28.8 percent was somewhat above the economywide average. Of these 28.8 percentage points, the contribution of overall growth was 21.1; that of differential rates of growth among industries, 4.4; and that of changes in occu- pational staffing ratios within industries, 3.3. That is, clerical employment in the aggregate was boosted by 4.4 percent because industries that were relatively large employers of clerical workers expanded more than other industries. This 4.4 percent, however, represents the average of widely differing contributions that the various industries made to clerical employment growth. As can be seen in Table 3-1S, those contributions range from—22.9 to +49.1. An increase in the clerical staffing ratio from 17.4 percent in 1972 to 18.5 percent in 1982 caused a further 3.3 percent rise in clerical employment. This, too, represents the average effect of changes in staffing ratios in all industries. Hence, in the 10 years from 1972 to 1982, neither changing staffing ratios nor differential rates of industry growth were major contributors to overall growth in clerical em- ployment, although both factors were modestly positive during the period; rather, overall economic growth accounted for the bulk of clerical employment growth. The differential experiences in the various industries shed light on sources of change; an understanding of them is likely to be helpful in considering future employment growth. Table 3-1S shows the decomposition of the growth in clerical employment for each of the 11 major industrial sectors (from Hunt and Hunt, 1985a). In retail trade, for example, there was a total increase of 741,000 clerical workers (or 35.3 percent) in the industry between 1972 and 1982. Eco- nomic growth at the same rate as the economy as a whole would have increased clerical employment by 4~44, 000 (21.1 percent) in this industry for this period; the more intensive use of clerical workers in the retail trade sector over the decade increased clerical jobs by 252,000 (12.0 percent); and the retail sector's above-average growth increased clerical employment by 45,000 (2.1 percent). A particularly striking result shown in this table is that staffing ratios for Staffing ratios may change for reasons other than the use of innovations; organizational change or job title change with no change in job content may occur, for example. It should be understood that any time an individual occupational staffing ratio changes, all of the remaining staffing ratios in that industry will also change. A very large increase in the need for professional or assembly-line workers in an industry could reduce the staffing ratio for clerical workers even though no change affected the nature of their jobs directly.

102 Go - C~ Dig lo, c o - C) U) m Cal ~4 o It .c C) o o cO; E to o u, ~ C) .. 4_ Cal of In ~ o - , — ~ ,~ Cal Pt U. Cal ~4 Ct ~ 3 ~ o V ~ o 3 ~ o to o UP ~ X UP ~ ~ o ~ . . . . . . . . . V) ~ UP Go ~ ~ o ~ ~ — 1 ~ ~ 1 1 . . . o ~ 1 ~ ~ ~ o ~ Do ~ ~ Go ~ ~ ~ . . . . . . . . . . . . Go ~ _ Cal d- ~ ~ ~ 1 ~ ~ ~ 1 1 1 1 1 _ - - - - - ~ - - - - - - == O ~ = ~D o0 MO ~ ~ ~ - ~ X \0 ~ \0 . ~ ·- ~ o ~ ~ ~ - ~ ~ l ~ ~ ~ ~ ~ ~ c~ ~ v - ~ ~ o~ .= r ~o ~ ~ ~ _ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ "o co - ~ ~ ~ ~ ~ ~ ~ ~ ~ c O.~- ~ o ~ o 3 ~ ° ~ ~ ~ -° ~ ~ ~ ~ ~ ~ ~ o <: ~ V - =e', .= ~ = ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ o. Ct ~' t— ~ ~ ~ ~t — ~ — ~ ~ ~ oo X X ~ V ~ ~ ~ ~ _ ~ ~ _ ~ ~ ~ ~ ~ ~ e~4 0 1 u~ ~) {~N C~ — d- ~ O — ~ ~ ~ O -0 ca C~ 0~ _ X C C~ ~ O ~ ~ 00 — ~ ~ ~ ~ O O ~ ~ ~ _ _ u, oo c~ ~ ~ t~ ~ ~ d ~ t— ~ ~ O ~ ~ ~ O ~ ~ ~ O ~ X ;o~ ~ ~,L] =- _ _ _ ~ ~ l~ —X ~ ~ C) ~ 0^ ~ 00 ~ ~ ~ O ~ ~ ~ ~ ~4 00 ~ ~O ~ _ u2 ~ l~ ~D ~ ~ O Do c~ Q ~ [_ ~ ~ .= ~ ~ o ~ ~ ° ~ ~ ° ° ~ ~ ~ ~ V ~ .L) ~ _ _ _ ~ ~ ~ ~ ~ ~ ~c c 5 e-~ 0 = c.~.c~ c ~ O ct c~ o c~ ct . ~ x c, - ct E~ . . ct oo - - ce . . o

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 103 clerical jobs are falling in a number of industries that employ large numbers of clerical workers. For example, in finance, which has experienced above-aver- age growth in clerical employment over the 10-year period (37.0 in finance compared with 28.8 percent for the whole economy), a decline in the staffing ratio would have reduced clerical employment by 6.9 percent. This decline suggests that the adoption of the new technologies in finance, which has been regarded as a leader, may be having a negative effect on the relative employ- ment of clerical workers. At the same time, the rapid technological change in this industry probably contributed to its rapid growth (through lower prices and expanded services and products), suggesting net positive effects of technology on employment. Staffing ratios for clerical jobs also fell in three other important industries utilities, wholesale trade, and public administration. Technology may have been a factor in these declines. In utilities, for example, the employment of telephone operators declined substantially as automatic switching equipment became more advanced. Less is known about developments in wholesale trade. As Hunt and Hunt (1985a) note, the decline in public administration is puz- zling. In recent years government has been adopting office automation on a large scale, although it was a slow starter. The postal service has automated numerous clerical jobs in the mail-sorting operation over the past several dec- ades. It is also true that government grew relatively slowly during this time period. Perhaps government administrators, when faced with tight budgets and rising demands for services, economize more on clerical jobs than on other positions. Interestingly, changes in staffing ratios contributed to increases in clerical employment in those industries that are relatively small employers of clerical workers: agriculture, mining, construction, and durable and nondurable manu- facturing. These changes may point to sources of clerical job growth. Clearly, more study of these trends is called for in order to better understand the reasons for changes in staffing ratios. OUTLOOK FOR CLERICAL EMPLOYMENT OVERALL GROWTH The critical question for considering future growth in clerical employment involves changes in its three underlying components. How much will the econ- omy grow overall? Will the fast-growing industries that have employed the majority of clerical workers continue to experience above-average growth? How will staffing ratios for clerical occupations in the various industries change? Obviously, projecting the future of these factors is not easy. While definitive conclusions are not possible, however, alternative estimates can re- duce uncertainty about possible outcomes.

104 at or Do cr. IS: C) a .o Ct o C) .~ U' . - C~ . Cal V, o o m Cal m - 00 _ Cal _ CL 00 O ._ _ U2 O ~4 .o C; U. ~ ~ . . o C) Cal c Ct o lo: C) a U. C) A: CD U. Ct - o ~ ~ O — — . . . . ~ 00 0 ~ 1 Ct ._ - 0 ~ ~ O =~ REV CD C~ _ ._ ~ =° 3 ~ o~ "o ~ ~ V C~ 04 C) C¢ ~ o~ U2 ~ . O |— r ~ _ Ct O Ct Ce ~ ~ _ ~ C~ U, ._ ~ O ~ ~ ~ 0 ~_ C~ ~:5 a, ~ U, 3 ~ o o ~ ,= C) .C) O4 .= E C a, ~ ~ _ . . . . ~ ~ C~ ~ 1 1 1 - ~ _ ~ ~ ~ ~ ~ oo . . . . . . . . o ~ ~ — o 1 1 1 _ _ _ _ _ _ _ _ . . . . . . . . oo oo oo oo oo oo oo oo C~ C~ o oo — o ~ ~ oo _ o~ ~ U~ C~ ~ — 1 1 1 1 =\ ~ — In ~D ~ ~ O o oo 1 — — ~ U) c~ 1 1 oo C~ ~ _ ~ oo V) ~ ~ o _ ~o ~ oo — ~ ~ ~ — o~ C ~, ~ U, C ~o ~o ox ~ ~ ~ ~ o C~ ~ ~ o _ Ch ~ ~ o ~ ~ =^ — `: ;^ ~ ~ ~ ~ ~ ~ ~o o ~ ~ V~ o ~ ~ ~ C~ ~ ~ r _ ~ t_ ~ - ^ 00 1_ ~ ~ ~ {_ O O O t— ~ ~ ~ ~ /> 1— ~ `~ =~ ~ — ~ —- _ v~ ~ ~ — ON oo — . . . . . . . . 00 0 ~ O ~ — 00 00 ~ ~ ~ ~ — ~ ~ ~ ~ _- O ~ — \0 0 ~ ~ O ~ 00 0 00 ~ O O ~) Ch t_ _^ ~ ~ _C~ ~ t~ ~ 00 0 C-l ~ ~ — e 3 O E O ~ ~ 1 a., ° on ~ ._ ~ 0 u: ~ 0 c ct ~ 0 ce ce 0 ~ . - 5 ct ~ U2 ~ _ °E ~ 0 ~ c' ~o . ° u C`2 c~ ._ _ 0 ,= _ ~ ~ 0 ~o _ _ x ce ct 0 _ r: ~ ~ 3 ~4 u ct ~ C~ . - ~i c,0 o o ~ o E~ c~ _ . ~ ~ ~ E~ ~ . - ~ ~ o ~ ~ c~ ~ ~ - c) c~ ~ ~ c) , ~ _ t_ o c ~ c~ u~ 3 O D c) O U, ~ V) ~ .= C~ ~ ~ Ct O ~ Ct Z C~ o ._ C) a, .E - a~ o ~L - Ct o ._ - e~ C) o V) m - o 1 ~0 ~0 Ct - Ct - Ct o U, Ct D . ~ C~ D Ct E~ . . Ct U~ oo C~ ~: e~ - . . o

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 105 Projections of occupational employment begin with the concept that the de- mand for labor is a derived demand. That is, employers wish to hire workers in order to produce the goods or supply the services that they sell to consumers. As a first step in developing projections, the BLS forecasts aggregate economic activity and industry demand for a given year in the future. In a second step, it estimates total employment by industry, based on forecasts of productivity growth over the period. In a third step, it estimates how many workers in each occupational category will be needed, adjusting the occupational staffing ratios in each industry on a judgmental basis to account for any changes in occupa- tional demand anticipated as a result of technological change or other factors. As Hunt and Hunt (1985a:5.6-7) note: Technological change actually enters the system in at least three places. First, the indus- t~y output projections should account for anticipated changes in demand induced by technological change. Secondly, the estimated productivity gains forecast for each in- dust~y should be influenced by technological change. Finally, the staffing patterns them- selves are altered directly to account for technological change. In other words, techno- logical change will have specific effects on some occupations, it will have an overall impact on the productivity of workers, and it will affect the demand for goods and services generally.... [T]his system involves a considerable amount of judgment, especially in anticipating the effects of technological change. There are no simple equa- tions that predict changes in staffing ratios within an industry. In fact, the BLS staff has found that trends in industry employment levels can be predicted more accurately than the changes in occupational employment.... This is due in large part to the difficulty of projecting specific occupational impacts of technological change. Table 3-16 summarizes the Hunt and Hunt (1985a:Chapter 5) analysis of the Bureau of Labor Statistics projections for 1995. The projections that Hunt and Hunt worked with used 1982 as the base year and were released by BLS in 1983. More recent projections for 1995, with 1984 as the base year, were re- leased by BLS late in 1985; the panel's necessarily brief analysis of the newer projections appears below. Hunt and Hunt's analysis of the 1983 BLS projections shows expected em- ployment increases in eight broad occupational categories, including decompo- sition of their sources of growth.5 In the past the industries that employed the most clerical workers have been faster growing than the average. That differen- tial began to erode in the 1970s, and it is not expected to reemerge by 1995. Differential rates of industry growth are expected to increase clerical employ- sThe BLS does not actually forecast occupational employment growth at the major group level, but it is still helpful to analyze the projections at this level of aggregation to provide an overview of the system. It also enables us to compare those projections with the historical CPS data reviewed above.

106 COMPUTER CHIPS AND PAPER CLIPS ment by only 1.6 percent between 1982 and 1995 (see Table 3-16), compared with 4.4 percent between 1972 and 1982. Of the large clerical-employing industries, growth is expected to be smallest in state and local government and education and in the federal government. Employment by the federal government is not expected to increase at all, while employment in state and local government and education is expected to recover from the declines that occurred in the 1980 and 1982 recessions and to grow, but significantly more slowly than the national average. The fastest-growing industries among the top employers of clerical workers are expected to be credit agencies and commodity brokers, hospitals, miscellaneous business services, and banking (Hunt and Hunt, 1985a). Of these, the growth anticipated by BLS for banking and hospitals is surprising. As noted in Chapter 2, both banking and health services are experiencing significant structural change. In banking, fac- tors that have contributed to speculation that employment growth may slow are deregulation, the increased use of automatic teller machines, the closing of branch and satellite banks, and declines in employment among several of the largest banks in the nation. In health services, employment is shifting away from hospitals toward other providers, and employment growth overall seems to be slowing. In these projections, average employment growth is expected to be 28.1 percent; only three of the eight broad occupations shown in Table 3-16 are expected to grow slower than the average operatives, laborers, and clerical workers, with clerical workers showing only slightly less than average growth. The variation of growth rates for the occupations around the average growth rate is relatively small, from 19.1 to 38.5 percent, although the range actually experienced between 1972 and 1982 was from—5.9 to +46.9 percent. And between 1972 and 1979 (omitting the recession years, which might distort the data), growth rates ranged from 8.5 to 35.4 percent. The Hunt and Hunt (1985a) analysis suggests that BLS projects less change in the relative impor- tance of broad occupations over the 13 years from 1982 to 1995 than actually occurred during the 7 years from 1972 to 1979. This finding probably also reflects BLS estimation methods, which make it difficult to project dramatic shifts, a natural caution that is probably not misplaced. In keeping with these results, the decomposition shown in Table 3-16 indicates that the relative ef- fects of changing staffing ratios and differential rates of industry growth are small for all occupations. The decomposition does reveal a significant change, however: the impact of staffing ratios on clerical employment is expected to be negative, the only reversal projected by BLS from the existing trends in the historical data. In its later projections for 1995 (released in November 1985), BLS reduced its estimates of overall growth and further reduced the expected relative size of clerical employment (Silvestri and Lukasiewicz, 19851. In the 1985 projections

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 107 of 1995 employment (which use 1984 rather than 1982 as the base year), em- ployment growth over the 11-year period is anticipated at 15 percent in the moderate trend series (meaning economic growth would be moderate) in com- parison with 25 percent in the earlier projection for the 13-year period, and the growth of clerical workers is expected to be 9.5 percent, well below the econo- mywide average. The share of clerical workers in total employment would thus fall from 17.5 percent of the labor force in 1984 to 16.7 percent in 1995. In their discussion of the new projections, Silvestri and Lukasiewicz point out that growth in clerical employment was about average between 1973 and 1984 and that, despite slower projected growth, clerical employment would still remain the largest category of employment and would add 1.8 million jobs. They sug- gest that the main reason for the reduced projection is revision in the bureau's estimate of the impact of office automation, not only its direct impact on the tasks performed by clerical workers but also on a shift in tasks from clerical workers to professional, technical, and managerial workers (which are ex- pected to be especially fast-growing occupational categories). In the earlier (1983) projections, the growth of employment in clerical work, as analyzed by Hunt and Hunt (see Table 3-16), was expected to be only 1.6 percent less than overall employment growth (26.5 compared with 28.1 for the 1982-1995 period). In the later (1985) projections, as analyzed by Silvestri and Lukasiewicz, the difference is much more dramatic: clerical employment is expected to grow 9.5 percent from 1985 to 1995, compared with a growth in total employment of 14.9 percent; BLS now projects clerical work to grow 5.4 percent less than and only about two-thirds as much as—total employment. Although the lack of time prevented a thorough analysis of these new projec- tions similar to that performed by Hunt and Hunt on the earlier ones, a first-cut analysis suggests that the dramatic projection of much slower than average growth in clerical jobs is actually largely the result of classification changes: for example, cashiers have been reclassified as sales workers. And because BLS projects the occupation of cashiers to grow very rapidly (more rapidly than projected in 1983), removing them—a very large category from the clerical classification has the effect of slowing the growth rate of clerical workers more than would have occurred without the change. A comparison of the 1983 and 1985 BLS projections appears in Table 3-17 along with some alternative projections. The 1983 projections are taken from Hunt and Hunt (1985a) and are based on the OES tape; they are not the same as the published data, since the tape excludes unpaid family workers and self- employed people. The 1985 BLS projections have been adjusted to correspond to the Hunt and Hunt 1983 version; the data are from the tape, and the panel has restored the classification (approximately) to its 1983 version. The category of clerical workers was made more similar to its 1983 form: cashiers, 1,570,000 in 1982 and projected to be 2,469,000 in 1995 (Silvestri and Lukasiewicz,

108 COMPUTER CHIPS AND PAPER CLIPS TABLE 3-17 Projections of Occupational Employment Growth, 1982-1995 Paneld 1985 Projection Clerical employment 1982 1995 Difference (1982-1995) Percent change (1982-1995) BLsa 1983 BLSb LeontiefC Panels 1985 -Duchin 1983 18,717 18,717 23,673 22,680 4,957 3,963 26.5 21.2 18,032 17,786 -246 -1.4 18,717 18,717 21,641 20,674 3,167 1,957 16.9 10.5 Total employment 1982 91,950 91,950 107,284 91,950 91,950 1995 117,745 112,361 143,753 117,745 112,341 Percent change (1982-1995) 28.1 22.2 34.0 28.1 22.2 Clencal workers as a percent of total 1982 20.4 20.4 16.8 20.4 20.4 1995 20.1 20.2 12.4 18.4 18.4 Percent change (1982-1995) (clerical staffing ratio) - 1.5 - 1.0 - 26.2 -9.9 -9.9 al983 projections from BLS, as presented in Hunt and Hunt (1985a). bl985 projections from BLS, unpublished data, adjusted by panel to correspond to 1983 BLS occupational classifications. CData from Leontief and Duchin (1984), as presented in Hunt and Hunt (1985a). Computed by the panel using a larger decrease in the clerical proportion of total employment, but retaining the other projections from BLS for the respective years. NOTE: Data for both 1983 and 1985 Bureau of Labor Statistics projections are from the unpub- lished Occupational Employment Statistics tape and differ from published figures because unpaid family workers and the self-employed are not included here. 1985), were added back, along with shipping packers, another fairly large cate- gory that had been removed, while order fillers were removed from the 1985 projections since they had been added to the clerical category between 1983 and 1985. As Table 3-17 shows, the result of this adjustment is that, overall, BLS has not altered its estimate of clerical employment growth. As in 1983, clerical employment growth is expected to be slightly less than average, so that the proportion of clerical employment to total employment falls slightly (indicating a small negative staffing ratio, shown on the last row in the table). However, between 1 983 and 1985, BLS changed some of its employment projections for specific clerical jobs substantially. Apparently, forecast increases (e.g., for cashiers) almost offset forecast decreases (e.g., for secretaries). These changes in specific occupations are examined further below. The BLS projections on balance suggest substantial growth in clerical work. Other forecasts have projected a much more pessimistic outlook for clerical employment. Perhaps the best known of these is the 1984 study by Leontief and Duchin, which can be seen in comparison with the BLS 1983 and 1985 projec-

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 109 lions in Table 3-17. While BLS projects increases in clerical employment over the period only slightly below the national average, Leontief and Duchin pre- dict an actual decrease of 1.4 percent in clerical employment for the 1982- 1995 period. Leontief and Duchin also predict relatively slow growth in employment of managers, 11.3 percent, in comparison with the above-average 38.5 percent 1983 BLS projections. If Leontief and Duchin are correct, even in part, it could mean not only displacement for large numbers of clerical workers but also difficulties for managers and other workers seeking to enter management posi- tions in the office. It is highly likely, however, that the Leontief-Duchin projections overesti- mate the negative effect of technology on clerical employment. For example, Leontief and Duchin project a 7.2 percent decline in secretarial employment, in comparison with the 30 percent increase forecast by BLS in 1983 (or the 26 percent increase forecast by BLS in 19854. The Leontief-Duchin estimate is based on the study with the largest estimate of productivity gains resulting from use of word-processing equipment, even though other studies suggest consider- ably smaller effects. They further assume that word-processing equipment cre- ates no "new" work, such as repeated drafts. They also assume fairly rapid diffusion of the new technology and do not take into account that productivity gains may vary by industry, being highest where the work to be done is routine and repetitive and lower where it tends to be more unique and less repetitive. Finally, they consider no technological changes over the period other than com- puter-based technologies, distorting their estimates of the relative importance of clerical jobs. The overall decline in the staffing ratio for clerical workers projected by Leontief and Duchin, 26.2 percent, is very large, even more than the decrease for operatives of 22.3 percent actually experienced between 1972 and 1982 (calculated from data in Hunt and Hunt, 1985a). Are Leontief and Duchin right about their projected large decline in clerical employment? The panel thinks not. Comparison between the largest actual decline in a staffing ratio, that for operatives, and the Leontief and Duchin projection for clerical workers can illuminate the issue. First, the decline ob- served for operatives has been occurring for some time, while Leontief and Duchin predict a rapid about-face for clerical work. Second, the trend observed for operatives in the 1972-1982 period includes the effect of the 1981-1982 recession, the steepest since the Great Depression of the 1930s. Since blue- collar employment falls steeply during a recession, the large change observed for operatives probably reflects this cyclical effect. Such a cyclical effect may not recur, and, even if it did, it would likely be smaller for clerical workers, whose employment traditionally does not fall as much as that of blue-collar workers in a recession. (Indeed, only in the last recession, 1981-1982, did clerical employment actually decline; in all others it continued to grow.) Third, the drop in operatives may also be due to a classification change, which resulted

110 COMPUTER CHIPS AND PAPER CLIPS in some operatives being reclassified as technicians. Hence, the actual drop in the employment of workers who carry out those functions may not have been as great as the numbers indicate.6 For all these reasons, the large decline that Leontief and Duchin predict in the staffing ratio of clerical workers seems highly unrealistic. It is worth noting, however, that, if the secretarial function actually disappears, or declines dramatically, as Leontief and Duchin predict, then employment opportunities for women will become a serious problem, especially if their training is not easily transferable to other jobs. As an illustration of the effects of staffing ratio changes that are large but more plausible than Leontief and Duchin's, the panel has computed two alter- native projections, corresponding to the 1983 and 1985 BLS projections. These projections use a larger decrease in the clerical staffing ratio than BLS fore- casts, but retain BLS figures for projected total employment. As noted above, several industries experienced declines in the clerical staffing ratio between 1972 and 1982; the largest rate of decline occurred in public administration and wholesale trade. The 9.9 percent decline in the clerical staffing ratio shown in the bottom row of Table 3-17 for the 1982-1995 period is computed by apply- ing the annual rate of decline in the clerical staffing ratio actually experienced by these two industries between 1972 and 1982 to the whole economy. As can be seen in Table 3- 17, under this assumption, clerical employment increases by 16.9 percent in the 1983-based projections and by 10.5 percent in the 1985- based projections. These increases are sizable, even though they are consider- ably below the average for the economy as a whole. (The primal cause of the difference in the two projections is the slower overall employment growth esti- mated by BLS in 1985 than in 1983.) Moreover, it might be argued that even this figure could be an overestimate of the reduction in growth in these jobs, since the rate of decline in the staffing ratios in public administration and wholesale trade may very well have been influenced by factors other than tech- nological change. Of course, it could be an underestimate if the rate of techno- logical change were to accelerate over the next 10 years. Although the growth in clerical employment in the panel's estimates is much larger than that pro- jected by Leontief and Duchin, it is also substantially smaller than that pro- jected by BLS because the change in the clerical staffing ratio is substantial and negative. In the panel's judgment, these estimates reflect the "most plausible 6Similar shifts are occurring in the classification of clerical workers, with uncertain effects on observable future employment. One possibility is that Leontief and Duchin's predictions will be supported by future observations of decline in such clerical categories as secretaries, but that the decline may turn out to be an artifact of reclassification. If all that happens is that secretaries take on other titles, a decline in the category "secretary" will not significantly affect employment opportu- nities for women, as they will be able to find employment in newly titled but fundamentally the same jobs.

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 111 worst case" as large a negative effect for clerical employment as is likely to occur. Based on this analysis and its best judgment concerning various trends, the panel concludes that the rate of growth of clerical employment will decline, most probably falling somewhere between the 1985 BLS and the 1985 panel estimates. In the panel's judgment, there is insufficient evidence to support the much more negative outlook of the Leontief-Duchin model. The panel does not foresee dramatic declines in the need for most clerical functions. Indeed, those that emphasize human interaction are likely to continue to grow. The panel does not, therefore, foresee a major unemployment problem for clerical workers. It is interesting to note that in its 1985 projections for 1995, reflecting its latest thinking, BLS approaches the panel's previous most plausible worst case fore- cast; BLS's revised 1985 projections for clerical employment fall between the 1983 BLS and the 1983 panel projections. As Table 3-17 shows, the 1985 BLS projection is 22.7 million clerical jobs in 1995 compared with the panel's pro- jection of 20.7 million. In all the panel and BLS estimates, at least 2 million clerical jobs will be created by 1995, while Leontief and Duchin project a loss of 250,000 clerical jobs. Although the number of clerical jobs will continue to grow overall, slower aggregate growth, and the inevitable shifts that will occur among the many subfields of clerical work, may well cause some structural unemployment. Pro- jections of some of the more detailed clerical occupations are examined next to see what occupational shifts can be expected. OCCUPATIONAL SHIFTS The 1983 BLS projections for 95 specific clerical occupations, as analyzed by Hunt and Hunt ~ 1985a), are shown in Table 3- 18, ranked in order of clerical staffing ratio change. Projected employment growth varies from +76.1 percent for computer operators over the 13-year period (1982 to 1995) to—20.0 percent for central office telephone operators. The proportion of the employment change that can be attributed to changes in staffing ratios, the more or less intensive use of specific occupations, varies from +38.4 for computer opera- tors to—55.6 for central office telephone operators. Clearly, the diversity of change anticipated in clerical suboccupations is large. In the OES classification, the three largest occupations are general office clerks, secretaries, and cashiers. No negative effects of technological change on employment trends in these occupations are discernible in these projections. The staffing ratio for cashiers is expected to increase significantly- despite the increased use of "intelligent cash registers" contributing to the overall 48.2 percent growth forecast for that occupation. The other two occupations are expected to experience only slightly more growth than the average growth for

112 C) .= - v at ._ Cal cr Do ~ C`O 2 ~ ~ vO ~ . o ~ ~ I — o Ct 4_ .o ~ Cal ~ g I: C) Ce o v ·= ~ ~ o - .O U. · - ~ ~ o o Cal o Ct a' m ~ 00 1_ ~ 1 0 ~ ~4 m ~ ~ C) ~ O to 5 a- V) Ct ._ ~ O ~= ~ ~ ~ e ~ . ~ O O At: ~ 00 C) C A ~ ~ ~ O ~ ~ ca ~ ~ Cal ~ ~ ~ '£ ~ o o ,= V _ C' 04 o ~ 3 ~ ~ ~ ~ U. .= £ ~ ~ e~ 0 1 u, ~ CL ~ o `, 3 ~ ~ ~ C~ £ ~ ox £ 5: _ ~ _ £ o oo ~ o ~ `,~ =- oo 1 ~ ~1 ~ ~o ~ ~ o · . . . 1 1 _ _ _ · oo oo oo oo ~ C~ et ~ oo · . . o X oo _ U) 1 oo C~ - oo ~ _ o ~ ~ V~ o . . . . . . . . oo ~ ~ ~ ~ ~ oo 1 1 1 — — _ _ _ _ _ _ _ . . . . . . . oo oo oo oo oo oo oo ~ ~ C~ ~ C~ ~ ~ - - _ _ . . . oo oo oo o ~o ~ o _ ~ o ~ _ 1_ ~ ~ o ~ ~ ~ ~ ~ ~ oo ~o ~ ~ . . . . . . . . . . . . . a~ e%, ~ 1 1 1 1 1 ~ —~ ~ aN ~ ~ ~ ~ ~ ~ o ~ ~ - oo . . . . . . . . . . . . . o oo o u) '] ~ ~ - - u) ~N - u) - - oo ~ ~ ~ ~ - - ~ ~ ~ o . ~ ~ ~ o ~ - ~ oo oo ~ u~ ~ - ~ oo . . . . ~ ~ ~ o u) ~ ~ ~ ~ ~ oo ~ ~ ~ ~ . . . . . . . . . . . ax r~ ~ oo ~ ~ 0 ~ ~ 0 ~ ~ ~ ~ ~ - - oo ~ - c~ ~ - q ~ o ~ ~ · ~ o v~ ~ - - \~o - - y ~ ~ o ~ o .~, c) ~J ~ =- C ~ O -¢ V ~ ~ . . ~ O - _ _ . . C~ 0- _ ~ 00 Ct Ct ~ ~ O C) £ ° ~ ~ ~ =- ~s ~ ~ ~ 3 ~ _ ~l c~ ° ~, ~- ~ C-. 3° c;, U, ) ~ c ~ C CO ._

113 ___ _____ _______ . . . . . . . . . . . . . . . oo oo oo oo oo oo oo oo oo oo oo oo oo oo oo C~ ~ ~ ~ ~ ~ ~ ~ ~ C~ ~ ~ CN ~ OX ~ ~ ~ 1— ~ C~ . . . . . . . . . . . . . . . . . ~ ~ o _ o o ~ ~ ~ o ~ ~ o o o o c~ c<~ _ ~ o ~ ~ ~ ~ . . . . . . . . . . . . . . . . . . . . . . . . o o o _ _ ~ ~ o _ _ ~ o o ~ o ~ o ~ _ _ o o o _ 1 V~ 1 1 1 1 — - 1 1 1 1 oo ~ ~ o ~ ~ oo ~ ~ ~ o ~ U~ ~ ~ ~ ~ ~ ~ ~ U~ o . . . . . . . . . . . . . . . . . . . . . . . . oo — ~O r— - , C'l a~ o _ _ _ ~ ~ ~ ~ c~ _ c _ _ ~ 0 ~ ~ ~ oo ~ oN ~ ~ oo ~ O ~ ~ ~ O ~ O ~ ~ r~ ~ ~ . . . . . . . . . . . . . . . . . . . . . . . . c~ ~ t— t— k.C' ~1 oo ~ ~ — oo ~ ~ — ~ oo ~ ~ — ~ ~ O ~ O _ ~ oo _ _ _ ~ _ ~ ~ C') _ _ ~ _ _ ~ ~ _ ~ ~ oo ~ _ o _ o oo ~ ~ ~ o _ oo ~ o V) ~ ~o o . . . . . . . . . . . . . . . . . . . . . . . . t— ~~o oo c~ —} (~ <5, tN a~ ~ ~ — ~ ~— ~ — ~ oo ~ os ~ ~ 0 0 C~ ~ ~ \0 ~ — ~ — — ~ 0~` ~ ~D ~ ~ — — ') ~ — C~ — C~ oo ~ ~ —) ·0 ~ — —] ~ U~ ~ ~] \,0 — ~ —) — ~ ~— ON C~ ~D ~ oo . . . . . . . . . . . . . . . . . . . . . . . . ~ ~ 0 ~ ~ oo — ~ ~ O ~ ~ 0 ~o 0 ~ u~ — oo ~ O ~ ~ a~ (~) c~] — — o,o oo d- ~] — ~ — ~ °o O O ~ ~ — ~ — - ) ~ _ ~ _ u ~ e - ~ U ~ ~ ~ ~ ~ ° ~ a ~ C

114 - - a' s v - o o c To .. _ C) C) V so . . - C) s v - ~o Ct . ~ C-' o ~O . Ct o~ C :S o A_ Do m 0£ .0 Ce Ct a ~ V Ct ._ ~ O ~ S 3 ~ CI: ~ $° a ~ s v 00 ~ S ~ o 3 -5: ~ O Cal OF Cal ·—~ sol s V, ~ V — U. ~ c C= ~ ~ O ~ .e a ~ _ "c C ~ lo. ~ ~ Cal C 0. V 3 ~ s c: in o ~ ~ - s: o ~ ~ ~ s - ~ ~ o c~ c) o - o ~ oo ~ ~ o o ~ ~ ~ . . . . . . . . . . . - - o o o o o o o o o 1 1 1 - ~ ON ON c~ ~ ~ ') ~ ~ - . . . . . . . . . . . cr~ ~ ~ 0 ~ 0 ~ ~ — ', _ — 1 1 — — 1 1 1 1 1 1 - - - - - - - - - - . . . . . . . . . . x oo oo oo x x oo x oo oo ~ ~ ~ ~ ~ ~ ~ ~ c~ ~ - oo C'I e~) —) 00 — — O O ~ O ~ . . . . . . . . . . . ~ o o ~ o o o o o o ~ — 1 1 ~ ~ ~ ~ ~ ~ ~ ~ o ~ o . . . . . . . . . . . oo ~ - o o ~ o ~ ~ o ~ — 1 1 ~ 1 1 1 1 1 — \0 0 0 °O ~ d — . . . . . . . . . . . ~ ~ x ~ ~ ~ ~ ~ ~ ~ ~ ) ) ~s~ ko . . . . . . . . . oo - ~ ~ ~ ~ o - ~ ~ - ~ ~ c~ - ~ - ~ ~ c~ . . ~ - ~ v~ ~ ~ c~ ~ ~ ~ ~ . . . . . . . . . . . — ~ t— V) ~ ~ d. ~ oo — — vo ~ — ~ ~ c~ r~ — ~ ~ c~ O . . . . . . . . . . . - ~ t - ~ ~ oo oo ~ v) oo ~ ~ ~ ~ ~ ~ ~ ~ - ~ ~ o ~ ~ ~ - oo o u~ o oo ~ ~ oo o ~ ~ . . . . . . . . . . . ~ ~ o ~ ~ ~ ~ ~ ~ oo o o ~ ~ ~ - ~ - ~ ') c~ ~ ~ c~ ~ - ,~,¢ O c) — =0 ~ y 3 0 3 ~ 93 ~ =. j c~ D O ~ Y _ ~A y

IIS 00 _ O ~ ~ ~ 0\ ~ _ v~ o ~ ~ o a~ o ~ c~ ~ . . . . . . . . . . . . . . . . . ~ O — C-l ~ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 s, o ~ ~ _ ~ ~ ~ . . . . . . . ~ o~ o ~ ~ ~ ~ _ _ 1 ~ V~ ~ ~ ~ ~ _ . . . . . . . o _ ~ ~ ~ o ~ 1 _ 1 ' 1 __ _ ____ _ _____ __ _ __ . . . . . . . . . . . . . . . . oo oo oo oo oo oo oo oo oo oo oo oo oo X oo oo oo ~ 0 ~ 0 r~ ~ oo ~D ~ oo oo ~ u~ oo ~ oo . . . . . . . . . . . . . . . . . _ _ _ ~ o - , o _ o o ~ C~ o o oo ~ ~ _ 1 1 1 1 — U~ 1 1 1 1 - 1 1 1 ~ 1 ~ 1 ~ oo — ~ — m} ~ c~ — ~ ~ ~D — O ~ ~ ~ v~ . . . . . . . . . . . . . . . . . . oo o ~ o oo ~ o o o ~ ~ o ~ ~ V~ o — '~ ~ 1 — 1 1 1 — 0 ~ ~ ~ oo ~ ~ — O — cr 0 ~ ~ oo — ~D . . . . . . . . . . . . . . . . . . X ~ ~ ~ ~ ~ ~ ~ ~ U~ ~ ~ C~ ~ ~ oo d- ~ ~ ~ - ) ~ — ~ — O _ ~ _ CN ~ oo ~ o o o o~ ~ ~ ~ ~ oo ~ o ~ o ~ ~ . . . . . . . . . . . . . . . . . . a, ~ ~ ~ 0 0 ~ u~ ~ \0 00 — ~ ~t Oo C~ 00 m) _ ~ ~ _ _ 00 00 0 ~ 00 C~ ~ CN ~ — 00 ~ ~ — ~ - ) d" . . . . . . . . . . . . . . . . . . O C~ ~ O — ~ C~ ~ — ~ ~ ~ — O ~ O C k O~ — ~ ~ ~ —) ~— — ~ O~ —) ~t — ~ ~ . . . . . . . . . . . . . . . . . . _ ~ ON O ~ 00 ~ ~ ~ ~ ~ ~ o0 C~ O ~ ~ — ~ ~ o ~ - oo ~ ~ ~ o ~ - ~ ~ oo ~ ko oN ~ ~ o%' N — ~ ~ 0 — oo ~ ~ ~ oo a~ O~ ~— — ~ — t— ~s~ — oo oO O~ t~ (~\ c~ ~ — c~ d ' — ~ —} ~ — - ~ ~ ~ ~ oo ,o. ~ ct: o y, o - ~ o ~ c ~ o ~= O — ~ .= ~ 3 C~D ·— o Ct —~ ~ o "o U: C~ · c c o o ~ ~ =: c~ ~ ~ ~ 'e n 3 ° 2 e _ e =3 o2 i: v' ~ v, V

116 V) _` to E So I:: o ._ . - U) C) _` .S _' Do m . . - C) a 04 O .. _ U: C) Ct U: 0C O V) - - O ~= 2 PI ~ V . Cal o o ~ REV U: ~ He C5 ~ to U: - ~ o Cal ~ ^ o ~ ~ ~ o o ~ . ~ C ~ c, 2 ~ V — .' OF 0 1 C) c' E as c .= C at == 0C 0 - te U. C) ~4 V - & o ~ & ~ ~ o ~ c~ o ~ ~ ~, ~ ~ ~ CN ~ ~ ~ 00 . . . . . . . . - ~ ~ ~ ~ ~ ~ ~ - - - - - - - - 1 1 1 1 1 1 1 1 1 . . . u) ~ ~ ~ - - 1 1 1 ~ oo ~ ~ o - ~ ~ ~ o o oo . . . . . . . . . . . . o v) o o o ~ v) ~ 1 ~ — 1 1 1 - - - - - . . . . . oo oo Oo oo oo ~ ~ ~ c~ . . . . a~ as — 0 1 1 °° ~ ~, 1 ~ 1 1 ' o o ~ o oo . . . . . o v~ ~ ~ ~ ~ ~ ~ u~ ~ . . . . . - ) ~ o ~ ~ - - - - - - - . . . . . . . oo oo oo oo oo co oo oo o oo . . . o - o 1 1 1 . . . o ~ - — 1 1 ~ oo O ~ — ~ ; . . . . . . . o ~ - o ~ o c~ 1 1 ~ 1 v~ ~ ~ ~ a~ ~ 0 . . . . . . . - - - oo oo ~ ~ ~ - ~ —] ~— oo ,5\ — ~ o oo oo ~ o . . . . . . . . . . . . Oo ~ ~ ~n ~ o~ ~ ~ ~ oo O _ ~ _ _ 1 M° 1 — — — ~ o ~ c~ ~ . . . . . o ~ - ~ - - ~ ~ - _ _ 1 ~ oo ~ ~ ~ ~ ~ . . . . . . . o ~ o ~ ~ ~ ~ ~ - 0 ~ 0 Oo ~ ~ ~ ~ u~ ~ . . . . . . . . . . . . ~D ~ O oo ~ ~ — ~ ~ ax 0 0 o ~ ~ c~ - ~ ~ - ) - co - cN o - - o ~ ~ ~ — ~ ~ ~ oo ~ v~o — . . . . . . . . . . . . v~ ~ oo ~ ~ ~ ~ ~ ~ r~ ~ r~ oo ~ r~ ~ oo a~ ~ oo c~ 0 C e ~ 0 _ e 0 ~ ° 0 ;, 1t ~ j u ~ ~ ~ '~ 3 e O u 3

~ ~S~ ~ ~ ~ — U) ~ a~ ~ ~ o~ ~ 0 ~ . . . . . . . . . . . . . . . . . oo ~ ~ O — ~ c~~ d Ir) ~r - - - - - ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ~ _ ~ ~ O u~ ~D ~ ~ oo O ~ a~ ~o ~ ~ ~ . . . . . . . . . . . . . . . . . o u~ ~ ~ ~ ~ o cM — — 1 _ ~ _ ~ ~ 1 u~ _ 1 1 - - - - ~ - - - - - - - - - - ~ ~ . . . . . . . . . . . . . . . . . oo oo oo oo oo co oo oo oo oo oo oo oo oo oo oo oo r~ M . . . . . . . . . . . . . . . . . oo — O~ oo —) ~ t— — ~ ~ oN ~ ~ ~ ~ ~ o 1 ~ 1 - 1 1 1 - 1 1 1 1 1 1 ~ 1 ~D o0 ~D oo O ~ u~ _ ~ ~ ~ O ~ ~ ~ ~ ~ ~ . . . . . . . . . . . . . . . . . r~ ~ 0 0 ~ ~ ~ ~ ~ ~ — ~ — 0 - ~ - ~ - - - oo oo ~ oo ~ r - - ] ~ o~ - ~ o ~ ~ - oo oo ~ ~ o . . . . . . . . . . . . . . . . . o ~ ~ ~ u~ ~ ~ oo ~ ~ ~ ~ ~ oo o ~ o 1 ~ — — — — ~ ~ ~ — 1 ~ — 1 ~ 1 1 1 ~ r~ 0 ~ ~ ~ ~ oo oo ~ oo u~~ ~— ON — m] 00 . . . . . . . . . . . . . . . . . o - ~ ~ c~ ~ ~ ~ ~ ~ ~ ~ o - ~ o - 1 ~ — ~ ~ ~ ~ ~ 1 1 1 1 1 ~ —, ON oo oo — —) — ~ ~ oo ~) ON a~ ~o ~ ~ . . . . . . . . . . . . . . . . . oo — d. O r~ ~ ~ ~ 0 0 — ON t - ~ - - ~ ~ ~ ~ ~ ~ o ~ oo oo ~ - ~ ~ ~ ~ Ko 5' - ~ '] . . . . . . . . . . . . . . . a~ ~ ~ ~ ~ ~ _ ~ ~ ~ ~ u~ ~ oo ', ~ _ - , ~ ~ _ — 0 a~ ~o — - ~ ~ ~ ~ ~ ~ ~ c ~ ~ c ~ ~ ~ ~ ~ ~ ~ o ~ ~ ~ ~ O O 'O C O O C ~ O "c ~ ~— ~ 3 — ~ .— ~ ~— — o ~ "c : 3 0 = D, O C O O o C ~ et ~ O E~ ~ ~ CQ ~ ~ ~ ~ m ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~ O ~ a.> 04 .= ~ .° 1OI ~ 0 ce ~ — 0 ct c~ c, ~ `: c' .° ce _ c~ 0 ~ O ce . _ . c~ ~ c' o ~ ~ ~ cO~ u°, o o - ~ -~O :: 0 cq ~ ° ca ~ ~ '~ · ~ ce · _ ~ CL o " ~ - ~ o ~ o ~ ~ m ' u' ~ ~ O x ce 0 t4 c~] E ~ c ~ 3 04 c~ - 0 ct .O c~ u =3 ~ ~ ~, ~ U, o o O D - 4 C) . ._ ~ ~ U~ ~ ~ O O ca _ ~ ~~ .. _ ~ U ._ t_ ~ ~ 3 ~ o ~ ~ ~ ~ ~ C Z ~ ~ V) 117

118 COMPUTER CHIPS AND PAPER CLIPS all clerical workers, with a change in the staffing ratio having a slightly negative effect for secretaries (as it does for all clerical workers in this projection) and an even smaller positive effect for general office clerks. In the 1983 BLS esti- mates, the projected growth rates in these two large clerical occupations, and for clerical workers generally, are approximately equal to those projected for total employment. The projected growth for secretaries is entirely consistent with the historical data (except for the apparent decline of secretaries in 1982, which may have been caused by the recession). In the 1985 projections, BLS reduced the estimated growth of secretaries from 1984 to 1995 to 9.5 percent, identical to the growth projected for all clerical work. The projected growth of secretaries and clerical employment generally is only about two-thirds the growth rate projected for total employment. The new estimate represents a decline of about 100,000 secretaries from the 1983 projection, but it would still result in an increase of almost 300,000 secretaries by 1995. Clearly the econ- omy will still need new workers trained in secretarial skills. In the 1983 projections, the fastest-growing clerical jobs are computer opera- tors, peripheral electronic data processing (EDP) equipment operators, medical insurance clerks, credit clerks in banking and insurance, credit authorizers, insurance checkers, receptionists, claims adjusters, cashiers, and survey work- ers. Employment growth in these occupations is expected to range from 48.2 to 76.1 percent between 1982 and 1995. Many of these occupations are expected to experience staffing ratio changes equivalent to employment increases of 20 percent or more. This list reflects in part the obvious technological impacts of computers, but it also reflects the continuing or increased importance of inter- action between a worker and a customer being served. Cashiers, for example, may replace other sales workers. As Hunt and Hunt put it (1985a:5-171: "A world of both high-tech and high touch is anticipated." Although various elec- tronic office technologies have the capacity to replace some aspects of human interaction for example, by automatic bank tellers, partially automated tele- phone number announcements, or computerized ad takers at newspapers cus- tomers may be resistant to using these devices, or the variety in transactions may make them less widely applicable than now anticipated. As Hunt and Hunt note, the fact that such devices can be developed does not guarantee that they will be, or that they will prove to be profitable if they are developed. The occupations expected to experience the largest percentage declines by 1995 are central office telephone operators, postal service clerks, data-entry operators, stenographers, security workers and purchase and sales clerks, and postal mail carriers. The effects of staffing ratio changes are expected to be large and negative in these occupations. Projected declines in employment for postal mail carriers (a loss of 11,000 jobs) and postal service clerks (a loss of 55,000 jobs) reflect both the large, negative effects of declining staffing ratios (largely due to technological change) and well below average industry growth

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 119 (probably due to increased competition from other forms of communication services and other mail services). The effect of a large negative staffing ratio change is expected to reduce employment for data-entry operators (despite a positive industry effect) by about 10 percent, or 34,000 jobs. This decline re- flects both the use of new technologies to perform the same work (for example, optical character reading) and the capability of new technologies to shift work to others (for example, consumers or professional and managerial staff). Other occupations expected to experience large, negative effects of staffing ratio changes include in-file operators, file clerks, credit reporters, brokerage clerks, and survey workers. Some of these occupations will nevertheless experience positive employment growth because of strong industry demand and overall economic growth. The number of survey workers, for example, is expected to grow substantially (48.2 percent), although the effect of staffing ratio changes alone would be negative ~—22.1 percent). It seems likely that rapid technologi- cal change is contributing to increased demand for surveys (along with de- creased labor input per dollar of output). The occupations of credit reporters (34.4 percent growth) and brokerage clerks (23.0 percent growth) may reflect similar changes. Other occupations that will experience slow growth, such as file clerks, are simply continuing a decline begun in the 1960s or 1970s. Many of the declining and slow growing occupations are back-of~ce jobs that require little or no direct contact with the customer and may have ready technological explanations: file clerks, stenographers, data-entry operators, and central of- fice telephone operators. Many of these occupations are held disproportion- ately by minority women—postal service clerks, file clerks, data-entry opera- tors, and telephone operators. These anticipated shifts among clerical suboccupations make it clear that the slowdown in projected growth for clerical workers will not affect all such work- ers equally. Some will be in even greater demand, some less so. Some occupa- tions will decrease, but it is important to note that none of the absolute decreases shown in Table 3- 18 is expected to be especially large. A preliminary comparison of 1983 and 1985 BLS projections for 1995 em- ployment in detailed clerical occupations shows that in several occupations BLS has projected larger declines than those suggested by the slowdown in employment growth overall (which is a 3.4 percent decrease, based on data published in Silvestri et al., 1983, and in Silvestri and Lukasiewicz, 1985~. In many occupations, however, greater growth is now projected. But comparisons are extremely difficult because of changes in the classification of OES data on which the projections are based. Table 3-19 shows comparisons for those occu- pations that could be straightforwardly matched and in which the projected changes in 1995 employment were significant and negative. The largest per- centage changes between the 1983 and 1985 projections include stenographers, statistical clerks, directory assistance operators, central office operators, and

120 Cal o · _ .: PA U. .o Cal ._ V) so o ~9 Cal o Cal m lo: ._ Us v o C) Cal ._ 3 lo: o ._ ~ Do C) C) U. At oo ·c V 3 ~ a.' m me ~ o -em C) C) ~ ~ -A ~ ~ _ £ au UP on o o .O _. oo ~ Be, ~ £ oo ~ ~ _ oo ~ U~3 ~ oo ON ', ~4 . . . . . . . . . . . . C'l oo t— t— — '] ~ O 1 1 1 1 1 1 1 1 1 1 1 1 o oo o o o o o o o o oo o ~ ~ c~ ~ ~ ~ ~ - ~ oo ~ ~ o ~ - ~ - ~ ~t ~ ~ - ~ OY — 1 1 1 1 1 - 1 1 1 1 1 o o o oo o o o o oo o o o _ O ~D oo ~ ~ ~ ~ ~ ~ ~ oo ~ ~ ~ — O ~ ~ ~ ~D \0 - - ~ o ~ ~ ~ ~ o~ - ~ o o oo o o o o o oo o o o o ~ ~ ~ ~ ~ ~ oo - - ~ - In - ~ \`o oo - ~ o`' ~t <~ ~ ') - ~ o o o oo o o o oo o o o o t_ ~ ~ _ ~ ON — a~ ~— ~ ~ ~ o ~ ~ ~ ~ oo ~ ~ ~ oo o o o o o o o o o oo oo o oo oo ~ ~ - o ~ ~ - ~ - ~ ~ ~ o o ~ ox ~— OY c~ ~ - ~ ~ ~ ~ ~ ~ ~ ~ c~ o ~ c ~ ~ ~ o ·= ·s m ~ ~ ^~ ~ ~ .~ ~ ~ ~ =.3 ~ E ~= ~ C) "D ;^ :: o ,.' ~ C~ U~ oo _ .~ .O 3 .O U, C5 ~: ~i Ct .,_ U2 ;> C~ c5 oo - - ct c) .c u, c) - .= e ~o ct ct .. v)

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 121 payroll and timekeeping clerks. For secretaries, teachers' aides, and computer operators, the lower levels of 1995 employment are in keeping with BLS esti- mates of slower growth overall. Which clerical occupations are likely to experience the largest increases in employment, especially for women? The latest published projections (Silvestri and Lukasiewicz, 1985) combined with data on the proportion of females in each occupation from the 1980 census yield the employment increases for women shown in Table 3-20. As the table shows, several of the slower-grow- ing, but large, female clerical occupations provide the largest job growth: sec- retanes, general office clerks, and bookkeeping, accounting, and auditing clerks. A fair degree of uncertainty surrounds all these projections, however. The impact of technological change could be either more or less than is now supposed. Projections for many occupations have been substantially changed by BLS between 1983 and 1985, generally downward. Substantial absolute declines are projected for stenographers (87,000 jobs for women, a 40.3 per- cent decline), postal service clerks (lO,OOO jobs, an 8.5 percent decline), statis- tical clerks (9,OOO jobs, a 12.7 percent decline), payroll and timekeeping clerks (9,000 jobs, a 5 percent decline), and central office operators (8,000 jobs, an TABLE 3-20 Clencal Occupations with Largest Projected Job Growth for Women, 1984-1995 Change in Female Employment,C Employmentb 1984-1995 (thousands) Number Occupation (and Percent Femalea) 1984 1995 (thousands) Percent Secretaries (98.8) 2,797 3,064 265 9.6 General office clerks (82.1) 2,398 2,629 190 9.6 Bookkeeping, accounting, and auditing clerks (89.7) 1,973 2,091 106 6.0 Switchboard operators (91.0) 347 447 91 28.7 Teachers' aides and educational assistants (92.7) 479 566 82 18.3 Receptionists and information clerks (93.4) 458 542 78 18.2 Computer operators, excluding peripheral equipment operators (58.9) 241 353 65 46.1 Order clerks, material, merchandise and services (67.4) 297 355 38 19.2 Billing, posting, and labeling machine operators (87.1) 234 272 33 16.2 Billing, cost, and rate clerks (80.7) 216 254 31 17.5 aPercent female from 1980 census data (Hunt and Hunt, 1985a:Table 2.4). bData from Silvestri and Lukasiewicz (1985:Table 2). CEstimates of job growth for women are conservative; because the percentage female is likely to grow by 1995 in many of these occupations, these numbers underestimate job growth for women.

122 COMPUTER CHIPS AND PAPER CLIPS 11.5 percent decline). In sum, slow growth is now expected to be even slower, and declines in some occupations will be significant. One area of anticipated rapid growth for clerical workers is the temporary help industry. Between the trough of the recession in 1982 and December 1985, the number of employees working for firms that supply temporary personnel nearly doubled, making it one of the fastest-growing industries with more than 50,000 workers (Carey and Hazelbaker, 1986~. Of the industry's 735,000 em- ployees, more than half are involved in office occupations (other significant areas of temporary personnel include industrial, medical, and engineering oc- cupations). Continued strong growth is expected, but at less than the very rapid rate experienced between 1982 and 1984 when the industry probably benefited from the recovery of the recession; as the recovery continues, employers may be more likely to hire workers on their regular payrolls. The BLS moderate- trend projections estimate annual growth for the industry at 5 percent through 1995, higher then the 4.2 percent annual rate estimated for all business services industries, and much higher than the overall estimated annual increase of 1.2 percent. Employment is expected to be 1,060,000 by 1995 (Carey and Ha- zelbaker, 19861. The employment figure can be thought of as the average num- ber of daily placements. The number of people working for temporary agencies during a year is much larger; one estimate put it at more than 5 million for 1984 (Appelbaum, 1985~. According to Appelbaum's study, the dominant motive of employers in hiring temporary rather than permanent workers is cost cutting: hiring, training, and fringe benefit costs are generally reduced. Appelbaum believes the dominant motive of temporary workers, especially women work- ers, is to obtain flexible scheduling. In Appelbaum's view there are disadvan- tages as well as advantages to temporary help, for both employers and employ- ees. Employees receive fewer fringe benefits and are unlikely to experience the earnings growth that normally accompanies seniority with an employer. Em- ployers, Appelbaum believes, may be forgoing the opportunity to restructure work in the most efficient and productive manner in the long run as they opt for short-term cost savings. Whatever the advantages and disadvantages of tempo- ra~y work, observers agree that employment growth in this sector is likely to remain strong. Given the overall slow growth predicted in clerical occupations generally, the rapid increase predicted in temporary employment constitutes a shift in clerical employment from permanent to temporary work. Because it is of interest to compare growth in clerical jobs with opportunities elsewhere in the economy, Table 3-21 presents the 20 occupations that are expected to experience the largest growth of jobs by 1995 throughout the entire labor market. In percentage terms their growth rates for the 1984-1995 period vary from 9.6 percent for secretaries and general office clerks to 71.7 percent for computer programmers. A number of the top 20 occupations are related to computers, and several others are related to preparing and serving food, to sales

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS TABLE 3-21 Occupations with the Largest Job Growth, Bureau of Labor Statistics Projections, 1984-1995 123 Employment (thousands) Occupation Cashiers Registered nurses Janitors and cleaners, including maid and housekeeping cleaners Truck drivers Waiters and waitresses Wholesale trade sales workers Nursing aides, orderlies, and attendants 1984 1995 1,902 2,469 1,377 1,829 Change in employment 1984-1995 Number (thousands) Percent of Total Job Growth Percent 1984-1995 3.6 2.8 2,940 3,383 2,484 2,911 1,625 2,049 1,248 1,617 Salespersons, retail Accountants and auditors Teachers, kindergarten and elementary Secretaries Computer programmers General office clerks Food preparation workers, excluding fast food 987 Food preparation and service workers, fast food 1,201 Computer systems analysts, electronic data processing Electrical and electronics engineers Electrical and electronics 1,204 1,552 2,732 3,075 882 1,189 1,381 1,662 2,797 3,064 341 586 2,398 2,629 308 520 390 566 452 443 428 424 369 348 343 307 281 268 245 231 1.205 219 1,417 215 212 597 206 29.8 32.8 15.1 17.2 26.1 29.6 28.9 12.6 34.8 20.3 9.6 71.7 9.6 17.9 68.7 52.8 2.8 2.7 2.7 2.3 2.2 2.2 1.9 1.9 1.7 1.5 1.4 1.4 1.4 1.3 1.3 technicians and technologists 404 607 202 50.0 1.3 Guards 733 921 188 25.6 1.2 Automotive and motorcycle mechanics 922 1,107 185 20.1 1.2 SOURCE: Office of Technology Assessment (1986: Table 8A-1). in general, and to health care. Janitors and cleaners, truck drivers, accountants and auditors, and teachers are also included. Clearly the large-growth occupa- tions include both those that require substantial education and training (regis- tered nurses, electrical and electronics engineers) and those that do not (cash- iers, fast-food workers). The slowest-growing of these occupations are the clerical occupations.

24 COMPUTER CHIPS AND PAPER CLIPS JOB LOSS AND DISPLACED WORKERS The slower growth projected for clerical workers overall, combined with the anticipated decreases in some clerical subfields and the specific employment losses observed in some instances of automation (see, e.g., Appelbaum, 1984; Gutek and Bikson, 1985), indicate the need for considering programs that will help workers shift occupations as necessary. The relative disappearance of back-off~ce jobs and continued growth in jobs with greater customer contact suggest that workers may benefit from help in identifying their oral communi- cation skills and developing them for transition to jobs with more customer contact. It is difficult to know how the slower growth overall and the occupational shifts anticipated in clerical work will translate into unemployment, if at all. Some of those who lose specific jobs will remain with their employers; even if many are displaced (laid off more or less permanently), substantial unemploy- ment may not result if those who lose jobs find new ones quickly. Because of intense interest in the question of displaced workers during the recessions of the early 1980s, BLS added a special supplement to the January 1984 Current Population Survey (Flaim and Sehgal, 19851. Respondents from about 60,000 households were asked whether any adult member of the house- hold had experienced job loss since 1979 because of a plant closing, an em- ployer going out of business, or lack of recall from a layoff. The results show that from 1979 to 1983, nearly 1 1 million nonagricultural workers lost jobs; this represented about 12 percent of annual employment based on payroll data, a proportion much larger than previous estimates obtained from less complete information (Podgursky, 1986~. Podgursky's analysis of a sample ofthe displaced workers (those aged 20-61 and displaced no later than December 1982) finds that a substantial proportion found jobs within 15 weeks (about 40 percent) and a substantial proportion remained jobless for more than 52 weeks (about 26 percent). Relative to their representation in the total labor force, the displaced workers were dispropor- tionately blue-collar workers from manufacturing; white-collar and service workers constituted 39 percent of those displaced. Whether white-collar or blue-collar, women remained jobless longer than men; white-collar workers had less joblessness than blue-collar workers, and more women worked in white-collar than in blue-collar jobs. The industries that are especially large employers of clerical workers (for example, government and finance) had much less displacement than others. An analysis of a somewhat different sample of the displaced workers in the same survey (Flaim and Sehgal, 1985) showed that clerical workers fared somewhat better than average in replacing their former earnings. In general, however, reemployment incurred earnings losses for nearly everyone (average

EMPLOYMENT LEVELS AND OCCUPATIONAL SHIFTS 125 losses for full-time workers were 12 percent in white-collar and 15 percent in blue-collar jobs; Podgursky and Swaim, 19861. Both women and blacks had longer periods of joblessness, and workers with below-average education had larger earnings losses on reemployment. Better-paid and more senior workers faced larger losses, as did those who were reemployed in different occupations or industries. Some 15 to 20 percent of the displaced workers had participated in education or training programs, most often paid for either by themselves or their employers. CONCLUSION This review of recent and future trends in labor force growth and clerical employment suggests that technological change is likely to contribute to em- ployment problems for women, but that massive job loss is unlikely to occur. Clerical jobs will experience slower growth in the aggregate than they have in the recent past, and shifts in the demand for various clerical occupations will occur. Some increased structural unemployment may result. Between 1958 and 1968, clerical employment grew considerably more rap- idly than total employment: clerical workers increased their share of total em- ployment from 14.5 to 17.5 percent. Between 1970 and 1980, however, growth was slower: clerical workers increased only from 17.5 to 18.5 percent. And, since 1980, their share of total employment has remained the same. In the "most plausible worst case" scenario developed by the panel the largest plau- sible negative impact of information technology on clerical employment, based on the largest historic negative effect in any industry using the November 1985 BLS projections of slower employment growth overall, clerical employ- ment by 1995 would have lost at most 2 percentage points of its share of total employment. Clerical employment would increase by 2.0 million jobs, or an increase of 10.5 percent, between 1982 and 1995. This growth rate is one-third that of the historic growth rate of clerical employment between 1972 and 1982 (about 0.7 percent per year compared with 2.3 percent per year). Labor supply is also expected to grow more slowly after 1985, but not as slowly as clerical employment would in the most plausible worst case. Labor force growth will decline to about one-half its growth in the previous decade. Thus, women are expected to enter the labor force somewhat faster than clerical jobs are now expected to grow. In this case, a larger proportion of women than in the past would have to find work in nonclerical occupations. Since women have been integrating formerly male-dominated occupations, the panel does not think such a change would be difficult. In any event, the panel does not foresee massive technologically induced unemployment among clerical work- ers or would-be clerical workers, and if the economy grows at an average rate, opportunities in other areas should be sufficient for women to shift occupations.

126 COMPUTER CHIPS AND PAPER CLIPS This finding, coupled with the panel's review of the shifts among clerical sub- fields, supports the need for programs that will assist workers with transition. As noted above, back-off~ce jobs appear to be declining relatively, and in some cases absolutely, while those that require greater contact with customers are increasing relatively. This differential change is likely to pose a particular hard- ship for minority women, who hold relatively more of the declining jobs. For all women, the slower growth and likely shifts point to the need for sound basic education in core competencies, such as reasoning ability, problem solving, and communication, to prepare workers for the jobs likely to be created. Since technical occupations, such as those related to operating computer equipment, will also grow relatively rapidly, good technical training will also be important.