Editorial Board:
Editor:
George H. Conklin,
 North Carolina
 Central University

Board:
Bob Davis
 North Carolina
 Agricultural and
 Technical State
 University

Richard Dixon,
 UNC-Wilmington

Ken Land,
 Duke University

Miles Simpson,
 North Carolina
 Central University

Ron Wimberley,
 N.C. State University

Robert Wortham,
 North Carolina
 Central University

® 
Volume 4, Number 1
Spring 2006
 

Women’s Occupational Mobility After Work Interruption

by

Megumi Omori
Bloomsburg University of Pennsylvania
and
Shelley A. Smith
University of South Carolina
 

Introduction 

    Female labor force participation has increased steadily for nearly five decades.  By 2002, approximately half of currently employed workers were women, which represented a ten percent increase since 1970.  The increase was attributable mainly to significant increases in the numbers of women with young children who work.  In 1975, less than 40 percent of women with preschool children, and about 34 percent of women with children aged three years or younger, were in the labor force.  By 2002, those percentages had increased to 64 and 60 percent respectively.  By and large, women’s labor force participation today is less constrained by children than in the past. 

    For the majority of women, the age at which they bear children is coterminous with the years they are economically active in paid employment.  They are therefore at greater risk of intermittent labor market attachment and its attendant impact on job continuity.  Our objective in this paper is to evaluate the impact of work interruption on occupational mobility.  Specifically, to what jobs do women return after a work hiatus?

    Women on average hold eight different jobs between the ages of 18 and 32, although the number of job changes decreases with age (Bureau of Labor Statistics, 1998).  When women experience work interruptions, they often experience a change from the occupation held before the interruption upon labor force reentry – occupational mobility.  The occupations obtained by women upon labor force reentry are often lower than previously in terms of status, prestige, and wages (Appelbaum, 1981; Felmlee, 1995; Robinson, 1986).  Intermittent labor force participation has a significant negative impact on women’s occupational attainment and wages.  As Felmlee points out, “a single break in employment will have an immediate detrimental effect on a woman’s attainment upon labor-force reentry (1995:171). 

    Intermittent labor force participation may occur for a number of reasons, but childbearing is key:  both participation and reentry are strongly connected to childbirth (Wenk and Rosenfeld, 1992).  Childbearing significantly raises the likelihood that a woman will exit the labor force (Felmlee, 1993; Stinebrickner, 2002), and most women withdraw from labor force after childbirth for at least a short period (Desai and Waite, 1991; Joesch, 1994).

    Compared to childless women, substantial downward occupational mobility occurs during women’s childbearing periods (Dex, 1990).  The relationship between fertility, and both occupational status and wages, is significantly negative (Maume, 1999; Dex, 1990).  Maume (1999) finds that, on average, women who experience a childbirth during a given year are significantly less likely to receive a wage promotion, and more likely to be jobless, compared with those with no childbirth experience.   Specific examples are revealed by Dex and Shaw (1986): following childbirth, roughly one third of teachers experienced downward occupational mobility – ten percent became clerical workers, eight percent semi-skilled, five percent non-manual, and three percent sales workers.  And like teachers, approximately one-third of clerical workers lost ground after childbirth.  In short, childbirth appears to be the crucial factor in downward occupational mobility among women.

    We investigate the observed price paid by women in terms of their occupational status following childbirth, by focusing on occupational choice and occupational characteristics.  Human capital theory offers an explanation in terms of occupational choice and atrophy rates specific to certain types of occupations.  The approach assumes that individuals choose occupations with the aim of maximizing lifetime earnings and minimizing penalties associated with the desired lifetime participation in the labor force (Polacheck, 1979).  The suggests that women choose occupations with a minimum of skills depreciation if they exit the labor force, and reenter at a later date.  It further suggests that women choose occupations which provide greater scheduling flexibility in order to accommodate the demands of childbearing and child rearing (Anker, 1997).

    Polacheck defines the penalty of skills depreciation as an "atrophy rate," the "loss of earnings potential that can be attributed to periods of work intermittence (1981:62)."  His estimates show that the atrophy rates for professional, managerial, and craft occupations – typically considered male-dominated and sex-typed – are low.  This implies that the wage penalty for intermittence is higher for male-typed occupations than for female-typed occupations.  Moreover, the longer the time spent out of the labor force, the lower the probability of holding managerial or professional occupations.  Polacheck argues that women who show less attachment to the labor force choose occupations with low atrophy rates and little skills depreciation (1981).  This explanation suggests that women in high atrophy rate occupations are less likely to experience work interruptions, and when they do, they will spend shorter periods detached from the labor force, hence minimizing skills depreciation.  If this is the case, there is little reason for women to avoid occupations with high atrophy rates should they plan uninterrupted labor force participation.  However, women in occupations with high atrophy rates are also expected to experience downward occupational mobility upon returning to work.   It can be concluded, therefore, that women will concentrate in occupations sex-typed as female, and female-dominated, which feature, on average,  limited promotion opportunities, relatively low pay, and have few career-continuity requirements (Wolf and Rosenfeld, 1978).  Although almost half of all workers are women, they are still concentrated in a smaller number of occupations than their male counterparts, and frequently those occupations are sex-typed as female.  About 98 percent of kindergarten teachers and dental hygienists are women, as are 97 percent of receptionists.  In contrast, twelve percent of engineers, 19 percent of dentists, and 30 percent of lawyers are women (U.S. Bureau of Labor Statistics, 2004).

    Female-typed occupations are easier to re-enter after interruption than male-typed occupations (Wolf and Rosenfeld, 1978).  They are easily reentered because they require little-to-no on-the-job training or specialized human capital, but rather require general human capital prior to job attainment (Wolf and Rosenfeld, 1978).  General human capital, that acquired through general training, is more easily transferable than specialized human capital, which is unique to specific jobs.  This is demonstrated in occupation-specific studies of post-interruption occupational mobility in male- vs. female-typed jobs.  The amount of downward mobility between pre- and post-childbirth was significantly less in nursing, a female-dominated occupation, than semi-skilled factory work, a male-dominated occupation (Dex and Shaw, 1986).  Further detail is provided by Appelbaum (1981).  Women who held high status jobs prior to childbearing experienced a decline in job prestige following an interruption, whereas women in medium-high and medium-low status jobs were more likely to maintain the same prestige level. 

     This study further inquires into the issues raised above.  Specifically, how is female occupational mobility after an employment interruption related to the characteristics of the occupation held prior to the interruption, especially its gender composition?  And, are women who have a childbirth experience during the interruption (compared to women who do not) more at risk of downward occupational mobility upon reentry into the labor force?

    We analyze data from the Panel Study of Income Dynamics (PSID).  (We used Public Release II, 1988-1993, the PSID 1985-1997 Childhood and Adoption History File, and the Dictionary of Occupational Titles).   These data comprise a longitudinal survey of U.S. individuals and families, collected annually since 1968.    The PSID Public Release II files provide detailed information on respondents’ employment status and occupation.  Detailed information on respondents’ employment status was gathered from 1988 to 1993 on a month-by-month basis, hence our analysis is limited to that period.  For the childbirth information, we use the PSID 1985-1997 Childbirth and Adoption History File, which contains retrospective histories of childbirth and adoption.  The Childbirth and Adoption History File and the Public Release II data files share a unique identifier for each observation, allowing us to match respondents’ information from each file.

    The sample consists of women between the ages of 18 and 45 years, who experienced an employment interruption during the observed period.  Since we focus on occupational mobility, we only include employed women for whom pre- and post-interruption occupation is reported. Although women who do not return to the labor force may be thought of a downwardly mobile, for our purposes we consider non-returnees as unpaid, therefore "non-occupational" (Payne and Abbott, 1990).  Our focus is on the movement between paid occupations in the conventional labor market. 

    The only occupation information in the PSID data is the 1970 three-digit census occupation identifier code.  We appended occupational characteristics from other sources to each PSID respondent’s record.  The occupational characteristics are Specific Vocational Preparation  (SVP) scores, occupational prestige scores, and occupational percent female.  SVP is "the amount of time required to learn the techniques, acquire the information, and develop the facility needed for average performance in a specific job-worker situation.  SVP includes training acquired in a school, work, military, institutional or vocational environment, but excludes schooling without specific vocational content (Roose and Price 1981:9)."

    The SVP scores and percent female were obtained using the Dictionary of Occupational Titles (England and Kilbourne, 1988).  The occupational prestige scores are those from the rating system originally developed at NORC and updated by Hauser and Warren (1997).  The prestige scores vary from 0 to 100, high scores representing high prestige.   The SVP scores range from 1 to 9, where higher scores indicate more preparation.  The most recent SVP and prestige scores are 1980 and 1989 respectively, whereas the PSID occupational codes are 1970-based.  Hence we converted the PSID occupational codes to their 1980 counterparts using the “crosswalks” provided by the NOICC Crosswalk & Data Center.

    We focus on the determinants of occupational mobility, which is measured two ways.  In both, the key observation is if downward mobility occurred, or not.  The first measure of mobility uses the SVP score.  We measure mobility by taking the difference between a woman’s occupational SVP score before, and after, the work interruption.  A drop in the SVP indicates downward mobility.  The second indicator of mobility uses the occupational prestige scores, and is used like the SVP scores.  Occupational mobility measured with prestige scores is based on the difference between the pre- and post-interruption scores of a respondent’s occupation.

    Key independent predictors of mobility are the characteristics of the respondent’s occupation prior to an employment interruption.  Since we predict mobility two ways, using SVP and occupational prestige scores, each measure is included in the model predicting mobility as measured by the other.  Occupational percent female is included as a predictor in both models.  We also measure the effects on mobility of human capital characteristics – full-time employment experience since age 18, earnings from prior to the work interruption, education and time out of the labor force.  Finally, we control for individual demographic characteristics known to affect labor market outcomes: age, race (white vs. non-white), household headship, a birth experience event during the work interruption, and number of preschool children. 

    The analysis is predicting the probability of a binary outcome – whether a woman experienced downward mobility (or not) after an employment interruption.  We use a logistic regression to model the relationship between that outcome and the predictors.  The probability of a downward mobility vs. no mobility event is an additive function of the vector of independent variables listed above – the set of  occupational characteristics, human capital characteristics and individual demographic characteristics. 

    The basic descriptive statistics for the sample are presented in Table 1.  For descriptive purposes, we divide the sample into four subgroups: those who did, and did not, experience downward mobility based on the SVP score measure, and those who did, and did not, experience downward mobility based on the occupational prestige measure.

Roughly thirty percent of the 1,129 women in the sample encountered downward mobility, regardless of the measure used. 

    Both tables show significant differences in occupational percent female between downwardly mobile and not mobile women.  Those downwardly mobile upon labor force reentry were in occupations with between five and ten percent more women as those not mobile.  In contrast, both SVP scores and occupational prestige of occupations held prior to labor force exit were marginally, though significantly, higher among downwardly mobile women.  These statistics show initial support for the impact of atrophy rates associated with particular occupations.

    Among human capital characteristics, few differences emerge with respect to the percent who have worked full time since age 18, but both tables show significantly fewer months out of the labor force among women who experienced no downward mobility upon reentry.  Finally, a notably greater percentage of women encountering no downward mobility had achieved higher levels of college-level education or more.

    Overall, the comparison of those who were downwardly mobile compared to those who were not, shows several similarities among the individual demographic characteristics: average age is about 33 years, between 83 and 85 percent are white, and the average number of preschool children is 0.6.  Women who had a childbirth during the interruption were somewhat more likely to experience downward mobility, although the percentages are not significantly different.  Roughly four percent more women who were downwardly mobile were household heads, compared to those not downwardly mobile. 

    Table 3 presents the results of the logistic regression analysis.  The first model reports mobility measured with changes in SVP (first two columns of results), the second model mobility as measured by changes in occupational prestige (second two columns). 

    Neither model shows a significant impact of individual demographic characteristics.  Most unexpected is the absence of any impact of experiencing a childbirth during the work interruption.  We expect that this unlikely result may be that the impact of a childbirth experience becomes attenuated upon controlling for occupational characteristics.  The impacts of those characteristics are clear, significant, and unambiguous.  Moreover, they are consistent with our expectations.  The likelihood of downward mobility, when measured using SVP scores, is significantly reduced in occupations with proportionately more women.  The result is also negative for the prestige model, but not significant.  The models further show a positive relationship between occupational prestige, and SVP scores, and the probability that a woman is downwardly mobile upon labor force reentry.  Overall, the impact of occupational characteristics on downward mobility indicates that skills demands of occupations with high atrophy rates will clearly penalize women who have work interruptions, regardless of whether a childbirth occurred  the interruption.

    Among human capital characteristics of respondents, education and length of interruption both affect the likelihood of downward mobility.  Having either some college or a college degree or more reduces the likelihood of downward mobility, the latter having a stronger reducing impact than the former.  Higher levels of education are typically associated with higher occupational rewards, which we would also expect to be associated with higher rates of atrophy.  Other things equal, however, it appears that higher education itself may somewhat offset the impact of other occupational characteristics.  Lastly, the longer a woman is out of the labor force, the greater is her likelihood of experiencing downward mobility upon reentry.  We surmise this to be the result of skills atrophy during a work interruption.

    Women’s labor force participation patterns are often repetitions of exits and returns to active employment.  Intermittent employment is argued to be associated with downward occupational mobility (Appelbaum, 1981; Felmlee, 1995).  In the study reported here, childbirth experience, number of preschool children, and other individual demographic characteristics typically associated with labor market outcomes are shown to have no resounding impact on occupational mobility.  The overriding determinants of mobility in our results are occupational characteristics, and to a more limited degree, selected human capital characteristics.

    Our models measured mobility using two indicators – SVP scores and occupational prestige scores.  Both approaches rendered consistent results, boosting our confidence in these findings.  The impact of occupational characteristics on mobility supports the premises of human capital theory:  women concentrate into occupations that offer the fewest skills depreciation penalties should they temporarily withdraw from the labor force.  The results further illustrate the importance of education to labor market outcomes.  While education is most often associated, positively, with occupational rewards such as earnings, it also has a significant and important impact on mobility following a work interruption.  In short, education reduces the negative impact of occupational characteristics.  Education may be considered a form of general human capital, which comprises basic and transferable skills (Tam, 1997), and is slow to depreciate over time (Mincer and Ofek, 1982).  By way of contrast, specialized human capital skills specific to occupations are less transferable, and more quickly become obsolete, during a labor force exit (Zaloker, 1998). 

    Our analysis result that the childbirth experience in and of itself has no appreciable impact on occupational mobility suggests that its effect may be largely indirect.  In related research, we explore this by examining the hypothesis that women who have a childbirth during an interruption stay out of the labor force longer than their counterparts who do not give birth.  Our results above suggest that the skills depreciation impact of months out of the labor force translates the impact of a childbirth experience into a greater likelihood of downward mobility.  Specifically, months out of the labor force is positively and significantly correlated with depreciation of occupation-specific human capital.

Anker, R.  1997.  “Theories of Occupational Segregation by Sex: An Overview.”  International Labour Review, 136:315-349.

Appelbaum, E.  1981.  Back to Work: Determinants of Women’s Successful Re-entry.  Boston, MA: Auburn House.

Becker, G.  1981.  A Treatise on the Family.  Cambridge, MA: Harvard University Press.

Desai, S. and L. J. Waite.  1991.  “Women’s Employment during Pregnancy and After the First Birth: Occupational Characteristics and Work Commitment.”  American Sociological Review, 56:551-566.

Dex, S.  1990.  “Occupational Mobility over Women’s Lifetime.”  Ch. 10 in Payne, G. and Pamela Abbot (eds.). The Social Mobility of Women: Beyond Male Mobility Models.  London: The Falmer Press.

Duncan, K. C. and M. J. Prus.  1992.  “Atrophy Rates for Intermittent Employment for Married and Never-married Women: A Test of the Human Capital Theory of Occupational Sex Segregation.” Quarterly Review of Economics and Finance, 32:27-37.

England, P.  1982.  “The Failure of Human Capital Theory to Explain Occupational Sex Segregation.”  The Journal of Human Resources, 17:358-370.

Felmlee, D. H.  1995.  “Causes and Consequences of Women’s Employment Discontinuity, 1967-1973.”  Work and Occupations, 22:167-187.

Felmlee, D., H.  1993.  “The Dynamic Interdependence of Women’s Employment and Fertility.”  Social Science Research, 22:333-360.

Hauser, R. M. and J. R. Warren.  1997.   “Socioeconomic Indexes for Occupations: A Review, Update, and Critique.” Sociological Methodology, 27:177-298.

Hodge, R.W., P. M. Siegel, and P. H. Rossi.  1964.  “Occupational Prestige in the United States, 1925-1963.”  American Journal of Sociology, 70:268-292.

Joesch, J. M.  1994.  “Children and the Timing of Women’s paid work after Childbirth: A Further Specification of the Relationship.”  Journal of Marriage and the Family, 56:429-440.

Kleinbaum, D. G., L. L. Kupper, K. E. Muller, and A. Nizam.  1998.  Applied Regression Analysis and Other Multivariable Methods.  3rd Edition.  Pacific Grove, CA: Duxbury Press.

Maume, D. J. Jr.  1999.  “Occupational Segregation and the Career Mobility of White Men and Women.”  Social Forces, 77:1433-1459.

Mincer, J. and H. Ofek.  1982.  “Interrupted Work Careers: Depreciation and Restoration of Human Capital.  Journal of Human Resources, 17:3-24.

Polachek, S.  1981.  “Occupational Self-selection: A Human Capital Approach to Sex Differentiation in Occupational Structure.”  The Review of Economics and Statistics, 63:60-69.

Robinson, P.  1986.  “Women’s Occupational Attainment: The Effect of Work Interruptions, Self-selection, and Unobserved Characteristics.” Social Science Research, 15:323-346.

Roos, P. A. and J. Price.  1981.  Dictionary of Occupational Titles (DOT), Fourth Edition.  National Academy of Science, Committee on Occupational Classification and Analysis.

Stinebrickner, T. R.  2002.  “An Analysis of Occupational Change and Departure from the Labor Force.”  Journal of Human Resources, 37:192-216.

Tam, T.  1997.  “Sex Segregation and Occupational Gender Inequality in the United States: Devaluation of Specialized Training.?” American Journal of Sociology, 102:1652-1692.

U.S. Bureau of Labor Statistics.  2004.  Women in the Labor Force: A Data Book.

U.S. Bureau of Labor Statistics.  2001.  “Civilian Labor Force Participation Rates by Sex, Age, Race, and Hispanic Origin, 1980, 1990, 2000, and projected to 2010.”  http://www.bls.gov/emp/home.htm#data

U.S. Bureau of Labor Statistics.  2001.  “Employment Status of the Civilian Noninstitutional Population 16 years and over by Sex, 1969 to Date.”
http://www.bls.gov/opub/mlr/mlrhome.htm

U.S. Bureau of Labor Statistics.  1998.  Number of Jobs, Labor Market Experience, and Earning Growth: Results from a Longitudinal Survey. See:
<Http://ftp.bls.gov/pub/news.release/history/nlsoy.062598.news>

Wenk, D., and R. A. Rosenfeld.  1992.  “Women’s Employment Exit and Reentry: Job-leaving Reasons and their Consequences.” Research in Social Stratification and Mobility, 11:127-150.

Wolf, W. C. and R. A. Rosenfeld.  1978.  “Sex Structure of Occupations and Job Mobility.”  Social Forces, 56:823-844.

 Return to Sociation Today, Spring 2006
 

©2006 by the North Carolina Sociological Association