HETEROGENEOUS JOBS ON WOMEN LABOR SUPPLY
The relationship between female labor supply and its various implications like marriage, fertility, family earnings distribution and male-female wage differentials poses questions on how the behavior of female work force distribution is influenced by job heterogeneity (Female Labor Supply A Survey Killingsworth, Mark R. Heckman, James J. 1986). In order to understand female labor supply, it would more sufficient to look at its significant facts as manifested in Western economies in the modern period.
Furthermore, there is a need to use theoretical and empirical models in order to come up an appropriate conclusion which would satisfy the question as to how heterogeneous jobs play an important role in defining the nature of female labor supply particularly in Western nations.
Brief Literature Review
Women Labor Participation in the Twentieth Century
In most developed economies, the growth of women participation in the labor market had become a striking feature in the twentieth century. Participation growth began at various times but it was only in 1960s when it became eminent (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). In the United States (1890 to 1980), women participation was highest among females age 20-24(67.8), second within 25-44(64.9) and third within 45-64(50.5) (U.S. Department of Commerce, Bureau of the CensusCensus of Population 1890-1980 Cited in Killingsworth, Mark R. Heckman, James J.). In Canada (1911 to 1981), female labor force participation was highest among age 25-44(65.2), second was 19 and below (61.2) and third within 45-64(46.3)(Census of Canada 1911-1981 Cited in Killingsworth, Mark R. Heckman, James J.). Great Britain (1891 to 1981) recorded its highest among women age 20-24(69.3), second within 25-44(59.5) and third age 20 and below (56.4)(Department of Employment and Productivity, British Labour Statistics Historical Abstract 1886-1968, London HMSO Cited in Killingsworth, Mark R. Heckman, James J.). In Germany (1895 to 1981), women participation is highest among females age 20-24(71.0), second within 25-44(58.4) and third within 14-19(40.4) (Statistiehes Jahrbuch 1962 Yearbook of Labour Statistics 1975 Cited in Killingsworth, Mark R. Heckman, James J.). With the exception of Germany (by which female participation rate changed minimal since 1946), the rest showed gradual increase in female labor participation until recent recorded periods.
The participation growth was dominantly attributed to the increasing number of involvement among married women in the labor market. Relatively, the participation of single women decreased in the United States and Britain. However, participation rate of the nonmarried remains higher than married females in total. In contrast to women labor supply, male participation has been decreasing since the first quarter of the twentieth century (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). Weekly hours worked, however, decreased in both men and women.
Qualitative Changes as Induced by Heterogeneous Jobs
Females Working Character
Apart from the quantitative changes of labor supply in women, there also occurred qualitative modifications. The increase of women labor supply in respect to heterogeneous jobs had resulted to changes with respect to the females working character more and more women started to hold white collar, so to speak clerical jobs at the onset of the twentieth century (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). What became more surprising was that the rate of women participation in white collar jobs occurred even faster relative to men. This implied how significant white collar jobs are to most people (regardless of gender) as the economy grew wider. In 1900, 20.2 percent of the female workers in the United States held white collar jobs, way lower compared to 65.6 percent in 1980 (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). The growth rate was higher compared to mens, whose white collar job participation increased by only 2.3 percent relative to womens at 10 percent. As women concentration in clerical jobs increased from periods 1900 to 1980, women participation in service and blue collar jobs decreased whereas, mens participation in both was relatively increasing but in a slow phase.
Education, Marriage and Ethnicity
One of the factors which stimulated women participation in highly competitive jobs is education. In both United States and Britain, educational attainment among women increased that is, women who finished four or more years in college rose to 50 percent (even though women educational attainment growth was slow in the United States from 1926 to 1930). In fact in 1980, 20.5 percent of women (aged 25-29) in the United States had finished 4 years in college (Table 2.19 United States Schooling completed by the female population, by age, 1980 Cited in Killingsworth, Mark R. Heckman, James J.). In Great Britain (1981), 8.1 percent of women (aged 25-29) finished higher education (Table 2.20 Great Britain Highest educational qualification attained by female population in 1981 by age Cited in Killingsworth, Mark R. Heckman, James J.)
If there is one aspect which had implied dramatic fluctuations in women labor involvement (1890 to 1980), was mainly because of female distribution by marital status (in spite the increased participation of the nonmarried). In 1890, the proportion between married and nonmarried women was equal however, it varied from 1890 to 1980. Within the period, 50 percent of females aged 20 to 24 were married (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). In 1960, however, 70 percent of the women within the said age bracket were married.
In the United States (1890 to 1950) alone, fertility rose in 1920 but declined in 1940 (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). Fertility rate in 1950 was below and almost similar to that of 1910 (perhaps due to the Great Depression). Older works (such as Mincers in 1966) suggested that participation among married women was procyclical in the United States and moreover, teenage participation is sensitively influenced by cyclical variation (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). In short, this section argues that women labor participation is closely influenced by educational attainment and marital status among females. The higher the educational attainment is, the higher the probability of acquiring a job.
Other factors are embodied by the fact that white single women (age 65 and below) had higher participation rate relative to black single women (age 65 and below). However, black single and black married women have a higher participation rate than their white counterparts. Also, the number of children or dependents in the family is one of the determining factors which influence women job involvement. The higher the number of dependents, the greater probability it is for both men and women to find sources of income.
Economic Theory and Model
Theory on Utility
There were some approaches used to analyze labor supply using the job heterogeneity model. In 1956, Tinbergen emphasized the relationship between the choice of the amount of job characteristics and the desirable job characteristics (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). The assumption states that these variables caused income reduction but increased utility, assuming that all types of work required similar number of hours.
The first approach to utility theory focused on the combined determination of labor supply and continuous job characteristics (treated as consumer goods). In 1982, Atrostic specified utility as a function of consumer good, leisure time and job characteristics (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). If desirability of job characteristics stimulates compensating wage differentials, then wage rate becomes its dependent variable (similar to how consumer expenditure works).
The second approach to utility theory focused on the combined determination of labor supply and the individuals discrete choice given the availability of different jobs (Female Labor Supply A Survey Killingsworth, Mark R. Heckman, James J. 1986). In 1985, Killingsworth specified and defined utility as a dependent variable to the job which an individual has (wage rate and exogenous income remained equal) and that wage rate may vary from one job to another that is, an individual may receive different wage rate depending on ones choice of job (again wage rate, job choice and working hours are endogenously treated).
Heckmans Regression Model
The inclusion of job variables in labor supply functions creates a dilemma simply because they are considered choice variables (factors like job characteristics) and thus, cannot be provided with a consistent estimate due to its endogenous nature. As a result, job variables are determined along with the number of hours worked. Variations in job variables are accorded with wage differentials compensation and so, wage rate becomes a choice variable too (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). In order to do this, we consider the regression of hours of work (H) on the wage (W), exogenous income (R), a background characteristics vector (X) and a job variable (J) which may be represented as H a bW cR kX jJ e (where e is an error term). However, the estimate of j cannot be consistently provided by fitting least squares because J is endogenous and thus, it is determined along with H. In line with this, compensating wage differentials (W) which is now a choice variable, influence differences in J, so least squares estimates may also yield biased estimates of b. Furthermore, if the individuals choice of J depends on X elements (schooling and age), e and elements in X shall be correlated and thus, may again yield biased estimates of k on X elements. To sum it up, there is a need to revise or extend existing estimation strategies. Whereas by ignoring J, the model now appears to be H a bW cR kX u (where u is a composite error term given by u e jJ) and fitting by least squares results to biased estimates of all its parameters. This is due to the fact that J and W are jointly determined (based on conventional compensating differentials). As a result, labor supply is optimized and consequently, H is added to the list of endogenous variables. Therefore, u shall be correlated with W, R and X. In short, H a bW cR kX jJ e is a labor supply function while H a bW cR kX u is a labor supply locus.
Rather than statistical, the basic issue appears to be behavioral in nature. Wages, number of hours worked and job characteristics become an individual option (endogenously chosen) with the availability of heterogeneous jobs. In spite the consistency of parameter estimates, choice of hours worked becomes a dependent variable of the choice of job characteristics. How endogenously-chosen variable changes affect exogenous changes is not represented. For instance the following is overseen if wage and job variables remained correlated, changes in exogenous income pose a direct effect on hours worked these changes may, however, trigger modifications on wage and job variables which may both affect hours worked indirectly and directly.
Parameter Estimation
Little has been done on the analysis of labor supply using labor supply models addressing job heterogeneity. More often than not, related studies were only used to understand areas like compensating wage differentials (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986). The fact that the study of job variables implied no interest to most related works (provides limited information on preferences), it makes itself useful in understanding labor supply and how it is affected by heterogeneous jobs.
Job variables help provide estimation in compensating wage differential needed by a worker if one wished to change a certain job or particular job characteristics. The job heterogeneity model may be used to supply information as to what factors stimulate wage differential compensation. Wages and data on labor supply may be used to consistently estimate supply parameters and at the same time consider job variables of an individual prior to compensating wage differentials. Labor supply information within job variations is produced by similar preference structure. This structure dictates job choice and compensating wage differentials and thus, more data is acquired out of wage, labor supply and job choice analysis rather than wage analysis alone (Female Labor Supply A SurveyKillingsworth, Mark R. Heckman, James J. 1986).