Determinants of female labor force participation

Female Labor Force

Determinants of Female Labor Participation Rate

The rate of participation of females in the Korean labor force is a subject in need of extensive study, as certain large-scale trends present educational and employment trajectories seemingly at odds with each other. While female enrollment in tertiary school has increased along with rapid economic growth in the country (and the region), many Korean women still work as unpaid laborers in family businesses and neither have nor are trying to obtain an individual and direct means of income or regular employment. Despite the fact that economic, social, and political developments in the latter half of the twentieth century and into the current era have led to profound changes in the status and culturally acceptable roles of women in Korea, the People's Daily Online (2011) reports that South Korea still has the lowest proportional level of educated females in the regular workforce of any OECD country. Compared with a country that has achieved a great deal more success in the integration of females into the workplace, such as the United States, South Korea still has fairly restrictive cultural values in terms of individual rights, freedoms, and equality, all of which could have a potential impact on the rate of female participation in the workforce.

By better understanding the factors that contribute to females' decisions of whether or not to join the workforce, practical changes that could achieve greater levels of equality and independence for women in South Korea might be suggested. This research set out to examine which factor or factors have the biggest impact on female employment in an effort to address the seeming inconsistencies between female's participation in education and employment in South Korea. It is hypothesized that issues such as marriage and education, which are seen as major influential factors in female's workforce entrance and exit decisions in the United States, will have the largest influence on employment decisions, yet other factors are also examined. Previous studies and forecasts of labor fore participation incorporate a range of effects including historical changes in behavior, differences across individuals and cohorts in terms of the effects of policies and rewards for work, and the culture that is being examined, and this research also examined several variables in these areas to develop a comprehensive and accurate model for the prediction of female workforce participation.

In the following sections, previous research and existing theoretical frameworks will be presented and discussed as background for the current research presented in the latter sections of this report. The sources of data and the definitions of the basic statistics examined in terms of female employment decision-making influences will then be presented, followed by a section providing an overview of the regression analysis techniques used to determine the model of best fit for the selected statistics and the degree of influence each independent variable was seen to have. These results are then discussed in context of the direct findings of the research and the larger-scale implications they have in the context of female employment in South Korea, with the model of best fit and a quantifiable analysis of various effects on labor force entrance and exit decisions delivered. Finally, this research concludes with a brief and general summary of the findings and results, and suggests manners in which South Korea can encourage greater participation by females in the labor force.

Section 2: Theory and Prior Research

Though specific research detailing the specific factors that influence the decision by cohorts of female to enter or exit the labor force has not been especially extensive, issues involving or related to females in the labor force have been the subject of much previous attention in recent decades. Cultural factors and the way in which certain life events or situations are perceived have been identified as major contributors to the decisions made in this regard; for example, maternal status (i.e. whether or not a woman has children) is known to have a significant influence on the perception of the decision to seek or maintain employment although acceptance of working mothers has grown significantly (Smith, 1981). Age, on the other hand, is and remains a negative influence on female labor force participation, with an increase in age correlating to a lower likelihood of employment (Waite, 1981). Research such as this makes it clear that there are a variety of influences on females' decision whether or not to join the labor force, and also shows that determining these influential factors and their degree of influence can be quite complex in a dynamic environment due to the shifting of cultural influences on choice and perception (Smith, 1981; Waite, 1981).

Other research shows the complex relationship between economics and education in influencing women's participation in the Korean labor force. On the one hand, women that have achieved at least a middle-school education are more economically active -- i.e. participate to a greater extent in the labor force -- yet at the same time women from lower socioeconomic backgrounds are also more likely to seek and achieve regular employment as part of the labor force than women from a higher economic standing and cultural status (Nam, 1991). Generally speaking, women of a higher socioeconomic background are more likely to achieve higher levels of education, meaning that these two influences are slightly at odds with each other and create especially complex conflicts for educated women of a high economic/cultural standing, as they have both an individual impetus and familial ability to achieve an education at a high level, yet also face cultural inhibitions against their employment due to a lack of need and it being seen as an irregularity in Korean society (Nam, 1991). More recent research has found that despite increasing equality in employment access leading up to the Asian economic crisis a decade ago, wage disparities remain substantial and the crisis itself setback employment equality substantially (Kim & Voos, 2007).

Other influences on employment that have been identified as especially important in Korean culture include familial issues and other class motives, including intrinsic cultural pressures such as perceptions regarding employment generally as well as certain types and areas of employment specifically (Kim, 1997). For factory workers, very near the bottom of the socioeconomic adder in South Korea, there is often a divide between traditional family values and expectations and the desire for upward class and economic ability that a factory job is seen as potentially providing, creating a conflict of influences and motivations in the lower classes similar to that observed between education and economic prosperity in the middle and upper classes (Kim, 1997). For countries that have only experienced market economies and economic liberty and modern growth for the past few decades or half-a-century, including South Korea, there are a host of other economic, political, and cultural complications hat do not exist in more developed countries or those that have experienced greater levels of democratization and market capitalism for longer periods (Aslanbeigui & Pressman, 1994). These complications have even made it difficult to determine whether or not women are benefiting from industrialization and increased economic opportunities at all, let alone what the various factors influencing this issue might be; in South Korea particularly, the high degree of growth and rapid increase in prosperity has made it difficult to accurately and reliably trace employment trends, with controversy surrounding most official and many academic reports (Aslanbeigui & pressman, 1994). It is against this backdrop of mixed findings and controversial statistics regarding South Korea's labor force and female participation that this research is being undertaken.

The underlying framework or the conducting of this research finds its precedent in other quantitative measures of the same basic variables, and in largely similar contexts (Fernandez & Rodriquez-Poo, 1997). Essentially, models have been created that position employment status as a dependent variable, with a changing though relatively focused set of independent variables multiplied by unknown coefficients in order to determine this employment status (Fernandez & Rodriquez-Poo, 1997). Progit and Logit analysis have been regularly utilized and well vetted in terms of identifying the appropriate coefficients for these independent variables and thus the full scope of the models that provide the best fit to observable employment statistics and other measures (Fernandez & Rodriquez-Poo, 1997). This background research and previously established framework were utilized to ensure the validity of this research's results.

Section 3: Data Sources, Definitions, and Descriptive Statistics

Given the limitations of the researcher's capabilities due to time and resource constraints, no new data was collected for analysis in this research. Instead, a public use dataset from the Panel Study of Income Dynamics was obtained and utilized, as it provided all of the essential statistical data necessary for further analysis and incorporation into various models. This dataset is also more extensive and more reliable than any that could likely have been created through individual original primary data collection, enhancing the validity and the reliability of the findings of this analysis. In order to constrain the dataset and the actual population of statistics analyzed in this research, the summary statistics of Gander/Sex were used to ensure that only females were incorporated in the study; the dataset only consisted of individuals of an age high enough to be legally employable, so although age was used as an independent variable in testing it was not needed as a constraining factor in actually determining the dataset/population of statistics to be utilized.

Panels of data from each of the odd-numbered years between 2000 and 2010 -- 2001. 2003, 2005, 2007, and 2009 -- were used in this research, and though these datasets included many of the same individuals from year to year each individual was counted separately in each separate year. This was considered appropriate for two reasons: first, as this research was not longitudinal in nature, and is attempting to determine specific influences on female labor force participation currently rather than how these forces might change over time, an analysis that controls for individual factors over an extended time period was not deemed necessary; second, because individual circumstances can change, the decision of each individual in each year can be seen as the result of the changing factors, which are for all intents and purposes unique (as is the decision-making process itself) each year. This created a pooled dataset that allowed for a more detailed and comprehensive analysis of the relationship between the dependent variable of employment status and the independent variables of age, health, marital, status as a student, and two variables measuring educational level.

The research included several dummy variables, including the dependent variable of employment status/labor force participation that is the key focus of this research. health, marital status, and status as a student were also dummy variables assigned either a zero or a one depending on whether the condition was met or not (e.g. A health person would receive a "1" for health as opposed to a "0" for an unhealthy person; a married individual has a "1" while an unmarried individual has a "0," etc.). The existence of dummy variables, and particularly of a dummy variable for the dependent variable, had a direct impact on the design of the research and the statistical tests that were eventually selected as appropriate and capable of providing meaningful and reliable results (Fernandez & Rodriquez-Poo, 1997). Income was also considered as a potential variable, yet did not allow for the creation of a clear model but rather clouded the data patterns found in the research. Basic descriptive statistics were calculated, but were essentially meaningless given the performing of Logit rather than a standard regression analysis on the data (regression was rejected due to the dummy variable status of the dependent variable) (Fernandez & Rodriquez-Poo, 1997).

Section 4: Analysis

LOGIT Pooled tests were conducted on the data from each year of the panel survey utilized as the dataset in this research, with different models used in the LOGIT analysis to enable a comparison of different data configurations and interactions. After a host of such testing and analysis, the Model 4 Logit test for 2009 was selected for further analysis, interpretation, and discussion because a) 2009 was the most recent year for which data was available, and b) Model 4 contained the highest level of significance across all variables and provided the most comprehensive model of best fit for the observed data. Employment was the sole dependent variable in all models; in Model 1, independent variables included were age, health, and status as a student, Model 2 used these same independent variables with the addition of marital status; Model 3 added the first of two education measures; and finally Model 4 included all identified potential independent variables: age, health, status as a student, marital status, and two different measures of education. As the most comprehensive model and the model with the highest level of significance across the independent variables, Model 4 was deemed to be especially reliable in its predictive/correlative value.

Section 5: Discussion

The analysis performed on the data did not just yield a model of best fit, but also provided coefficients for each of the independent variables that help to determine their significance and level of influence on female participation. The independent variables had both positive and negative influences on levels of labor force participation, with a positive health status, and the second measure of education both tending to influence greater levels of female participation in the labor force, while increases in age, being married, being a student, and the other educational measure all had a negative effect on participation levels. More specifically, an increase in age of one year was correlated with a reduction of 0.3% in the likelihood that a woman would be employed, while being a current student had a much more dramatic effect, reducing the likelihood of employment by more than thirty percent. This is more than twice the impact that marital status had; women who are married are just under fifteen percent less likely to be employed than their unmarried counterparts, and finally the first of two educational measures led to a reduced likelihood of female participation in the labor force of just over eight percent. A positive health status, on the other hand, increased the likelihood of female labor force participation by twenty-eight percent, while the second educational measure increased labor participation by 0.5%.