تفاوت های جنسیتی در میزان جداسازی کار و ثبات استخدام : شواهد جدید از داده های کارفرما و کارمند
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|3743||2008||23 صفحه PDF||سفارش دهید||9219 کلمه|
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله شامل 9219 کلمه می باشد.
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Labour Economics, Volume 15, Issue 5, October 2008, Pages 915–937
I analyze the job separation process to learn about gender differences in job separation rates and employment stability. An essential finding is that employer-employee data are required to identify gender differences in job separation probabilities because of labor market segregation. Failure to recognize this may potentially lead to statistical discrimination. Three important empirical results are obtained from the analysis. First, women have higher unconditional job separation probabilities. Second, there are no gender differences in job separation probabilities for employees working in similar workplaces. Finally, women's employment stability is relatively low because they are more likely to move from a job and into unemployment or out of the labor force, and less likely to make job-to-job transitions.
Employee turnover has been documented to be high in most countries, see Davis, Haltiwanger and Schuh (1996) and Davis and Haltiwanger (1999). Even though it is no surprise that a substantial part of the workforce leaves their workplaces every year, it is not a simple task to pinpoint the employees who are most likely to leave, and the workplaces from which employees are likely to separate. Improved knowledge about the job separation decision is important for employers in order to make workforce adjustments and develop retention policies that impose minimal disruption to the production process. Employees also benefit from this information because it allows them to make informed choices about where to work. More generally, this knowledge will help shape public policy targeting employment stability. The primary goal of the analysis is to study the job separation process with the purpose of answering two important questions. First, do men and women who experience similar working conditions have different job separation rates? Second, does employment stability differ for men and women? Previous studies of job separation processes have provided some insights into these questions, but data limitations have restricted their focus to information on individuals or workplaces. For instance, studies focusing on the individual component have documented the effects of human capital and demographic variables on the probability that an employee separates from the job (Blau and Kahn, 1981, Light and Ureta, 1992, Lynch, 1992 and Royalty, 1998). Parallel to these studies, Anderson and Meyer (1994) have analyzed how workplace characteristics influence the job separation probability. In this study, I will integrate these two lines of research using a register-based employer-employee data set. Identification of gender differences in job separation rates is infeasible without simultaneous information about employees and workplaces because of labor market sorting. Sorting takes place when matches between employees and workplaces are non-random, i.e., employees make directed search when looking for a job, and employers are selective in their choice of workforce. The sorting process will naturally lead to a segregated labor market where distinct groups of individuals work in different types of workplaces.1 Empirically, I find the tendency that women work in smaller low-wage workplaces with relatively high levels of job separations.2 In the analysis conducted below, I show that failure to recognize this labor market segregation will lead to biased estimates and incorrect conclusions about gender differences in job separation probabilities, which potentially leads to statistical discrimination. Furthermore, I argue that conventional statistical methods, such as the random-effects or fixed-effects models, are unable to eliminate the bias induced by omitted variables when the labor market is segregated. Instead, consistent estimates can successfully be obtained from employer-employee data. The focus on gender differences provides a series of important empirical results. First, women have higher unconditional job separation rates than men. Women's separation rates are also estimated to be significantly higher conditional on a large set of individual characteristics. Taking these findings at face value, women will face statistical discrimination in the labor market. This result arises because labor market segregation is ignored. Thus, heterogeneity in job separation rates across workplaces due to differences in workplace characteristics are picked up by individual characteristics leading to biases. A more comprehensive analysis of the employer-employee data shows that adding information about the workplace, such as the size of the workplace and the payroll class, has the consequence that the gender coefficient becomes insignificant. Hence, there are no significant gender differences in the job separation rates for employees working in similar workplaces.3 If employers recognize this result, statistical discrimination due to gender based on concerns about costly job separations should be absent from the labor market. The fact that gender is an insignificant predictor for job separations conditional on working conditions does not imply that men and women experience the same employment stability. The main reason for this is that the stability of employment matches and the employee's destination state subsequent to a job separation vary due to differences in both workplace and employee characteristics. This implies that the employment prospects may differ substantially for men and women in a gender-segregated labor market. To address this issue further, I estimate a multinomial logit model using information about the destination states following an employment match. The parameters from this regression are used to predict the labor market outcomes for the population of individuals currently working. The key results are that currently employed women relative to employed men are more likely to separate from a job (two percentage-points), experience a spell of unemployment and withdraw from the labor market. In addition, they are less likely to make job-to-job transitions. Hence, women's employment stability is clearly below men's. A decomposition of the two percentage-points gap in job separation rates reveals that 25 percent can be contributed to differences in individual characteristics and the remaining 75 percent to differences in the workplace component. These results emphasize that future labor market policies intended to equalize employment stability between men and women should focus not only on removing gender differences in individual characteristics, such as education levels, but also have considerable focus on eliminating differences in workplace characteristics, i.e., to reduce labor market segregation. The data used in the analysis is the Integrated Database for Labor Market Research (IDA). The database contains information on all employees from all establishments in all sectors in Denmark in the years 1980 to 2000. Each year (on a specific day in November) all employees and workplaces are merged, providing a snapshot of all employment matches in the Danish economy. Further, employees and workplaces carry unique identification numbers which enable tracking over time. Thus, the data provides a unique opportunity to study mobility patterns in the labor market. In the following analysis I focus on a sub-sample containing all private sector workplaces and their employees, corresponding to 3,253,312 unique employees and 477,619 workplaces, or 29,069,419 November-employment matches over the 20-year period.4 The Danish labor market shares many of the characteristics found in the UK and US labor markets. For example, all three countries have highly liberal labor market policies. However, they differ in one respect, namely in the generosity of unemployment benefit levels. To the interested reader Appendix A provides a detailed discussion of the institutional settings in Denmark and a comparison with the labor markets in the US and the UK. In the next section I present descriptive statistics of the data used in the analysis. Two points are made in this section. First, the level of job separations is high. Second, men and women are somewhat segregated in the labor market in the sense that women tend to be employed in smaller low-paying workplaces with relatively higher turnover. The latter plays an important role for the empirical findings in this paper. The theoretical framework for modeling the job separation process and associated labor market flows is presented in Section 3, and the estimation strategy is proposed in Section 4. In Section 5 I discuss and illustrate the empirical consequences for the job separation process when important variables are omitted, and the labor market is segregated. This analysis is extended in Section 6 to include information on the individual's destination states. The extended framework is used to advance our knowledge of gender differences in job separation rates and employment stability. The results an policy implications are discussed in Section 7. Section 8 concludes the paper.
نتیجه گیری انگلیسی
In this paper, I have studied the job separation process to learn about gender differences in job separation probabilities and employment stability. In contrast to earlier studies, which have focused on the importance of either individual or workplace characteristics, I use employer-employee data to obtain a series of new empirical results. The main finding in this paper is that there is no significant difference in job separation probabilities between men and women with similar characteristics working in similar workplaces. This is an important result because employers are concerned about the costs imposed by job separations. Therefore, even small differences in the expected likelihood of job separations across gender would lead to statistical discrimination. Given there are no systematic differences due to gender, these concerns should no longer be present. Hence, the observed segregation of men and women into different types of workplaces should not be due to statistical discrimination related to job separation decisions. A second important finding is that women currently working in the labor market experience relatively low employment stability. There are two reasons for this. First, men and women are to some extent segregated in the labor market, i.e., women tend to work in smaller low-wage workplaces with relatively low retention probabilities. Further, women have individual characteristics that are associated with relatively high job separation rates. Therefore, they experience job separations more frequently. Second, women have more transitions out of the labor market and into unemployment and make fewer job-to-job transitions. The analysis provides some scope for economic policy targeting gender equality in employment stability. For instance, policies eliminating the gender differences in observable individual and workplace characteristics may reduce the stability gap. Such policies include improvements of women's human capital accumulation and better childcare. Finally, this paper expands on previous research. For this reason, the analysis has deliberately been constrained, such that it clearly illustrates the information advantage of employer-employee data over previous studies focusing on either individual or workplace characteristics. This implies that additional studies of the job separation process and employment stability are warranted. For instance, the effect of individual wages on the job separation process remains to be established. It would also be interesting to understand the consequences of the revealed heterogeneity in job separation rates across both individuals and firms on investments in human capital. Thus, in addition to presenting a series of new empirical results, this study paves the way for future research.