Gender Effect in Explaining the Mobility Patterns in the Labor Market: A Case Study of Turkey
Abstract
This paper examines the importance of gender for different job mobility patterns using an extensive household survey data from İzmir, third largest city in Turkey. The determinants of job-to-job and job-to-non-employment transitions are analyzed with the help of a multinomial logit estimation method. The results indicate that there is a distinction regarding the probability of job mobility patterns based on gender. It is more likely for women to be engaged in job-to-non-employment transition, whereas men tend to change jobs more often. Although gender plays a significant role in job mobility patterns, traditionally imposed social constraints associated with childcare and household duties provide us with mixed results considering the behavior of women in the job market. On the other hand, having high-paid and secure jobs decreases the probability of both patterns of job mobility.
The paper is on a very important subject addressing gender differences in job to job, or job to non-job transitions. On the whole, however, there are many weaknesses to be addressed. I will point out some of these in the hope that they will provide helpful guidance in improving the paper. It would be worthwhile for the authors to attempt to improve the paper given the importance of the subject and the richness of the data set they have access to.
There are many issues to be addressed, however, I will only point out several of them due to the limitation of the space:
1. On page 6, where they summarize the most important findings, they point out that better paid men are less likely to change to change their jobs, however, they use men’s own wages for the JJ category and wages are endogenous as they point out later in the paper. In the same sentence there is a reference to men’s job security, however, there is no variable that captures that in their estimation. How do they reach that conclusion?
2. They say that the result for JNE for married women is statistically significant, while the table indicates otherwise.
3. In the equations they have number of dependents and marital status. These two variables are likely to be correlated. I would suggest that they divide their sample into married and unmarried categories. It would also be desirable to divide the dependents by age groups as is done in most labor supply estimates.
4. The result that women who are household heads are more likely to leave the labor force is surprising. One would have expected that being a household head they would have more financial responsibilities thus they would stay in the job market.
5. The authors use education and skilled variables in the same equation. Being skilled and being educated usually are highly correlated. Instead of using skilled vs. unskilled categories it would be more desirable if they create an education variable into several categories by level of education.
6. Why the authors do chose to use a logit model instead of an alternative one such as probit model?
This is an interesting and throughly written paper that contributes to the literature of Turkey’s gender mobility in the labor market. As the paper stands, it could be a useful paper to the literature for some reasons.
First and foremost, the author(s) is certainly correct in calling out Royalty and Theodossiou for their appropriate comparison of disaggregating data by education. However, it would be a mistake to use these kinds of insentive elucidating to ignore the most important gender formation issue facing the labor market: what if the result will not change although the appropriate social policies are applied? But unfortunately, the primary voices rising about this concern in the literature are very limited and since the policies are not adopted, it is difficult for the authors and researchers to focus on this issue.
As is quite well known, the declining in job-to-job mobility rate for women is significantly the result of marginalizing of a large swath of working men. Without stable employment, it is quite understandable why many working women are willing to leave the labor force. All in all, economists and sociologists are concerned with these signals of gender inequality generated by occupational segregation and social justice. Therefore, my own approach to this pressing issue is to premeditate how to lower the ratio of job-to-non employment that many women face. On the other hand, in my opinion, the main point of this article is to encourage feminists, institutions and the government to view this “employment crisis” in working class women as not simply a revisiting of patriarchal, social and traditional values by regressive forces and to find ways of improving the chances of working class children by providing adequate child care and sufficient securities in the light of discrimination of women both at the workplace and home. As a consequence of women’s mobility from having a job to being unemployed, it could be indicated that women’s roles are trapped between home and labor market activities.
On the other hand, there is a critical issue in the paper as a reader. The data used in the text has to be updated. 2010 data was used in the analysis and the author(s) can update the data to make an effective comparison. However, the purpose of the paper is clear. But as the author(s) says “This is the first paper to account for both job-to-job and job-to-non-employment transition as separate mobility patterns in a regional Turkish labor market context”, paper does not provide unique results. Since the analysis is run with lack of data set like disaggregate data in education which is indeed one of the main indicators for women’s employment, unfortunately the results of the estimations are unilateral. Nevertheless, there are some riddles with incomplete sentences that make it extremely difficult to comprehend the points being indicated by the author(s). The reader is constantly left to guess the authors’ intended message.
Reading Eryar and Tekguc’s paper titled ‘Gender Effect in Explaining the Mobility Patterns in the Labor Market: A Case Study’ I was much encouraged by the policy implications they reach i.e. the need in Turkey for public provisioning of care, need for child care support and sufficient security in the presence of discriminatory practices both at the workplace and home. This is required reading for everyone who is working on the current debates: women’s labor force participation and employment in Turkey. It is placed high on top of policy agenda for economic growth, women as the untapped resource. I believe the findings of the paper are specifically important within this context.
Eryar and Tekguc explores whether gender matters in analyzing job mobility patterns in Turkey. They use a special household survey with a very large sample conducted in 2010 in Izmir, one of the major cities in the country. The paper contributes substantially to the discussions on labor market transitions between different states in Turkish labor market. Unlike prior studies in the literature, in this paper job-to-job (JJ) transitions is the main focus. Results they find suggest that women are more mobile than men in the labor market but most of the mobility is due to leaving the job market. They also find some mixed results for women with respect to the effects of traditionally accepted social roles such as household duties or child care (measured by the variables such as marital status and number of dependents) on mobility patterns. I think among the estimation results the following are of particular importance: being married emerges as statistically significant by raising the likelihood of women’s transition from job to non-employment position (JNE). In contrast, having dependents in itself does not affect the probability of leaving the job market. On the other hand, women with dependents tend to stay in the same job, while being married does not seem to be statistically important in this mobility pattern. With these they conclude results point to need for a wider conceptualization of gender roles not necessarily limited to and captured by variables such as marital status and the number of dependents.
Regarding the methodology used, I’d like to raise a couple clarification questions perhaps it could be useful if brief explanations are included in the paper. For example why did the authors selected common explanatory variables for both JJ adn JNE in their model. I think that the determining factors and conditions could be totally different. At the outset JJ requires controls for demand factors but as mentioned in the paper it was not possible to add demand factors due to a data issue. Beyond that JJ pool includes individuals who have found a new job independently of their state in the previous year in the job market. But depending on these different initial states, the household and invidiual characteristics of the individuals in this pool could vary significantly. A similar question could be asked about JNE pool. It is explained that based on the selected explanatory variables results do not show statistically significant differences but what would be the case if for example determining factors vary for the group moving from the unemployed state to employed and for out of labor force to employed. A comprehensive review of the literature on women’s employment in Turkey may provide some guidence about other country specific explanatory variables. It could also be useful if a more detailed information is provided about the household survey used in research, like the specific time period over which the data is collected, what percentage of the sample represents urban/rural areas and similar information. That could help understand certain findings better. On that note, I’d like to extend my congratulations to the authors for their contributions on such critically important issue of Turkish labor market both for researchers and more so for policymakers.