Deconstructing Job Search Behavior

Question

1. What are the detailed information available for job seekers in the dataset from www.trabajando.com?

2. How do job ad attributes attract applications?

3. How does the residual term o relate to ad unobserved features?

4. How do employment status and weighting methods affect job seekers' application frequency?

7. What is the relationship between misalignment in continuous dimensions and application decisions?

Accepted Answer

The dataset from www.trabajando.com provides detailed information on job seekers, including date of birth, gender, nationality, place of residency ('comuna' and 'region'), marital status, years of experience, years of education, college major, and name of the granting institution of the major. Additionally, the dataset includes information on the occupational area of the current or last job, salary, starting and ending dates, required level of experience, required college major, required skills, number of positions to be filled, occupational code, geographic information, firm size, and industry. Job seekers can also record their expected salary, which they can choose to show or hide from prospective employers. The dataset covers a sample of job postings and job seekers between January 1st, 2008, and December 24th, 2016, with more than 14 million single applications from around 1.5 million job seekers to around 270 thousand job ads.

Accepted Answer

This section analyzes how different job ad attributes attract applications from diverse job seekers. It focuses on determining relevant job ads for individuals in the sample, despite the lack of data on total searches or clicks. The study addresses the 'consideration sets' problem, which is common in marketing and industrial organization literature. The network-revealed preference approach is proposed as a potential solution, assuming access to databases that track consumer purchases. By understanding the factors that attract applications, employers can tailor job ads to better match job seeker preferences, potentially increasing application rates and improving the hiring process.

Accepted Answer

The residual term o may be correlated with ad unobserved features. This occurs when the applicant w and some others possess private information about the posted job that is not disclosed in the ad. Since we control for a large array of ad characteristics, this correlation may only occur if there is private information not disclosed in the ad. This is important to consider when analyzing the impact of ad characteristics on job applications.

Accepted Answer

Table 3 reveals the impact of employment status and weighting methods on job seekers' application frequency. Unweighed estimates show consistent signs but reduced magnitudes, suggesting that weighting methods diminish the influence of irrelevant job ads. Unemployed job seekers tend to apply more frequently to ads in their consideration sets compared to employed individuals. Among the employed, married individuals apply less than their non-married counterparts, although the difference is not statistically significant. Male job seekers, particularly the unemployed, apply more often to ads in their consideration sets, considering other applicant and ad characteristics constant. This data indicates that employment status and weighting methods play a significant role in job seekers' application behavior.

Accepted Answer

Individuals who explicitly state their wage expectations at the time of application apply more often, especially if unemployed. The table shows that an explicit wage in the job ad negatively affects the decision to apply on average, but the effect is weaker for the unemployed. This aligns with Banfi and Villena-Roldan (2019) findings, suggesting that ads with hidden wages attract more applicants due to potential wage flexibility or bargaining. Unemployed individuals significantly increase their application probability by 0.04% when the number of vacancies increases by one, while employed individuals do not show a significant reaction. The positive response to a marginally higher likelihood of receiving an offer indicates the importance of selection on the employer side, such as non-sequential employer search or signaling for poorer job conditions. The negative effect of perceived 'age' of job ads on unemployed individuals can be related to stock-flow matching behavior, where new job seekers apply to the stock of job ads. This behavior is consistent with the findings of Gregg and Petrongolo (2005) and Coles and Petrongolo (2008). The evidence also suggests that unemployed individuals react to 'phantom' ads, which may be filled positions by the time of application, as seen in Albrecht, Decreuse, and Vroman (2023) and Cheron and Decreuse (2016). The effects are not significant for the employed, indicating a different search pattern for this group on the website. The figure is computed using a polynomial of order 5 on each variable, with the rest of the regressors set at their sample mean.

Accepted Answer

The application probability for the unemployed increases with age, peaking at age 45 and then decreasing until mid-fifties. For individuals under 30, the application probability is higher, decreasing until mid-forties, and then slightly increasing. This pattern is observed in the left panel of the figure, where the unemployed apply more often to ads in their consideration sets at all ages. The evidence suggests that job finding rates and employment-to-employment transitions over the life-cycle may not fully align with these findings, as reported by other researchers. However, the self-selected sample of older workers using the online job board may differ from the average worker in the labor force of that age. Additionally, the application probability for employed jobseekers remains relatively flat, with a slightly increasing trend up to 400 weeks, or nearly eight years. This could be due to factors such as growing match-specific human capital and market-learning processes. Overall, the relationship between age and application probability is complex and influenced by various factors.

Accepted Answer

The relationship between misalignment in continuous dimensions and application decisions is represented by an inverted U-shaped relationship for education, experience, log wages, and a mostly decreasing line for regional distance. The application probability peaks at zero when the applicant is exactly aligned with the advertised requirements. However, the shapes of employed and unemployed are asymmetrical. Employed seekers are more ambitious and daring to apply to jobs for which they are underqualified in terms of education. For the unemployed, the experience dimension curves peak around -7 and show a steeper decline to the right, indicating that job seekers tend to have more than seven years of experience than the minimum required and do not refrain from applying if they are even more overqualified in experience. The application probability curve for the employed peaks a bit to the left of the one for the unemployed, and the gap between the two groups becomes wider for experience gaps between -15 and 0. The plot in the lower-left panel reveals that differences in log wages greatly affect application probabilities, with the unemployed fluctuating between 12% and 25% for, while for employed seekers, the range is wider, from around 7% to 25%. Overall, the unemployed curve surpasses the employed counterpart for negative gaps, indicating that the unemployed are more likely to apply to jobs for which they are overqualified in terms of productivity. For jobs for which applicants are underqualified, employed applicants apply more often to the right of the peaks. The lower-right panel depicts the predicted probability as a function of the distance between the regional capital of the applicant and the regional capital of the job in hundreds of kilometers, showing a decrease in the likelihood of an application as distance increments. The employed-unemployed gap is only noticeable in mid-distance applications.

Accepted Answer

The intensive margin of job search behavior is influenced by various factors. Unemployed individuals tend to apply more than employed individuals. Males, especially if unemployed, also apply more. Employed married individuals apply less compared to their single counterparts. The elapsed job ad posting duration negatively affects the likelihood of application for the unemployed, but not for the employed. The intensive margin increases with age for the employed, possibly due to an effort substitution effect or self-selection into job boards. Effort decreases as the job search duration increases, but tenure has a mild positive effect. The cyclicality of the intensive margin is nuanced, being insensitive to unemployment rates below the median but highly procyclical in regions with high unemployment. These factors contribute to a nuanced understanding of job search behavior.

Accepted Answer

Choice sets can be defined by segments based on job characteristics, clustering algorithms, or geographical location. However, the latter may not be sufficient for high-dimensional attributes like job ads. The definition of choice sets varies among researchers, with some using arbitrary sets of job characteristics and others using clustering algorithms. The classification of choice sets is not standardized and may not align with formal taxonomies such as SOC or ISCO. In the context of the Chilean labor market, choice sets may also be influenced by industry classifications created by website administrators, which may not match formal taxonomies like NAICS or ISIC. Overall, the definition of choice sets is subjective and can vary based on the researcher's approach and the specific context of the study.

Deconstructing Job Search Behavior

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What are the detailed information available for job seekers in the dataset from www.trabajando.com?

2. How do job ad attributes attract applications?

3. How does the residual term o relate to ad unobserved features?

4. How do employment status and weighting methods affect job seekers' application frequency?

5. How do explicit wage expectations affect application rates?

6. How does age affect application probability?

7. What is the relationship between misalignment in continuous dimensions and application decisions?

8. What factors influence the intensive margin of job search behavior?

9. How are choice sets defined?

References

Job Matching and the Theory of Turnover

Labor Market Institutions and the Distribution of Wages, 1973-1992: A Semiparametric Approach

Marketplaces and Matching

How Local Are Labor Markets? Evidence from a Spatial Job Search Model

Frictional Wage Dispersion in Search Models: A Quantitative Assessment

Related Papers (5)

Accumulation of technological knowledge, wage differentials, and unemployment

Wage equations, wage curves and all that

Dynamic wage bargaining if benefits are tied to individual wages

Wage curve, unemployment duration and compensating differentials

Cyclical Variations in Wage Differentials and Unemployment