The Impacts of Preference Signaling on Job Matching: Experimental Evidence from a Job Fair

Our intervention is a job preference signaling mechanism that allows job applicants to express their interest to few employers. This mechanism is essentially identical to the one that has been implemented by the American Economic Association in the Job Openings for Economists. Below are specifics about our intervention:
- Each jobseeker who participates in the job fair is endowed with two signals, which we call “High Priority Signals (HPSs).”
- Timing: Jobseekers can send HPSs when and after they apply for jobs. HPSs sent by jobseekers will be delivered to employers immediately.
- HPSs submitted cannot be withdrawn once employers log into their user account pages.
- A jobseeker can send at most one HPS to an employer. That is, s/he cannot send two HPSs to the same employer.

The setting in which the signaling mechanism is implemented is an online job fair. The fair is organized by a large online job portal in Bangladesh. We conduct another RCT that estimates the impacts of an online job fair. This RCT is registered as another study in AEA registry.

2021-11-09

2021-11-11

[Employer-level outcomes]
- Whether a position is filled; the number of hires
- Job match quality: (i) monthly salary; (ii) employer’s subjective assessments on hired workers’ skills and motivation; (iii) employer’s overall satisfaction on hired workers; (iv) worker retention
- Employer’s satisfactions on recruitment experience and outcomes

[Jobseeker-level outcomes]
- Being employed
- Job match quality: (i) monthly earnings; (ii) hourly wage; (iii) how often worker’s skills are used; (iv) how well worker’s skills fits his/her job; (v) whether a worker is under- or over-qualified; (vi) how interesting the current job is; (vii) how satisfied a worker is with his/her current job; (viii) job tenure

[Job application-level outcomes]
- Whether an applicant is hired

We will create an index of job match quality.

[Employer-level outcomes]
- The number of offers declined

[Jobseeker-level outcomes]
- The number of job applications viewed; the number of job applications shortlisted; the number of job offers received
- Reservation wage and aspired wage in 5 years
- Life satisfaction
- Whether looking for a job in last 4 weeks
- Self-perceived job search prospects: the percent chance of finding a full-time job within 6 months

[Job application-level outcomes]
- Whether an application is viewed
- Whether an applicant is shortlisted
- Whether an applicant is offered the job

Our study subjects are all jobseekers who are invited to and sign up to the job fair and apply for at least one job and all employers who participate in the job fair. All study subjects are users of the online job portal that organizes the job fair.

To make jobseekers understand the signaling mechanism, the organizer of the job fair holds online sessions and explain what it is and how to use it. The sessions are held multiple times prior to and during the job fair. The organizer explains the mechanism to employers individually.

Jobseekers can send HPSs to the maximum of two employers. They can send HPSs either when or after they submit job applications. Specifically, when they are about to submit job applications on the job fair website, they are asked whether they attach HPSs unless two HPSs are already exhausted. After they submit job applications, they can also send HPSs from their account page on the job fair website.

On employers’ account page, employers can view whether job applicants send HPSs to their jobs. (Note that since each employer advertises exactly one job in the fair, sending HPSs to employers is equivalent to sending HPSs to jobs.) More specifically, on the employer account page, employers can first view the list of all applicants. This list shows summary information about each applicant including whether s/he sent a HPS. Then, if employers click an applicant from the list, employers can open his/her application package and view his/her detailed information.

The data are the following:
1. Administrative data of the job fair: It includes the information about jobs, employers, jobseekers, applications, and HPSs.
2. Administrative data of the online job portal: The data includes the information about jobs posted by our study subject employers, the applicants to the jobs, and jobs that are applied by our study subject jobseekers.
3. Baseline survey data: The baseline jobseeker data is based on the online forms submitted by jobseekers to apply for the job fair. The baseline employer data is collected by the online survey too.
4. Endline survey data: We conduct endline surveys of jobseekers and employers twice, approximately 2 months and 9 months after the job fair. The modes of surveys are phone for jobseekers and in-person for employers. If in-person interviews are infeasible, interviews will be on zoom or phone.

[Randomization]
Our randomization is only across employers. There is no randomization across jobseekers. Every jobseeker is allowed to send HPSs to two employers. While HPSs can be sent to any employers, they are disclosed only to randomly selected employers, namely, treatment employers. HPSs that are sent to the other employers (control employers) are not disclosed to them. Control employers are not notified of the signaling mechanism at all, so they are presumably not aware of the mechanism. Jobseekers are not informed that employers are randomized or that HPSs are disclosed only to treatment employers.

[Source of Empirical Identification]
The impacts on employers will be estimated based on comparison between treatment and control employers. Those on jobseekers will be estimated by comparing jobseekers who send HPSs to treatment employers and those who sent HPSs to control employers, because the de facto number of HPSs is ex post greater for the former jobseekers than the latter.

[Analysis]
a) Employer level: We will regress outcome variables on employer’s treatment status.
b) Jobseeker level: We will regress outcome variables on the number of signals sent to treatment employers and the total number of signals sent. Essentially, among those who used the same number of signals, we will estimate whether the outcome variables differ depending on how many signals are ex post effective (i.e., sent to treatment employers) and how many are ex post wasted (i.e., sent to control employers.
c) Job application level: We will regress job application outcomes (e.g., whether an applicant is viewed, shortlisted, and offered the job) on the interaction between the dummy indicating whether the applicant sent a signal to the job and the employer’s treatment status. The control variables include the dummy whether the applicant sent a signal to the job and the fixed effects of employers. Essentially, among applicants for the same job, this regression estimates the differences in the probabilities of an applicant being viewed and shortlisted between those who sent a signal and those who did not when signals are disclosed to employers by controlling for the differences in the probabilities when signals are not disclosed.

[Ethical Consideration]
Jobseekers are treated equally ex ante since everyone is endowed with the same number of HPSs. HPSs sent to control employers are essentially wasted. However, jobseekers are still treated equally ex post in the sense that the competition between job applicants to control employers with and without HPSs is not affected by the signaling mechanism. Since the mechanism does not disclose HPSs to control employers, it does not affect selection decisions of control employers.

Randomization done in Stata

Firms

No

110 firms
1280 jobseekers

110 firms 1280 jobseekers

55 treatment firms; 55 control firms; 1280 jobseekers