Intervention (Hidden)
We propose to work with a firm to hire professional HR workers, who will be tasked with reviewing and screening resumes. For HR workers who respond to our job ad, our experimental design involves three dimensions of randomization. For each randomization, we describe the treatment and the outcomes we seek to link to the randomized treatment.
Our first randomization relates to the use of a noncompete in the job offer. The second randomization is how salient the noncompete is in the onboarding materials. And the third randomization is the wage offer. We describe the six treatment and control groups below.
1. No Noncompete, Low Wage: We offer the worker an employment contract including a nondisclosure agreement (NDA). No noncompete agreement would be included. The offered wage would be the 25th percentile of the wage distribution, which is approximately $25/hr.
2. No Noncompete, High Wage: Same as condition 1, except the offered waged is the 75th percentile of the wage distribution, approximately $60/hr.
3. Noncompete, Full Transparency, Low Wage: The worker receives an employment contract including an NDA and (separately) a clearly labeled noncompete agreement in a standalone document. The offered wage will be the 25th percentile of the wage distribution (~$25/hr). This condition includes mentioning the noncompete in the initial job post. We refer to this as our "full transparency" condition because there is no way the worker can sign the employment agreement without seeing the noncompete.
4. Noncompete, Full Transparency, High Wage: Same as 3, but with the high wage (~75th percentile of wage distribution, $60/hr).
5. Noncompete, Low Transparency, Low Wage: Same as 2, except the worker receives an employment contract that contains (in one document) both the NDA and the noncompete provisions. This is less transparent treatment as the noncompete is inside a larger contract, requiring the worker to notice it. The worker does not need to sign the noncompete separately.
6. Noncompete, Low Transparency, High Wage: Same as 5, except for the high wage.
In Phase 2, the poaching firm randomizes wages at $27/hr and $62/hr when soliciting the workers hired by Firm A.
Our experiment included a third phase related to releasing the workers early from their noncompetes.
The third phase of the experiment involves randomly releasing the workers from their noncompetes, and measuring their applications to another job. As workers are approximately 3-4 months into their 6-month noncompete, two cross-randomized treatment schedules begin. On each day, a few workers are randomly shown the job ad. Meanwhile, another set of randomly selected workers are sent a message releasing them from their noncompetes (if they have it) and reminding workers of their nondisclosure obligations. Workers who did not have a noncompete still received a message reminding them about their nondisclosure agreement.
This process continued until all workers were both i) sent the remind/release message, and ii) shown the job ad. Through this procedure, the sequence of the two events were randomized. Some were randomly released before seeing the ad, and some were randomly released after seeing it. We track who applies for the job posting every day. This helps measure whether noncompete release encourages workers to apply for jobs.
Due to a clerical error, our initial pre-registration did not include a passage about this phase of the experiment. However, the experiment above was approved by the Maryland IRB before the beginning of the third phase of the experiment, describing the release and goals.