The Behavioral Foundations of Labor Supply and Job Search: Experimental Evidence from an Urban Labor Market in Ethiopia

Last registered on March 30, 2018

Pre-Trial

Trial Information

General Information

Title
The Behavioral Foundations of Labor Supply and Job Search: Experimental Evidence from an Urban Labor Market in Ethiopia
RCT ID
AEARCTR-0002555
Initial registration date
March 19, 2018

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
March 19, 2018, 3:18 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
March 30, 2018, 12:12 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
University of Oxford

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2018-03-22
End date
2019-01-31
Secondary IDs
Abstract
This project seeks to provide new evidence on the behavioral foundations of labor supply, job search behavior, and occupational choice. I focus on impatience. I use two lab-in-the-field experiments and high-frequency phone surveys over thre months to elicit individual time preference parameters and a range of other individual characteristics, including cognitive control, non-cognitive ability, and aspirations, from a sample of young women who start their job in a light manufacturing firm in highly dynamic urban labor market in Ethiopia. I augment this data with rich administrative data from firm personnel records and bank account transactions. I use the resulting panel dataset to structurally estimate an on-the-job search model for workers with hyperbolic preferences. I hope to use the structural estimates to study counter-factual scenarios of labor supply and firm tenure, job search behavior, and occupational trajectories in the labor market under various degrees of present bias.

I have completed piloting the survey and intervention. No baseline data has been collected.
External Link(s)

Registration Citation

Citation
Meyer, Christian Johannes. 2018. "The Behavioral Foundations of Labor Supply and Job Search: Experimental Evidence from an Urban Labor Market in Ethiopia." AEA RCT Registry. March 30. https://doi.org/10.1257/rct.2555-3.0
Former Citation
Meyer, Christian Johannes. 2018. "The Behavioral Foundations of Labor Supply and Job Search: Experimental Evidence from an Urban Labor Market in Ethiopia." AEA RCT Registry. March 30. https://www.socialscienceregistry.org/trials/2555/history/27465
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
The focus of this study is to experimentally elicit individual time preference parameters and a range of other individual characteristics, including cognitive control, non-cognitive ability, and aspirations, from a sample of young women who start their job in a light manufacturing firm in highly dynamic urban labor market in Ethiopia.

Additionally, I randomly administer a light-touch information intervention in which a subset of surveyed workers is informed about their own time preferences as measured in baseline survey.

In the same context, I cross-randomize an unrelated intervention which is separately registered in the AER RCT Registry.
Intervention Start Date
2018-03-22
Intervention End Date
2018-08-31

Primary Outcomes

Primary Outcomes (end points)
My intermediate outcomes from the lab-in-the-field experiment are individual-level time preference parameter estimates for all my subjects.

I then use these parameters as inputs to a structural model of on- and off-the-job search to study the following final outcomes: (1) the labor supply decision on the new job, (2) on-the-job search behavior, (3) tenure at the firm, (4) job search behavior while unemployed after leaving the firm, (5) reservation wages while employed and while unemployed, and (6) occupational choice after leaving the firm.
Primary Outcomes (explanation)
Individual-level time preference parameters are estimated in a beta-delta framework from a fully incentivized convex time budget experiment (Andreoni and Sprenger 2012) with large stakes (on average one third of monthly income). As a secondary measures of time preference, I also estimate dynamic inconsistency in real effort (Augenblick, Niederle, Sprenger, 2015) from from two batteries of five unincentivized hypothetical questions. As a tertiary measure of time preference, I use the deferment of gratification (DGS) scale which measures the individual ability to control impulses and to plan and wait. The DGS scale is a simple additive measure.

I revisit all measures in the endline survey three months after the baseline.

Labor market outcomes are measured and constructed as follows (numbers refer to field above):
(1) labor supply decision on the job: Standardized daily productivity data from firm personnel records.
(2) and (4) job search behavior: I use high-frequency phone calls to collect data on various measures of job search effort (intensity, number of phone calls, amount of time). I then create an index of these measures.
(3) tenure at the firm: A binary variable based on high-frequency phone calls and personnel records.
(5) reservation wages: Directly elicited in baseline interview, high-frequency phone calls, and endline interview.
(6) occupational choice: High-frequency phone calls that collect data on sector, wage, and amount of time worked in potential next job, or else on time spent in unemployment.

Secondary Outcomes

Secondary Outcomes (end points)
While my main interest lies in the structural estimation of a behavioral job search model, my setup can speak to four secondary groups of outcomes that I am planning to study using observational data:

- Hypothesis Group A “Confidence and Job Search”: The role of baseline locus of control, aspirations, and self-confidence ("self-efficacy") in search effort and the potential psychological costs (beyond effort) from unsuccessful search.
- Hypothesis Group B “Cognitive control, cognitive depletion, and information processing”: The role of cognitive control and its depletion in job search.
- Hypothesis Group C “Industrial Work and Aspirations”: The effect of indus- trial work on aspirations and self-confidence.
- Hypothesis Group D “Industrial Work and Health”: The effect of industrial work on mental and physical health outcomes.
Secondary Outcomes (explanation)
The submitted Pre-Analysis Plan specifies further details on the measurement of these secondary outcomes.

Experimental Design

Experimental Design
The main focus of this study is to construct a longitudinal dataset of individuals starting employment at a large manufacturing firm. I collect data in four steps: First, I conduct a lab-in-the-field experiment and baseline survey. Second, I follow workers using biweekly phone calls. Third, I repeat the lab-in-the-field experiment together with an endline survey. Fourth, I obtain administrative data from firm personnel records and on workers' bank accounts.

At baseline, all study participants will be randomly assigned to one of three groups: (1) a SURVEY CONTROL in which I conduct baseline, high-frequency follow-up and endline surveys, (2) a TREATMENT group in which I additionally administer a light-touch information treatment, and (3) a PURE CONTROL group in which I do not conduct any surveys or administer any treatment, but only rely on administrative data on each subject.

Details are set out in the submitted Pre-Analysis Plan.
Experimental Design Details
Randomization Method
I randomize among all individuals starting employment in an industrial park outside Addis Ababa. Workers apply for work at the factory gate Monday through Saturday (t=-1), and I randomize them into treatment arms on the next day when they return for their first day of work (t=0). The randomization is done by the tablet computer used for interviewing.

All interviewed subjects are eligible to win the payouts from a convex time budget experiment with a 50 percent chance. Randomization for this payout is done using a coin flip at the end of the interview. If subjects win the coin flip, they draw one of the 15 decisions that they made to determine the payout dates and amounts. This is done using numbered cards in a bowl.
Randomization Unit
I randomize at the individual level among workers who start employment in an industrial park outside Addis Ababa.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
Target sample size of 450 garmenting workers as part of the main study, with an additional 400 workers for which I only collect administrative data.
Sample size: planned number of observations
Target sample size of 450 garmenting workers as part of the main study, with an additional 400 workers for which I only collect administrative data.
Sample size (or number of clusters) by treatment arms
Target sample size by treatment arm as follows:
- 225 workers in SURVEY CONTROL (with baseline, midline, and endline surveys)
- 225 workers in TREATMENT (with baseline, midline, and endline)
- 400 workers in PURE CONTROL (no surveys, no treatment, only administrative data)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
New York University Abu Dhabi
IRB Approval Date
2018-03-25
IRB Approval Number
#020-2018
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials