Group Identity and Dynamic Principal-Agent Problem: Theory and Experimental Evidence

Last registered on October 13, 2025

Pre-Trial

Trial Information

General Information

Title
Group Identity and Dynamic Principal-Agent Problem: Theory and Experimental Evidence
RCT ID
AEARCTR-0016938
Initial registration date
October 08, 2025

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
October 13, 2025, 10:13 AM EDT

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

Locations

Region

Primary Investigator

Affiliation
National Taipei University

Other Primary Investigator(s)

PI Affiliation
National Taipei University

Additional Trial Information

Status
On going
Start date
2025-10-01
End date
2025-10-31
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
We investigate the impact of group identity on the labor market using a lab experiment. Our experimental design is structured as a group classification procedure followed by a matching protocol that pairs employers and employees, who are played by subjects. We select a standard theoretical framework that incorporates group identity and analyze its properties to derive testable implications and to elicit subjects' strategic behavior during the game.
External Link(s)

Registration Citation

Citation
Chen, Chun-Ting and ChienHsun Lin. 2025. "Group Identity and Dynamic Principal-Agent Problem: Theory and Experimental Evidence." AEA RCT Registry. October 13. https://doi.org/10.1257/rct.16938-1.0
Experimental Details

Interventions

Intervention(s)
We vary the other players that the subjects will interact in the experiment.
Intervention (Hidden)
This investigation builds on a canonical principal-agent game described as follows. The Principal (P henceforth) first determines a wage offer for an Agent (A henceforth) based on the outcome (w50,w100). If the final outcome is 50, A receives the wage w50, and P receives (50-w50); if the final outcome is 100, A receives the wage w100, and P receives (100-w100). After A sees the offer, A can choose (i) whether to receive the offer and (ii) which action to take. If A chooses to reject the offer, both P and A receive 0. If A accepts, A can then choose either to take High effort action of Low effort action. High effort yields an 80% chance to get an outcome of 100 and a 20% chance to get an outcome of 50, while A needs to spend 20 as the cost. Low effort yields a 50% chance to get an outcome of either 100 or 50, while A needs to spend 13 as the cost. P cannot observe A's action directly.

We follow the Chen and Li (2009) grouping design to classify participants into two groups based on their preference for two famous artists, Paul Klee (1879-1940) and Wassily Kandinsky (1866-1944). There are seven rounds. For each round, subjects are asked to choose the preferred painting from a pair of paintings, one of which was painted by Klee and the other by Kandinsky. A participant who chooses four or more of Klee’s paintings will be classified into Group Klee; otherwise, he/her will be classified into Group Kandinsky. This classification will remain unchanged throughout the entire experiment. A participant will know which group they are in and the number of group members, as their screens display this information. To strengthen the group identity, we further ask the members of the same group to play 3 puzzles, where the group members can collaborate with a chat box.

Then we ask the subjects to play the principal-agent game as described. The subjects will be reminded whether they are paired with the same group, a different group, or an unknown group (not reminded), depending on the treatment condition of the session. We assume that the agents will do favors for the principals and are willing to choose high effort with a worse wage offer. Given this, the principals can take advantage of this and offer a lower price to agents of the same group.

To investigate the validity of this mechanism, we elicit the Agents' reaction when seeing offers from the Principals from the same/different/unknown group. Furthermore, we elicit the Principals' beliefs on how the Agents will react to see whether they exploit the behavior changes due to group identity. Lastly, to control the risk preference and other-regarding preference, we conduct a Holt-Laury style risk task and a dictator game.
Intervention Start Date
2025-10-01
Intervention End Date
2025-10-31

Primary Outcomes

Primary Outcomes (end points)
Principals: wage offers, beliefs of agents' actions
Agents: decisions on offers (reject decision and effort level decision)
All subjects: risk attitude, contributions in the dictator game
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Time spent on each page, responses in chatbox or surveys.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The experiment would undergo a group classification procedure followed by a matching protocol that mimics the employer-employee matching process in the labor market. Subjects are to play a canonical principal-agent game where a part of the subjects act as principals (employers) and the other part as agents (employees). The principals design the contracts, and agents decide whether to accept and the corresponding actions.
Experimental Design Details
Before each session starts, the experimenter will randomly decide whether this session will be a "same-group", a "different-group", or an "unknown-group" session. In a same-group session, the subjects will be reminded they are playing with the Principals/Agents from the same group assigned in the artist preference phase. Throughout the session, whenever they make decisions regarding other participants (offers, reactions to offers, or the dictator game), they are always matched with a member from the same group. The different-group is similar, while the subjects are paired with members from the other group. In an unknown-group session, the subjects will not be hinted about the opponent's group identity at all; they play the games as if there is no group assignment.
Randomization Method
Done by computer random draw before each group of sessions start. In consecutive sessions, if the first is assigned as the same-group, then the other is a different-group, and vice versa. An unknown group will be followed with a pair of same and different group treatments to ensure the subjects in the same/different treatments are adequate.
Randomization Unit
Experimental sessions
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
15 experimental sessions
Sample size: planned number of observations
300 subjects
Sample size (or number of clusters) by treatment arms
100 per each arm of same-group/different-group/unknown-group treatments.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
National Taiwan University Behavioral and Social Sciences Research Ethics Committe
IRB Approval Date
2024-09-24
IRB Approval Number
202407ES024
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