General Training and Reciprocity

We will have 4 treatments in a two-by-two design:
Positive-Exogenous, Negative-Exogenous, Positive-Endogenous, Negative-Endogenous

Endogenous (Exogenous) refers to whether the the training level is determined by Employers (randomized device)
Positive (Negative) refers to whether the worker's unobservable effort benefits (harms) Employer

The "unobservable effort" by Workers is operationalized through the completion of real effort tasks, represented by tables of work. In the positive intervention, each table contributed by a Worker increases Employer's chance of receiving a high-level additional income by 5%. Conversely, in the negative intervention, each table contributed effectively reduces this probability by the same margin.

2023-11-18

2024-12-31

We will collect the following data from each treatment:
Number of real-effort task completions by Workers.
Training level and wage decisions made by Employers.
Effort contribution and Minimum-Acceptable-Wage (MAW) decisions from Workers.

Effort contribution is measured by the number of real-effort task tables that Workers are willing to complete to benefit or harm Employers. Both the effort contributions and the Minimum-Acceptable-Wage (MAW) decisions are recorded across each of the five training levels.

Our study incorporates several phases, beginning with a practice session for Real Effort Tasks and a characteristic-based questionnaire to match subjects. The core of the experiment involves subjects assuming 'Employer' and 'Worker' roles, where each subject engages in a full strategy method, making decisions both as an Employer and as a Worker before knowing their actual role. In this phase, Employers invest in training at various levels, each associated with a specific cost, and simultaneously propose new wages. Correspondingly, Workers make unobservable efforts that influence Employer profits based on different training levels. They also determine their minimum acceptable wage for each training level, which affects the income of both parties.

Once roles are assigned, Workers proceed to the real-effort tasks phase. The experiment concludes with a post-experiment survey designed to assess participants' social value orientation and their inclination towards positive or negative reciprocity

Matching Mechanism: Subjects are paired using a greedy algorithm based on the scores from characteristic questions, where the smaller the score differences, the higher the similarity between paired subjects.

Detailed Parameters of Payoffs:
Employer's Payoff: Consists of two components. The productivity payoff, is calculated as 150 - W2 + 80X if the Worker stays, and 100 - 20X if the Worker quits (where W2 represents the new wage and X the training level). The extra payoff is stochastically determined, being either 800 or 0. The probability of obtaining 800 is contingent on the Worker's unobservable effort: in the positive intervention, it's 0.2 + 0.05e; in the negative intervention, it's 0.8 - 0.05e.
Worker's Payoff: Equal to 50 + W2 if they stay, and 100 + 100X if they quit.
A Worker stays if the minimum acceptable wage for a given X is less than or equal to the wage offered by the Employer.

Detailed Parameters:
1. The Worker's payoff from quitting includes the full return from training (100X), indicating the general nature of the training.
2. The discrepancy in the Employer's payoff between the Worker staying or quitting (50 + 100X) mirrors the Worker's stay-vs-quit payoff, implying no additional rent for bargaining between the two parties.

Comments about payoffs: 1, Worker's payoff from quitting includes the whole return of training, 100 X, reflecting that training is in fact general. 2, The gap of Employer payoff when Worker stays versus quits is 50 + 100X, which is exact Worker payoff when Worker stays versus quits. This implies that there is no additional rent for the two parties to bargain with.

Unobservable Effort: Workers report the amount to contribute to influence Employers' payoffs. If they actually did lower than this amount, Employers' payoffs are not affected, which depends on the reported/contributed amount, not the actual amount. However, Worker can earn an additional 20 points per table in the real effort tasks, incentivizing maximal effort rather than arbitrary reporting.

Randomization in our experiment is conducted in the lab using a computer.
Subjects are randomly assigned to treatments by computer.
Within a treatment, the role of each subject (either Employer or Worker) is randomly determined by the computer.

individuals

No

N/A

240 observations, corresponding to the 240 university students participating in the study

Planned Number of Participants: 240 university students, with 60 participants allocated to each of the four treatments

N/A