Essays on information structure and economic activities

• Main Author: Jing, Lin
• Other Authors: Yan Jubo
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Social sciences::Economic theory; Social sciences::Communication
• Online Access: https://hdl.handle.net/10356/159565
• Institution: Nanyang Technological University

This thesis consists of three self-contained essays on experimental studies on information structure and economic activities. The first essay examines the (relative) effectiveness of certainty and severity of punishment in deterring crime and corruption under asymmetric information. The literature mainly examines the deterrence of either crime or corruption separately, and the results are mixed on the relative effectiveness between certainty and severity of punishment. Utilizing a principal-monitor-agent framework with inbuilt asymmetric information developed by Ortner & Chassang (2018), we are able to investigate the deterrence of crime and corruption jointly and simultaneously. The experimental design features two different policy regimes: HP with high certainty (probability) of punishment (p) and low severity of punishment (E(W)) while LP with low p and high E(W). Within each regime, we examine whether there is a real deterrent effect by increasing p or E(W), and which one delivers a greater effect if there is any. One major contribution of this essay suggests that the (relative) effectiveness between certainty and severity of punishment is regime dependent. Specifically, in regime LP, neither increasing p nor increasing E(W) deters crime or corruption effectively. In contrast, in regime HP where the initial p is high enough, increasing p significantly deters crime and corruption while increasing E(W) only deters corruption significantly. Furthermore, an increase in p delivers a greater deterrent effect than that in E(W). In addition, we document the presence of the Cobra Effect in regime LP when we intend to deter corruption by increasing the expected wage of the monitor. Last but not least, we explore the changes in extensive and intensive margins of crime and corruption, and we find a difference in celerity of deterrent effect between increasing p and E(W) in regime HP, specifically, subjects are promptly responsive to changes in p while they are inertial to changes in E(W). The second essay examines theoretically and experimentally how information structures of the threshold effort level required to succeed in a joint group task affect cooperation and coordination in the group. We investigate three information structures, complete and symmetric (CS), incomplete and symmetric (IS), and incomplete and asymmetric (IA), as well as a signaling mechanism introduced under IA (IAS) to see what a difference it would make. The theoretical model predicts that, in equilibrium, the performance under IA is the lowest which is intuitive, and a signaling mechanism makes no difference. Furthermore, IS should deliver higher performance than CS in equilibrium. Our experimental results confirm the prediction that the performance under IA is the lowest among the three information structures, however, CS achieves significantly higher performance than both IS and IA, and IS just performs as poorly as IA. Furthermore, the signaling mechanism in IAS significantly promotes contribution and public goods provision compared with IA and IS. Between IAS and CS, the contribution and public goods provision are higher in IAS, but only the difference in contribution is marginally significant. Our results suggest that, in a cooperation and coordination problem, a simple signaling mechanism can completely overcome the inefficiencies caused by information asymmetry, and there is no need to try to eliminate information asymmetry at an absurdly increasing marginal cost. The third essay investigates the effectiveness of both leading-by-signaling and leader selection procedures on cooperation and coordination in group project implementation. Leadership is documented as a process characterized by influence, and communication is identified as one of the main channels for the leader to generate influence and achieve effective leadership. In addition, leadership has been considered as one important instrument that affects both the qualitative and quantitative levels of coordination and cooperation in group projects. Since the literature mainly investigates the effectiveness of leading-by-example, we in this paper experimentally investigate how leaders affect cooperation and coordination by sharing their privately observed information instead of by leading-by-example. In addition, we also compare the effectiveness of leadership among four leader selection rules: random selection, selection by ability, selection by historical contribution, and election. The experimental results show that the presence of the leader does not have a significant effect when she is randomly selected. However, when the leader is elected by her group members, the contributions, as well as the public goods provision ratio, are significantly higher, and the welfare also increases significantly compared with the control. In addition, we also show that the groups with elected leaders perform significantly better than those with leaders selected randomly or by historical contribution. Nonetheless, there is not a significant difference between groups with elected leaders and groups with leaders selected by ability. With separate examination on the differences in contribution behaviors of leaders and followers across treatments, we find that, elected leaders tend to be less manipulative against their followers, and the followers tend to be more compliant to coordinate at the message sent by the leader. This suggests that elected leaders receive the most legitimacy on average than leaders selected by other procedures.