Abstract From:
ESA 2018 world meeting in Berlin is now available at
https://www.economicscience.org/downloads/2018_ESA_Berlin.pdf
Are sunspots effective in a big crowd? : Evidence from a large-scale bank run experiment
Anita Kopanyi-Peuker University of Amsterdam
Author(s): Jasmina Arifovic (Simon Fraser University), Cars Hommes (University of Amsterdam), Anita KopanyiPeuker (University of Amsterdam),Isabelle Salle (Utrecht University)
ABSTRACT:
Coordination problems can have severe consequences for an economy. Whether currency attacks or bank runs eventually happen not only depend on the fundamentals, but also on agents' beliefs about others' actions. However, these beliefs might be affected by a sunspot variable (a commonly observed random variable, which is completely unrelated to the fundamentals), providing coordination device for agents. With this project we investigate whether sunspots effectively help people to coordinate in a bank run problem with a large number of participants, and how the payoff structure influences behavior. Our setup is similar to Arifovic and Jiang (2015) where subjects need to decide whether they wait or withdraw their money from the bank. Waiting gives higher payoff if enough people are doing so. Theoretically always waiting, always withdrawing, or coordinating based on the sunspot announcement are all equilibria of this repeated game. In this project we consider two group sizes: groups of 10 and 75-90. Furthermore, we vary the payoff of withdrawing, i.e. the relative attractiveness of the two actions. Our results show that large groups are more likely to end up in the 'run' equilibrium if payoffs are perceived too risky. However, coordination on waiting also happens in groups of 70-80 when waiting is sufficiently attractive. In that case all of our large and small groups coordinate on this action. In small groups we observe three types of behavior for the more risky environment: some groups coordinated on waiting, some on withdrawing, and some on the sunspot announcements.
ESA 2018 world meeting in Berlin is now available at
https://www.economicscience.org/downloads/2018_ESA_Berlin.pdf
Are sunspots effective in a big crowd? : Evidence from a large-scale bank run experiment
Anita Kopanyi-Peuker University of Amsterdam
Author(s): Jasmina Arifovic (Simon Fraser University), Cars Hommes (University of Amsterdam), Anita KopanyiPeuker (University of Amsterdam),Isabelle Salle (Utrecht University)
ABSTRACT:
Coordination problems can have severe consequences for an economy. Whether currency attacks or bank runs eventually happen not only depend on the fundamentals, but also on agents' beliefs about others' actions. However, these beliefs might be affected by a sunspot variable (a commonly observed random variable, which is completely unrelated to the fundamentals), providing coordination device for agents. With this project we investigate whether sunspots effectively help people to coordinate in a bank run problem with a large number of participants, and how the payoff structure influences behavior. Our setup is similar to Arifovic and Jiang (2015) where subjects need to decide whether they wait or withdraw their money from the bank. Waiting gives higher payoff if enough people are doing so. Theoretically always waiting, always withdrawing, or coordinating based on the sunspot announcement are all equilibria of this repeated game. In this project we consider two group sizes: groups of 10 and 75-90. Furthermore, we vary the payoff of withdrawing, i.e. the relative attractiveness of the two actions. Our results show that large groups are more likely to end up in the 'run' equilibrium if payoffs are perceived too risky. However, coordination on waiting also happens in groups of 70-80 when waiting is sufficiently attractive. In that case all of our large and small groups coordinate on this action. In small groups we observe three types of behavior for the more risky environment: some groups coordinated on waiting, some on withdrawing, and some on the sunspot announcements.