Here.
Two main classes of channels are studied as informational sources of financial contagion. One is a fundamental channel that is based on real and financial links between economies and the second is a social learning channel that arises when agents base their decisions on noisy observations about the actions of others in foreign markets. Using global games, I present a two country model of financial contagion in which both channels can be operative and I test its predictions experimentally. While the theory makes clear predictions about when each of these channels should lead to contagion, the experimental results show that subjects do not extract information optimally, which leads to two systematic biases that affect these channels directly. A base rate neglect bias makes subjects underweight their prior and thus weakens the fundamental channel. An overreaction bias strengthens the social learning channel, since subjects put too much weight on the information about the behavior of others, even when this information is irrelevant. These results have significant effects in terms of welfare that can be explained by the specific way in which these biases alter behavior.
From:
Isabel Trevino
University of California San Diego
Two main classes of channels are studied as informational sources of financial contagion. One is a fundamental channel that is based on real and financial links between economies and the second is a social learning channel that arises when agents base their decisions on noisy observations about the actions of others in foreign markets. Using global games, I present a two country model of financial contagion in which both channels can be operative and I test its predictions experimentally. While the theory makes clear predictions about when each of these channels should lead to contagion, the experimental results show that subjects do not extract information optimally, which leads to two systematic biases that affect these channels directly. A base rate neglect bias makes subjects underweight their prior and thus weakens the fundamental channel. An overreaction bias strengthens the social learning channel, since subjects put too much weight on the information about the behavior of others, even when this information is irrelevant. These results have significant effects in terms of welfare that can be explained by the specific way in which these biases alter behavior.
From:
Isabel Trevino
University of California San Diego