Theses Doctoral

Essays on Information Economics

Tangirala, Gowtham Kumar

In this doctoral dissertation, I broadly study the impact of information on economies from both a theoretical and an empirical perspective. Specifically, I study how strategic agents in a heterogeneous interacting network make decisions under incomplete information and how their actions are affected by the parameters that define the incompleteness of the information, with an emphasis on the social value of information. I then estimate the impact of information disclosure on the stock market by studying the specific example of the annual CCAR and DFAST bank stress tests conducted by the Federal Reserve. This dissertation consists of two chapters.

In the first chapter, I study a game of heterogeneous strategic interactions under incomplete information. I characterize the equilibrium actions and compare them to the benchmark constrained-efficient allocation. I parameterize the available information in terms of pairwise information commonality and accuracy and study how changing the said commonality and accuracy affects the social welfare. I also study how the structure of interactions between players affects the social value of information. I find that the extent of the inefficiency of the economy dictates the social value of information. I provide a complete characterization of the comparative statics of the social welfare with respect to commonality and accuracy for completely efficient economies. I find that when interactions are heterogenous, it is possible for social welfare to be non-monotonic with respect to information commonality, a behavior unseen in economies with homogeneous interactions. For inefficient economies, I provide sufficient conditions under which the social welfare exhibits monotonic behavior.

In the second chapter, I study the predictability of the results of the annual Comprehensive Capital Analysis and Review (CCAR) and Dodd-Frank Act Stress Test (DFAST) conducted by the Federal Reserve. I find that these results are highly predictable on year-to-year basis. I also find a high degree of predictability within the adverse scenario and the severely adverse scenario results within a given year. I find that that these predictable trends hold over time, from 2012 to 2020. I also try to ascertain the impact of the announcement of these results on the stock market and find no statistically significant effect. Lastly, I study the fixed effect impact of the disclosure events on the stock and options market. I find that while there are individual instances of significant impact, there is no significant impact across the years. I discuss potential implications of these patterns for the further development and application of stress testing.


  • thumnail for Tangirala_columbia_0054D_16545.pdf Tangirala_columbia_0054D_16545.pdf application/pdf 703 KB Download File

More About This Work

Academic Units
Thesis Advisors
Tahbaz-Salehi, Alireza
Glasserman, Paul
Ph.D., Columbia University
Published Here
June 1, 2021