Karthik Srinivasan
Welcome!
I am an Economics PhD candidate at the University of Chicago's Booth School of Business. I am on the job market during the 2023-24 academic year. I am an applied microeconomist working on topics in behavioral economics, labor economics, and political economy.
In my research, I study how non-monetary incentives and behavioral biases influence key civic institutions including the press, the judiciary, and social media platforms. Methodologically, my work combines causal inference designs with machine learning and natural language processing models. My job market paper, Paying Attention, documents that attention can be used to incentivize effort, and traces out some implications of this fact for the design and regulation of social media platforms.
Feel free to reach out to me at ks@chicagobooth.edu or take a look at my CV.
Thanks for visiting my website!
Working Papers
Can making decisions in extreme cases bias subsequent decisions? I study this question in a high-stakes field setting: felony sentencing. I estimate the effect of sentencing a first-degree murder on the length of sentences issued to subsequent defendants. I use data on the universe of felony sentencing decisions in Cook County to estimate a difference-in-differences design comparing judges in the same courthouse who have and have not recently sentenced a first-degree murder. Judges issue sentences that are 13% longer in the 10 days after they sentence a first-degree murder. Effects are twice as large for defendants who share the same race as the murderer and defendants who face high-class felony charges. A back- of-the-envelope calculation suggests that this bias affects 6% of defendants on an ongoing basis, because judges regularly sentence first-degree murders.
I study the scope of a principal-agent problem in the field. I analyze news firms and journalists with possibly misaligned preferences over the partisan slant of content, and find that the firm's ability to exert control is limited. I construct a dataset that links 2,700 journalists to firms, news articles, and Twitter profiles. I measure article slant with a machine learning algorithm I train to identify partisan phrases. Using a movers design, I find firm ideology does not change the slant of a journalist’s writing. In contrast, journalist ideology, estimated using the following decisions of Twitter users, is strongly correlated with article slant.
Work In Progress
Paying Attention (Job Market Paper) [Draft Coming Soon!]
Status: Experiment in progress.
Status: Pilot completed.
Teaching
I taught an intensive math camp to incoming Booth PhD students. I collaborated with my co-instructor Walter Zhang to design all course materials including lecture notes and problem sets. The course is comprised of 36 hours of lectures reviewing calculus, linear algebra, real analysis, probability, statistical inference, optimization and dynamic programming.