I am currently working on the following projects:
Ballot Stuffing: a Structural Approach with Kadir Atalay, R&R in the Economic Journal
We examine ballot stuffing, a form of electoral fraud involving the illegal addition of extra ballots. While common in autocratic regimes, this practice is increasingly cited in democracies as grounds for challenging election results. We propose a non-parametric structural model that links unobservable polling station characteristics to instances of ballot stuffing. By leveraging less fraudulent data--such as data from independent observers or exit polls--our model can identify ballot stuffing at the individual polling station. Applying this model to the 2011 Russian Parliamentary election, we find substantial evidence of ballot stuffing, with fraud rates varying across and within regions.
Causal Persuasion with Egor Starkov
We propose a model of causal persuasion, where a receiver is aware of a subset of real-world variables and their distribution, and tries to infer causal relationships. A better-informed sender knows the full causal graph and selectively reveals additional variables to persuade the receiver about some causal relationship. We show that persuasion is only possible when the receiver's model conflicts with the true causal structure. To reveal a true causal link, the sender often needs to disclose just one or two well-chosen variables. But to dispel a perceived link--to convince the receiver there is no causal relationship--every common cause must be disclosed. Our results highlight a fundamental asymmetry in persuasion: It is much easier to prove a connection than to disprove one.
Balancing Act: Quasi-Experimental Evidence on the Role of Non-Monetary Job Characteristics in Pension Reforms with Kadir Atalay and Ellen Stuart, under review
We study the role of non-monetary job characteristics in mediating the impact of pension reforms on labour supply. We examine a reform that increased the eligibility age for Age Pension, the first pillar of the Australian Pension System. We find that non-monetary elements of the work environment significantly influence retirement decisions. Individuals who self-report low engagement and stability in their job are more likely to extend their working years in response an increase in the Age Pension eligibility age, while those in more demanding occupations have weaker retirement responses. Our results contribute to the discussion about potential distributional effects of pension reform.
This paper studies how ambiguity affects identification of the state aggregation model -- a model with an agent that combines several states into an event in order to simplify decision-making. We show that the state aggregation is partially identified under Maxmin Expected Utility (MEU) and it is uniquely identified if α-MEU model is used. In addition, we offer testable restrictions of the model.