Demian

Published and Accepted Papers (for links see my CV { pdf }):

"On nonlinear ill-posed inverse problems with applications to pricing of defaultable bonds and option pricing," with X. Chen. Science in China Series A: Mathematics . Volume 52, No. 6, June 2009, pp 1157-1168.

"Efficient Estimation of Semiparametric Conditional Moment Models with Possibly Nonsmooth Residuals," with X. Chen. Journal of Econometrics . Volume 152, Issue 1, September 2009, Pages 46-60.

"Estimation and model selection of semiparametric copula-based multivariate survival functions under general censorship," with X. Chen, Y. Fan and Z. Ying. Journal of Econometrics . Volume 157, Issue 1, July 2010, Pages 129-142.

"Estimation of Nonparametric Conditional Moment Models with Possibly Nonsmooth Moments," with X. Chen. Econometrica . Volume 80, No. 1, January 2012, Pages 277-322.

"Sieve Quasi Likelihood Ratio Inference on Semi/nonparametric Conditional Moment Models," with X. Chen. Econometrica . Volume 83, No. 3, May 2015, Pages 1013-1079.

"Bootstrap Consistency for Quadratic Forms of Sample Averages with Increasing Dimension." Electronic Journal of Statistics . Volume 9, No. 2, 2015, Pages 3046-3097.

"Berk-Nash Equilibrium: A Framework for Modeling Agents with Misspecified Models,'' with I. Esponda. Econometrica . Volume 84, No. 3, May 2016, Pages 1093-1130.

"Sovereign Default Risk and Uncertainty Premia,'' with I. Presno. Accepted at American Economic Journal: Macroeconomics .

"Conditional Retrospective Voting in Large Elections,'' with I. Esponda. Accepted at American Economic Journal: Microeconomics .



Working Papers:

“On the Non-Asymptotic Properties of Regularized M-estimators”,
Updated: May 3rd, 2016 (submitted). { pdf }

Abstract: We propose a general framework for regularization in M-estimation problems under time dependent (absolutely regular-mixing) data which encompasses many of the existing estimators. We derive non-asymptotic concentration bounds for the regularized M-estimator. Our results exhibit a "variance-bias" trade-off, with the "variance" term being governed by a novel measure of the "size" of the parameter set. We also show that the mixing structure affect the variance term by scaling the number of observations; depending on the decay rate of the mixing coefficients, this scaling can even affect the asymptotic behavior. Finally, we propose a data-driven method for choosing the tuning parameters of the regularized estimator which yield the same (up to constants) concentration bound as one that optimally balances the "(squared) bias" and "variance" terms. We illustrate the results with several canonical examples of, both, non-parametric and high-dimensional models.

“Equilibrium in Misspecified Markov Decision Process”,
with Ignacio Esponda (Revision requested at Econometrica). Updated: May 14th, 2016. { pdf }

Abstract: We study Markov decision problems where the agent does not know the transition probability function mapping current states and actions to future states. The agent has a prior belief over a set of possible transition functions and updates beliefs using Bayes' rule. We allow her to be misspecified in the sense that the true transition probability function is not in the support of her prior. This problem is relevant in many economic settings but is usually not amenable to analysis by the researcher. We make the problem tractable by studying asymptotic behavior. We propose an equilibrium notion and provide conditions under which it characterizes steady state behavior. In the special case where the problem is static, equilibrium coincides with the single-agent version of Berk-Nash equilibrium (Esponda and Pouzo (2016)). We also discuss subtle issues that arise exclusively in dynamic settings due to the possibility of a negative value of experimentation.

“Retrospective Voting and Party Polarization”,
with Ignacio Esponda (submitted). Updated: June 14th, 2016. { pdf }

Abstract: We provide a new and favorable perspective on voter naivete and party polarization. We study a model where two parties compete by committing to policies and voters subsequently vote for their preferred party. We contrast sophisticated with naive voting. The former is embodied by Nash equilibrium while the latter is formalized using the notion of a retrospective voting equilibrium (Esponda and Pouzo, forthcoming). Retrospective voters do not understand the mapping between states and outcomes induced by a policy; instead, they simply vote for the party that has delivered the best performance in the past. We show that parties have an incentive to polarize under retrospective, compared to Nash, voting. Moreover, this polarization often results in higher welfare due to a better match between policies and fundamentals.

“Optimal Taxation with Endogenous Default under Incomplete Markets”,
with Ignacio Presno (Revision requested at AEJ: Macro). Updated: May 7th, 2016. { pdf }

Abstract: In a dynamic economy, we characterize the fiscal policy of the government when it levies distortionary taxes and issues defaultable bonds to finance its stochastic expenditure. Default may occur in equilibrium as it prevents the government from incurring in future tax distortions that would come along with the service of the debt. Households anticipate the possibility of default generating endogenous credit limits. These limits hinder the government's ability to smooth taxes using debt, implying more volatile and less serially correlated fiscal policies, higher borrowing costs and lower levels of indebtedness. In order to exit temporary financial autarky following a default event, the government has to repay a random fraction of the defaulted debt. We show that the optimal fiscal and renegotiation policies have implications aligned with the data.

“Learning foundation and equilibrium selection in voting environments with private information”,
with Ignacio Esponda. Updated: January 25, 2012. { pdf }

Abstract: We use a dynamic learning model to investigate different behavioral assumptions in voting environments with private information. We show that a simple rule, where players learn based on the outcomes of past elections in which they were pivotal but requires no prior knowledge of the payoff structure or of the rules followed by other players, provides a foundation for Nash equilibrium. In contrast, a rule where voters learn from all past elections provides a foundation for a new notion of naive voting where players vote sincerely but have endogenously-determined beliefs. Finally, we use the model to select among multiple equilibria in the jury model. We find that the well-known result that elections aggregate information under Nash equilibrium relies on the selection of symmetric equilibria which are unstable. Nevertheless, we show that there exist (possibly asymmetric) Nash equilibria that are asymptotically stable and aggregate information.