Climate change impact studies inform policymakers on the estimated damages of future climate change on economic, health and other outcomes. In most studies, an annual outcome variable is observed, e.g. agricultural yield, annual mortality or gross domestic product, along with a higher-frequency regressor, e.g. daily temperature. While applied researchers tend to consider multiple models to characterize the relationship between the outcome and the high-frequency regressor, a choice between the damage functions implied by the different models has to be made to inform policy. This paper formalizes the model selection problem and the policy objective in this empirical setting in light of current empirical practice. We then show that existing model selection criteria are only suitable for the policy objective under specific conditions. These conditions include a requirement that one of the models under consideration nests the true model. To overcome this restriction, we propose a new criterion, the proximity-weighted mean-squared error (PWMSE) of predicting climate change impacts. The PWMSE targets the policy objective of predicting the impact of projected climate change directly by giving higher weight to prior years with weather closer to the projected scenario. We show that our approach selects the best approximate regression model that has the smallest weighted error of predicted impacts for a future climate scenario. A simulation study and an application revisiting the impact of climate change on agricultural production illustrate the empirical relevance of our theoretical analysis.
Time Consistency and Duration of Government Debt: A Model of Quantitative Easing∗
All finite single-agent choice problem with ordinal preferences admit a compatible utility function such that: strict dominance by pure or mixed actions coincides with dominance by pure actions in the sense of Börgers (1993). With asymmetric preferences, Börgers’ notion of dominance reduces to the classical notion of strict dominance by pure strategies. The result extends to some infinite environments satisfying different assumptions. In all cases, the equivalence holds whenever the agent is sufficiently risk averse.
Other publications
What Price is Right? Cigarette Demand Has Become More Responsive to Prices
This paper proposes both point-wise and uniform confidence intervals (CIs) for an element theta_1 of a parameter vector theta in R^d which is partially identified by affine moment equality and inequality conditions. The CIs are based on an estimator of a regularized support function of the identified set and have closed–form. I provide examples in which my CIs are shorter than those in the existing literature. Unlike much of the existing literature, the proposed CIs can be computed as a solution to a convex optimization problem, which leads to a substantial decrease in computation time. My approach can be used, for instance, to compute a CI for the return to schooling using income bracket data without strong distributional assumptions.
Bias correction for quantile regression estimators
This paper studies small sample properties and bias of just-identified instrumental variable quantile regression (IVQR) estimators, nesting order statistics and classical quantile regression. We propose a theoretical framework for analyzing small sample properties based on a novel approximation of the discontinuous sample moments with a Hölder continuous process. Using this approximation, we derive remainder bounds for the asymptotic linear expansions of exact and k-step estimators of IVQR models. Furthermore, we derive a bias formula for exact IVQR estimators up to order o(1/n). The bias contains components that cannot be consistently estimated and depend on the particular numerical estimation algorithm. To circumvent this problem, we propose a novel 1-step adjustment of the estimator, which admits a feasible bias correction. Monte Carlo evidence suggests that our formula removes a substantial portion of the bias for sample sizes as small as n=50. We suggest using exact estimators, when possible, to achieve the smallest bias. Otherwise, applying 1-step corrections may improve the higher-order bias and MSE of any consistent estimator.
Secular rise and pro-cyclical variation in markups: Evidence from US grocery stores
We study the properties of projection inference for set-identified Structural Vector Autoregressions. A nominal 1-alpha projection region collects the structural parameters that are compatible with a 1-alpha Wald ellipsoid for the model's reduced-form parameters (autoregressive coefficients and the covariance matrix of residuals). We show that projection inference can be applied to a general class of stationary models, is computationally feasible, and it produces regions for the structural parameters and their identified set with both frequentist coverage and robust Bayesian credibility of at least 1- alpha.
A drawback of the projection approach is that both coverage and robust credibility may be strictly above their nominal level. Following the work of Kaido, Molinari, and Stoye (2016), we calibrate the radius of the Wald ellipsoid to guarantee that the robust Bayesian credibility of the projection method is exactly 1 - alpha. If the bounds of the identified set are differentiable, our calibrated projection also covers the identified set with probability 1 - alpha. We illustrate the main results of the paper using the demand/supply-model for the U.S. labor market in Baumeister and Hamilton (2015).
2015 11th World Congress Econometric Society, Montreal, Canada
2015 22nd International Symposium on Mathematical Programming, Pittsburgh, USA
2015 PSU-Cornell Macro Workshop, Pennsylvania State University, State College, USA
2015 Annual Conference of the Royal Economic Society, University of Manchester, UK
2015 Higher School of Economics, Moscow, Russia
2014 Latin American Meeting of The Econometric Society, University of São Paulo, Brazil
Identification in dynamic models using sign restrictions
Sign restrictions on impulse response functions are used in the literature to identify structural vector autoregressions and structural factor models. I extend the rank condition used for exclusion restrictions and provide a necessary and sufficient conditions for point identification for sign restrictions in this class of models. The necessary condition for point identification implies that as the number of sign restrictions grows a subset with sufficient number of sign restrictions becomes binding in the limit. However, one does not need to possess information about this subset to achieve point identification. So when exclusion restrictions are not justified by theory, sign restrictions can provide an alternative way to get point-identified impulse response functions. Also further, I present a closed form representation of the set of all impulse response functions satisfying a set of sign restrictions. I demonstrate that restrictions on responses to all shocks can dramatically shrink this set when compared to restrictions only on a small number of shocks.
2013 23rd Annual Meeting of the Midwest Econometrics Group, Indiana University, Bloomington
