Denis Kojevnikov

I am a PhD job market candidate at the Vancouver School of Economics, University of British Columbia. My field of interest is Econometric Theory. Currently I work on problems related to statistical inference with network dependent observations.

My recent paper focuses on the bootstrap for random variables lying on a given network whose stochastic dependence is governed by that network. It proposes a number of resampling methods that are useful for conducting inference on parameters of the underlying data generating process.

I will be available for interviews at the 2018 CEEE meetings in Toronto and the 2019 ASSA/AEA meetings in Atlanta.



Abstract: This paper focuses on the bootstrap for network dependent processes under the conditional -weak dependence. Such processes are distinct from other forms of random fields studied in the statistics and econometrics literature so that the existing bootstrap methods cannot be applied directly. We propose a block-based approach and a modification of the dependent wild bootstrap for constructing confidence sets for the mean of a network dependent process. In addition, we establish the consistency of these methods for the smooth function model and provide the bootstrap alternatives to the network heteroskedasticity-autocorrelation consistent (HAC) variance estimator. We find that the distribution of the modified dependent wild bootstrap and the corresponding variance estimator are consistent under weaker conditions than those of the block-based method, which makes the former approach preferable for practical implementation.

[go to paper]



Abstract: This paper considers a general form of network dependence where dependence between two sets of random variables becomes weaker as their distance in a network gets longer. We show that such network dependence cannot be embedded as a random field on a lattice in a Euclidean space with a fixed dimension when the maximum clique increases in size as the network grows. This paper applies Doukhan and Louhichi (1999) 's weak dependence notion to network dependence by measuring dependence strength by the covariance between nonlinearly transformed random variables. While this approach covers examples such as strong mixing random fields on a graph and conditional dependency graphs, it is most useful when dependence arises through a large functional-causal system of equations. The main results of our paper include the law of large numbers, and the central limit theorem. We also propose a heteroskedasticity-autocorrelation consistent variance estimator and prove its consistency under regularity conditions. The finite sample performance of this latter estimator is investigated through a Monte Carlo simulation study.

[go to paper]

Abstract: Econometric models of strategic interactions among people or firms have received a great deal of attention in the literature. Less attention has been paid to the role in inference of the underlying assumptions about the way the agents form belief about other agents. This paper focuses on a single large Bayesian game, and develops bootstrap inference methods which does not require a common prior assumption, and allows for the players to form beliefs differently from other players. By drawing on the main intuition of Kalai (2004), this paper introduces the notion of a hindsight regret which measures each player’s ex post value of other players’ type information, and obtains its belief-free bound. From this bound, this paper derives testable implications and develops a bootstrap inference procedure for the structural parameters. This paper demonstrates this result through Monte Carlo simulations.

Abstract: This note provides a conditional Berry-Esseen type bound for maxima of the normalized sum of a vector-valued martingale difference sequence adapted to filtration . In particular, we establish conditions under which the conditional distribution of the maximum of given some -field is approximated by the maximum of a mean-zero normal random vector having the same conditional variance given as the vector . In order to overcome the non-differentiability of the maximum function we approximate the latter by a version of the log sum of exponentials. In addition, this work relies on the anti-concentration bound for maxima of Gaussian random vectors established in Chernozhukov, Chetverikov, and Kato (2015).

[go to paper]


Teaching Assistant for Graduate Courses at UBC:

  • ECON 627 - Econometric Theory II (2017, 2018)
  • ECON 626 - Econometric Theory I (2017)
  • ECON 628 - Topics in Applied Econometrics (2015, 2016)
  • ECON 527 - Econometric Methods of Economic Research (2016)


Teaching Assistant for Undergraduate Courses at UBC:

  • ECON 425 - Introduction to Econometric (2018)
  • ECON 326 - Introduction to Econometric (2016)
  • ECON 325 - Introduction to Empirical Economics (2014)