nep-ecm New Economics Papers
on Econometrics
Issue of 2021‒05‒10
fourteen papers chosen by
Sune Karlsson
Örebro universitet

  1. A nonparametric instrumental approach to endogeneity in competing risks models By Jad Beyhum; Jean-Pierre Florens; Ingrid Van Keilegom
  2. Identification and Estimation of average marginal effects in fixed effect logit models By Laurent Davezies; Xavier D'Haultfoeuille; Louise Laage
  3. Semiparametric Forecasting Problem in High Dimensional Dynamic Panel with Correlated Random Effects: A Hierarchical Empirical Bayes Approach By Pacifico, Antonio
  4. Marginal Treatment Effects with Misclassified Treatment By Santiago Acerenza; Kyunghoon Ban; D\'esir\'e K\'edagni
  5. Automatic Debiased Machine Learning via Neural Nets for Generalized Linear Regression By Victor Chernozhukov; Whitney K. Newey; Victor Quintas-Martinez; Vasilis Syrgkanis
  6. Distribution and statistics of the Sharpe Ratio By Eric Benhamou
  7. Downside and Upside Uncertainty Shocks By Forni, Mario; Gambetti, Luca; Sala, Luca
  8. Frequency causality measures and Vector AutoRegressive (VAR) models: An improved subset selection method suited to parsimonious systems By Christophe Chorro; Emmanuelle Jay; Philippe De Peretti; Thibault Soler
  9. Conditional Rotation Between Forecasting Models By Timmermann, Allan; Zhu, Yinchu
  10. A Modified Randomization Test for the Level of Clustering By Yong Cai
  11. Estimation of High Dimensional Vector Autoregression via Sparse Precision Matrix By Template-Type: ReDIF-Paper 1.0; Benjamin Poignard; Manabu Asai
  12. Machine Collaboration By Qingfeng Liu; Yang Feng
  13. Spectral decomposition of the information about latent variables in dynamic macroeconomic models By Nikolay Iskrev
  14. Interactive R&D Spillovers: An estimation strategy based on forecasting-driven model selection By Georgios Gioldasis; Antonio Musolesi; Michel Simioni

  1. By: Jad Beyhum; Jean-Pierre Florens; Ingrid Van Keilegom
    Abstract: This paper discusses endogenous treatment models with duration outcomes, competing risks and random right censoring. The endogeneity issue is solved using a discrete instrumental variable. We show that the competing risks model generates a non-parametric quantile instrumental regression problem. The cause-specific cumulative incidence, the cause-specific hazard and the subdistribution hazard can be recovered from the regression function. A distinguishing feature of the model is that censoring and competing risks prevent identification at some quantiles. We characterize the set of quantiles for which exact identification is possible and give partial identification results for other quantiles. We outline an estimation procedure and discuss its properties. The finite sample performance of the estimator is evaluated through simulations. We apply the proposed method to the Health Insurance Plan of Greater New York experiment.
    Date: 2021–05
  2. By: Laurent Davezies; Xavier D'Haultfoeuille; Louise Laage
    Abstract: This article considers average marginal effects (AME) and similar parameters in a panel data fixed effects logit model. Relating the identified set of the AME to an extremal moment problem, we first show how to obtain sharp bounds on the AME straightforwardly, without any optimization. Then, we consider two strategies to build confidence intervals on the AME. In the first, we estimate the sharp bounds with a semiparametric two-step estimator involving a first-step nonparametric estimator. We derive the asymptotic distributions of these bounds under, mostly, a restriction on the cardinality of the support of the unobserved heterogeneity. The second, very simple strategy does not require any nonparametric estimation but may result in larger confidence intervals. Monte Carlo simulations suggest that both approaches work well in practice, the second being actually very competitive for usual sample sizes.
    Date: 2021–05
  3. By: Pacifico, Antonio
    Abstract: This paper aims to address semiparametric forecasting problem when studying high dimensional data in multivariate dynamic panel model with correlated random effects. A hierarchical empirical Bayesian perspective is developed to jointly deal with incidental parameters, structural framework, unobserved heterogeneity, and model misspecification problems. Methodologically, an ad-hoc model selection on a mixture of normal distributions is addressed to obtain the best combination of outcomes to construct empirical Bayes estimator and then investigate ratio-optimality and posterior consistency for better individual–specific forecasts. Simulations for Monte Carlo designs are performed to account for relative regrets dealing with correlated random effects distribution. A real case-study on the current COVID-19 pandemic crisis among a pool of developed and emerging economies is also conducted to highlight the performance of the estimating procedure.
    Keywords: Dynamic Panel Data; Ratio-Optimality; Bayesian Methods; Forecasting; MCMC Simulations; Tweedie Correction.
    JEL: C1 C5 O1
    Date: 2021
  4. By: Santiago Acerenza; Kyunghoon Ban; D\'esir\'e K\'edagni
    Abstract: This paper studies identification of the marginal treatment effect (MTE) when a binary treatment variable is misclassified. We show under standard assumptions that the MTE is identified as the derivative of the conditional expectation of the observed outcome given the true propensity score, which is partially identified. We characterize the identified set for this propensity score, and then for the MTE. We use our MTE bounds to derive bounds on other commonly used parameters in the literature. We show that our bounds are tighter than the existing bounds for the local average treatment effect. We illustrate the practical relevance of our derived bounds through some numerical and empirical results.
    Date: 2021–05
  5. By: Victor Chernozhukov; Whitney K. Newey; Victor Quintas-Martinez; Vasilis Syrgkanis
    Abstract: We give debiased machine learners of parameters of interest that depend on generalized linear regressions, which regressions make a residual orthogonal to regressors. The parameters of interest include many causal and policy effects. We give neural net learners of the bias correction that are automatic in only depending on the object of interest and the regression residual. Convergence rates are given for these neural nets and for more general learners of the bias correction. We also give conditions for asymptotic normality and consistent asymptotic variance estimation of the learner of the object of interest.
    Date: 2021–04
  6. By: Eric Benhamou (MILES - Machine Intelligence and Learning Systems - LAMSADE - Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision - Université Paris Dauphine-PSL - PSL - Université Paris sciences et lettres - CNRS - Centre National de la Recherche Scientifique, LAMSADE - Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision - Université Paris Dauphine-PSL - PSL - Université Paris sciences et lettres - CNRS - Centre National de la Recherche Scientifique)
    Abstract: Because of the frequent usage of the Sharpe ratio in asset management to compare and benchmark funds and asset managers, it is relevant to derive the distribution and some statistics of the Sharpe ratio. In this paper, we show that under the assumption of independent normally distributed returns, it is possible to derive the exact distribution of the Sharpe ratio. In particular, we prove that up to a rescaling factor, the Sharpe ratio is a non centered Student distribution whose characteristics have been widely studied by statisticians. For a large number of observations, we can derive the asymtptotic distribution and find back the result of Lo (2002). We also illustrate the fact that the empirical Sharpe ratio is asymptotically optimal in the sense that it achieves the Cramer Rao bound. We then study the empirical SR under AR(1) assumptions and investigate the effect of compounding period on the Sharpe (computing the annual Sharpe with monthly data for instance). We finally provide general formula in this case of heteroscedasticity and autocorrelation.
    Keywords: JEL classification: C12,G11 Sharpe ratio,Student distribution,compounding effect on Sharpe,AR(1),Cramer Rao bound
    Date: 2021–04–23
  7. By: Forni, Mario; Gambetti, Luca; Sala, Luca
    Abstract: An increase in uncertainty is not contractionary per se. What generates a significant downturn of economic activity is a widening of the left tail of the expected distribution of growth, the downside uncertainty. On the contrary, an increase of the right tail, the upside uncertainty, is mildly expansionary. The reason for why uncertainty shocks have been previously found to be contractionary is because movements in downside uncertainty dominate existing empirical measures of uncertainty. The results are obtained using a new econometric approach which combines quantile regressions and structural VARs.
    Keywords: Quantile regression; Skewness; uncertainty; VAR models
    JEL: C32 E32
    Date: 2021–03
  8. By: Christophe Chorro (Centre d'Economie de la Sorbonne, Université Paris 1 Panthéon-Sorbonne); Emmanuelle Jay (Fidéas Capital Quanted & Europlace Institute of Finance); Philippe De Peretti (Centre d'Economie de la Sorbonne, Université Paris 1 Panthéon-Sorbonne); Thibault Soler (Fidéas Capital and Centre d'Economie de la Sorbonne)
    Abstract: Finding causal relationships in large dimensional systems is of key importance in a number of fields. Granger non-causality tests have become standard tools, but they only detect the direction of the causality, not its strength. To overcome this point, in the frequency domain, several measures have been introduced such as the Direct Transfer Function (DTF), the Partial Directed Coherence measure (PDC) or the Generalized Partial Directed Coherence measure (GPDC). Since these measures are based on a two-step estimation, consisting in i) estimating a Vector AutoRegressive (VAR) in the time domain and ii) using the VAR coefficients to compute measures in the frequency domain, they may suffer from cascading errors. Indeed, a flawed VAR estimation will translate into large biases in coherence measures. Our goal in this paper is twofold. First, using Monte Carlo simulations, we quantify these biases. We show that the two-step procedure results in highly inaccurate coherence measures, mostly due to the fact that non-significant coefficients are kept, especially in parsimonious systems. Based on this idea, we next propose a new methodology (mBTS-TD) based on VAR reduction procedures, combining the modified-Backward-in-Time selection method (mBTS) and the Top-Down strategy (TD). We show that our mBTS-TD method outperforms the classical two-step procedure. At last, we apply our new approach to recover the topology of a weighted financial network in order to identify through the local directed weighted clustering coefficient the most systemic assets and exclude them from the investment universe before allocating the portfolio to improve the return/risk ratio
    Keywords: VAR model; subset selection methods; frequency causality measures; weighted financial networks; portfolio allocation
    JEL: C5 G11
    Date: 2021–04
  9. By: Timmermann, Allan; Zhu, Yinchu
    Abstract: We establish conditions under which forecasting performance can be improved by rotating between a set of underlying forecasts whose predictive accuracy is tracked using a set of time-varying monitoring instruments. We characterize the properties that the monitoring instruments must possess to be useful for identifying, at each point in time, the best forecast and show that these reflect both the accuracy of the predictors used by the underlying forecasting models and the strength of the monitoring instruments. Finite-sample bounds on forecasting performance that account for estimation error are used to compute the expected loss of the competing forecasts as well as for the dynamic rotation strategy. Finally, using Monte Carlo simulations and empirical applications to forecasting inflation and stock returns, we demonstrate the potential gains from using conditioning information to rotate between forecasts
    Keywords: finite sample bounds; Forecasting Performance; real time monitoring
    JEL: C18 C32 C53
    Date: 2021–03
  10. By: Yong Cai
    Abstract: Suppose a researcher observes individuals within a county within a state. Given concerns about correlation across individuals, at which level should they cluster their observations for inference? This paper proposes a modified randomization test as a robustness check for their chosen specification in a linear regression setting. Existing tests require either the number of states or number of counties to be large. Our method is designed for settings with few states and few counties. While the method is conservative, it has competitive power in settings that may be relevant to empirical work.
    Date: 2021–05
  11. By: Template-Type: ReDIF-Paper 1.0; Benjamin Poignard (Graduate School of Economics, Osaka University); Manabu Asai (Faculty of Economics, Soka University)
    Abstract: We consider the problem of estimating sparse structural vector autoregression (SVAR) processes via penalized precision matrix. Such matrix is the output of the underlying directed acyclic graph of the SVAR process, whose zero components correspond to zero SVAR coecients. The precision matrix estimators are deduced from the class of Bregman divergences and regularized by the SCAD, MCP and LASSO penalties. Under suitable regularity conditions, we derive error bounds for the regularized precision matrix for each Bregman divergence. Moreover, we establish the support recovery property, including the case when the penalty is non-convex. These theoretical results are supported by empirical studies.
    Keywords: sparse structural vector autoregression; statistical consistency; support recovery.
  12. By: Qingfeng Liu; Yang Feng
    Abstract: We propose a new ensemble framework for supervised learning, named machine collaboration (MaC), based on a collection of base machines for prediction tasks. Different from bagging/stacking (a parallel & independent framework) and boosting (a sequential & top-down framework), MaC is a type of circular & interactive learning framework. The circular & interactive feature helps the base machines to transfer information circularly and update their own structures and parameters accordingly. The theoretical result on the risk bound of the estimator based on MaC shows that circular & interactive feature can help MaC reduce the risk via a parsimonious ensemble. We conduct extensive experiments on simulated data and 119 benchmark real data sets. The results of the experiments show that in most cases, MaC performs much better than several state-of-the-art methods, including CART, neural network, stacking, and boosting.
    Date: 2021–05
  13. By: Nikolay Iskrev
    Abstract: In this paper, I show how to perform spectral decomposition of the information about latent variables in dynamic economic models. A model describes the joint probability distribution of a set of observed and latent variables. The amount of information transferred from the former to the latter is measured by the reduction of uncertainty in the posterior compared to the prior distribution of any given latent variable. Casting the analysis in the frequency domain allows decomposing the total amount of information in terms of frequency-specific contributions as well as in terms of information contributed by individual observed variables. I illustrate the usefulness of the proposed methodology with applications to two DSGE models taken from the literature.
    JEL: C32 C51 C52 E32
    Date: 2021
  14. By: Georgios Gioldasis (Università degli Studi di Ferrara); Antonio Musolesi (Università degli Studi di Ferrara); Michel Simioni (MOISA, INRA, University of Montpellier, Montpellier, France)
    Abstract: This paper reconsiders the international technology diffusion model. Because the high degree of uncertainty surrounding the Data Generating Process and the likely presence of nonlinearities and latent common factors, it considers alternative nonparametric panel specifications which extend the Common Correlated Effects approach and then contrasts the out-of-sample performance of them with those of more common parametric models. To do so, we extend a recently proposed data-driven model choice approach, which takes its roots on cross validation and aims at testing whether two competing approximate models are equivalent in terms of their expected true error, to the case of cross-sectionally dependent panels, by exploiting moving block bootstrap resampling methods and assessing forecasting performances of competing models. Our results indicate that the adoption of a fully nonparametric specification provides better performances. This work also refines previous results by showing threshold effects, nonlinearities and interactions, which are obscured in parametric specifications and which have relevant implications for policy.
    Keywords: large panels; cross-sectional dependence; factor models; nonparametric regression; spline functions; approximate model; predictive accuracy, international technology diffusion
    JEL: C23 C5 F0 O3
    Date: 2021–06

This nep-ecm issue is ©2021 by Sune Karlsson. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.