nep-ecm New Economics Papers
on Econometrics
Issue of 2012‒01‒10
eight papers chosen by
Sune Karlsson
Orebro University

  1. Robust Henderson III estimators of variance components in the nested error model By Betsabé Pérez; Daniel Peña; Isabel Molina
  2. A Duration Model with Dynamic Unobserved Heterogeneity By Botosaru, Irene
  3. Testing for Granger causality in a system of more than two variables By Theologos Pantelidis
  4. Panel-CADF Testing with R: Panel Unit Root Tests Made Easy By Lupi, Claudio
  5. What we can learn from pricing 139,879 Individual Stock Options By Lars Stentoft
  6. Nonparametric tests of panel conditioning and attrition bias in panel surveys. By Das, J.W.M.; Soest, A.H.O. van; Toepoel, V.
  7. Dependent default and recovery: MCMC study of downturn LGD credit risk model By Pavel V. Shevchenko; Xiaolin Luo
  8. Risk and Return: Long-Run Relationships, Fractional Cointegration, and Return Predictability By Tim Bollerslev; Daniela Osterrieder; Natalia Sizova; George Tauchen

  1. By: Betsabé Pérez; Daniel Peña; Isabel Molina
    Abstract: Common methods for estimating variance components in Linear Mixed Models include Maximum Likelihood (ML) and Restricted Maximum Likelihood (REML). These methods are based on the strong assumption of multivariate normal distribution and it is well know that they are very sensitive to outlying observations with respect to any of the random components. Several robust altematives of these methods have been proposed (e.g. Fellner 1986, Richardson and Welsh 1995). In this work we present several robust alternatives based on the Henderson method III which do not rely on the normality assumption and provide explicit solutions for the variance components estimators. These estimators can later be used to derive robust estimators of regression coefficients. Finally, we describe an application of this procedure to small area estimation, in which the main target is the estimation of the means of areas or domains when the within-area sample sizes are small.
    Keywords: Henderson method III, Linear mixed model, Robust estimators, Variance component estimators
    Date: 2011–12
  2. By: Botosaru, Irene
    Abstract: The paper considers a new class of duration models in which unobserved heterogeneity changes with time. The class addresses two main questions: How does the exit probability from a state vary when unobserved heterogeneity evolves through time? And do changes in unobserved heterogeneity have a timing effect? We show the non- and semi-parametric identification of the new class by solving a nonlinear integral equation with unknown kernel. Both the function of observed covariates and the mean of the distribution of unobserved heterogeneity are nonparametrically identified. Identifying timing effects and the distribution of unobserved heterogeneity requires stronger assumptions on either one of the two. An extension to the case when unobserved heterogeneity is a function of observed covariates is also identified. We show that sieve maximum likelihood estimators are consistent and present Monte Carlo simulations for both correct specification and misspecification. The paper also presents an empirical model of unemployment duration in which individuals exit unemployment when total accumulated losses due to unemployment cross over a self-imposed spending limit.
    Keywords: duration analysis, Levy process, dynamic unobserved heterogeneity, identification, mixture
    JEL: C14 C41
    Date: 2011–12–16
  3. By: Theologos Pantelidis (Department of Economics, University of Macedonia)
    Abstract: This paper provides useful guidelines to practitioners who investigate Granger causality within a system of more than two variables by means of the two-step procedure proposed by Cheung and Ng (Journal of Econometrics, 1996) and modified by Hong (Journal of Econometrics, 2001). First, a theoretical example highlights cases that can mislead the researcher into reporting false causal relations between the variables under scrutiny. Then, the size of the problem is revealed by means of Monte Carlo simulations. Finally, an empirical application that investigates causality-in-mean among six major European stock markets, illustrates the proper procedure to follow for correct inference.
    Keywords: Causality-in-mean; causality-in-variance; misspecification; simulation.
    JEL: C14 C22
    Date: 2012–01
  4. By: Lupi, Claudio
    Abstract: This paper presents the R implementation of the panel covariate augmented Dickey-Fuller (panel-CADF) test proposed in Costantini and Lupi (2011), as well as the implementation of the tests advocated in Choi (2001) and Demetrescu, Hassler, and Tarcolea (2006). A panel-CADF extension of the test suggested in Hanck (2008) is also discussed and its size and power properties are investigated via Monte Carlo analysis. The simulation results show that the panel-CADF tests have interesting properties in terms of size and power. The R implementation illustrated here is part of the ongoing work on a new R package named punitroots (Kleiber and Lupi 2011).
    Keywords: panel data, unit root, R
    JEL: C33 C12
    Date: 2011–12–29
  5. By: Lars Stentoft (HEC Montréal, CIRANO, CIRPEÉ, and CREATES)
    Abstract: The GARCH framework has been used for option pricing with quite some success. While the initial work assumed conditional Gaussian innovations, recent contributions relax this assumption and allow for more flexible parametric specifications of the underlying distribution. However, until now the empirical applications have been limited to index options or options on only a few stocks and this using only few potential distributions and variance specififications. In this paper we test the GARCH framework on 30 stocks in the Dow Jones Industrial Average using two classical volatility specififications and 7 different underlying distributions. Our results provide clear support for using an asymmetric volatility specifification together with non-Gaussian distribution, particularly of the Normal Inverse Gaussian type, and statistical tests show that this model is most frequently among the set of best performing models.
    Keywords: American options, GARCH models, Model Confidence Set, Simulation.
    JEL: C22 C53 G13
    Date: 2011–12–21
  6. By: Das, J.W.M. (Tilburg University); Soest, A.H.O. van (Tilburg University); Toepoel, V. (Tilburg University)
    Date: 2011
  7. By: Pavel V. Shevchenko; Xiaolin Luo
    Abstract: There is empirical evidence that recovery rates tend to go down just when the number of defaults goes up in economic downturns. This has to be taken into account in estimation of the capital against credit risk required by Basel II to cover losses during the adverse economic downturns; the so-called "downturn LGD" requirement. This paper presents estimation of the LGD credit risk model with default and recovery dependent via the latent systematic risk factor using Bayesian inference approach and Markov chain Monte Carlo method. This approach allows joint estimation of all model parameters and latent systematic factor, and all relevant uncertainties. Results using Moody's annual default and recovery rates for corporate bonds for the period 1982-2010 show that the impact of parameter uncertainty on economic capital can be very significant and should be assessed by practitioners.
    Date: 2011–12
  8. By: Tim Bollerslev (Duke University and CREATES); Daniela Osterrieder (Aarhus University and CREATES); Natalia Sizova (Rice University); George Tauchen (Duke University and CREATES)
    Abstract: The dynamic dependencies in financial market volatility are generally well described by a long-memory fractionally integrated process. At the same time, the volatility risk premium, defined as the difference between the ex-post realized volatility and the market’s ex-ante expectation thereof, tends to be much less persistent and well described by a short-memory process. Using newly available intraday data for the S&P 500 and the VIX volatility index, coupled with frequency domain inference procedures that allow us to focus on specific parts of the spectra, we show that the existing empirical evidence based on daily and coarser sampled data carries over to the high-frequency setting. Guided by these empirical findings, we formulate and estimate a fractionally cointegrated VAR model for the two high-frequency volatility series and the corresponding high-frequency S&P 500 returns. Consistent with the implications from a stylized equilibrium model that directly links the realized and expected volatilities to returns, we show that the equilibrium variance risk premium estimated with the intraday data within the fractionally cointegrated system results in non-trivial return predictability over longer interdaily and monthly horizons. These results in turn suggest that much of the existing literature seeking to establish a risk-return tradeoff relationship between expected returns and expected volatilities may be misguided, and that the variance risk premium provides a much better proxy for the true economic uncertainty that is being rewarded by the market.
    Keywords: High-frequency data, realized volatility, options implied volatility, variance risk premium, fractional integration, long-memory, fractional cointegration, equilibrium asset pricing, return predictability.
    JEL: C22 C32 C51 C52 G12 G13 G14
    Date: 2011–12–21

This nep-ecm issue is ©2012 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.