|
on Econometrics |
| By: | José M.R. Murteira (Faculdade de Economia, Universidade de Coimbra, and CEMAPRE); Joaquim J.S. Ramalho (Departamento de Economia and CEFAGE-UE, Universidade de Évora) |
| Abstract: | The present article discusses alternative regression models and estimation methods for dealing with multivariate fractional response variables. Both conditional mean models, estimable by quasi-maximum likelihood, and fully parametric models (Dirichlet and Dirichletmultinomial), estimable by maximum likelihood, are considered. A new parameterization is proposed for the parametric models, which accommodates the most common specifications for the conditional mean (e.g., multinomial logit, nested logit, random parameters logit, dogit). The text also discusses at some length the specification analysis of fractional regression models, proposing several tests that can be performed through artificial regressions. Finally, an extensive Monte Carlo study evaluates the finite sample properties of most of the estimators and tests considered. |
| Keywords: | Multivariate fractional data; Quasi-maximum likelihood estimator; Dirichlet regression; Regression-based specification tests. |
| JEL: | C35 |
| Date: | 2013 |
| URL: | http://d.repec.org/n?u=RePEc:cfe:wpcefa:2013_05&r=ecm |
| By: | José M.R. Murteira (Faculdade de Economia, Universidade de Coimbra, and CEMAPRE); Esmeralda A. Ramalho (Departamento de Economia and CEFAGE-UE, Universidade de Évora); Joaquim J.S. Ramalho (Departamento de Economia and CEFAGE-UE, Universidade de Évora) |
| Abstract: | This paper shows that a test for heteroskedasticity within the context of classical linear regression can be based on the difference between Wald statistics in heteroskedasticity-robust and nonrobust forms. The test is asymptotically distributed under the null hypothesis of homoskedasticity as chi-squared with one degree of freedom. The power of the test is sensitive to the choice of parametric restriction used by the Wald statistics, so the supremum of a range of individual test statistics is proposed. Two versions of a supremum-based test are considered: the first version does not have a known asymptotic null distribution, so the bootstrap is employed to approximate its empirical distribution. The second version has a known asymptotic distribution and, in some cases, is asymptotically pivotal under the null. A simulation study illustrates the use and .nite-sample performance of both versions of the test. In this study, the bootstrap is found to provide better size control than asymptotic critical values, namely with heavy-tailed, asymmetric distributions of the covariates. In addition, the use of well-known modifications of the heteroskedasticity consistent covariance matrix estimator of OLS coefficients is also found to benefit the tests'overall behaviour. |
| Keywords: | Heteroskedasticity testing; White test; Wald test; Supremum. |
| Date: | 2013 |
| URL: | http://d.repec.org/n?u=RePEc:cfe:wpcefa:2013_06&r=ecm |
| By: | Lijuan Huo (School of Economics, Yonsei University, Korea, Department of Applied Mathematics, School of Science, Changchun University of Science and Technology, China); Tae-Hwan Kim (School of Economics, Yonsei University, Korea); Yunmi Kim (Department of Economics, Kookmin University, Korea) |
| Abstract: | Quantile regression (QR) models have been increasingly employed in many applied areas in economics. At the early stage, applications took place usually using cross-section data, but recent development has seen a surge of the use of quantile regression in both time-series and panel datasets. However, how to test for possible autocorrelation, especially in the context of time-series models, has been paid little attention. As a rule of thumb, one might attempt to apply the usual Breusch-Godfrey LM test to the residuals from the baseline quantile regression. In this paper, we demonstrate by Monte Carlo simulations that such an application of the LM test can result in potentially large size distortions, especially in either low or high quantiles. We then propose two correct tests (named the F-test and the QR-LM test) for autocorrelation in quantile models, which do not suffer from any size distortion. Monte Carlo simulation demonstrate that the two tests perform fairly well in finite samples, across either different quantiles or different underlying error distributions. |
| Keywords: | Quantile regression, autocorrelation, LM test |
| JEL: | C12 C22 |
| Date: | 2013–02–13 |
| URL: | http://d.repec.org/n?u=RePEc:yon:wpaper:2013rwp-54&r=ecm |
| By: | Hess, Wolfgang (Department of Economics, Lund University); Persson, Maria (Department of Economics, Lund University); Rubenbauer, Stephanie (Department of Statistics, Ludwig-Maximilians-University Munich); Gertheiss, Jan (Department of Statistics, Ludwig-Maximilians-University Munich) |
| Abstract: | When analyzing panel data using regression models, it is often reasonable to allow for time-varying covariate effects. We propose a novel approach to modelling timevarying coefficients in panel data regressions, which is based on penalized regression techniques. To illustrate the usefulness of this approach, we revisit the well-known empirical puzzle of the 'death of distance' in international trade. We find significant differences between results obtained with the proposed estimator and those obtained with 'traditional' methods. The proposed method can also be used for model selection, and to allow covariate effects to vary over other dimensions than time. |
| Keywords: | Penalized Regression; Lasso-type Penalties; Varying Coefficient Models; Gravity; Death of Distance; Missing Globalization Puzzle |
| JEL: | C23 C52 F10 |
| Date: | 2013–02–19 |
| URL: | http://d.repec.org/n?u=RePEc:hhs:lunewp:2013_005&r=ecm |
| By: | Pingel, Ronnie (Department of Statistics, Uppsala University); Waernbaum, Ingeborg (IFAU - Institute for Evaluation of Labour Market and Education Policy) |
| Abstract: | In observational studies the overall aim when fitting a model for the propensity score is to reduce bias for an estimator of the causal effect. For this purpose guidelines for covariate selection for propensity score models have been proposed in the causal inference literature. To make the assumption of an unconfounded treatment plausible researchers might be tempted to include many, possibly correlated, covariates in the propensity score model. In this paper we study how the efficiency of matching and inverse probability weighting estimators for average causal effects change when the covariates are correlated. We investigate the case with multivariate normal covariates and linear models for the propensity score and potential outcomes and show results under different model assumptions. We show that the correlation can both increase and decrease the large sample variances of the estimators, and that the corrrelation affects the efficiency of the estimators differently, both with regard to direction and magnitude. Moreover, the strength of the confounding towards the outcome and the treatment plays an important role. |
| Keywords: | Efficiency bound; observational study; propensity score; variable selection |
| JEL: | C13 C40 C52 |
| Date: | 2013–02–13 |
| URL: | http://d.repec.org/n?u=RePEc:hhs:ifauwp:2013_005&r=ecm |
| By: | Neil Shephard |
| Abstract: | I discuss models which allow the local level model, which rationalised exponentially weighted moving averages, to have a time-varying signal/noise ratio. I call this a martingale component model. This makes the rate of discounting of data local. I show how to handle such models effectively using an auxiliary particle filter which deploys M Kalman filters run in parallel competing against one another. Here one thinks of M as being 1,000 or more. The model is applied to inflation forecasting. The model generalises to unobserved component models where Gaussian shocks are replaced by martingale difference sequences. |
| Keywords: | Auxiliary particle filter, EM algorithm, EWMA, forecasting, Kalman filter, likelihood, martingale unobserved component model, particle filter, stochastic volatility |
| JEL: | C01 C14 C58 D53 D81 |
| Date: | 2013–02–10 |
| URL: | http://d.repec.org/n?u=RePEc:oxf:wpaper:644&r=ecm |
| By: | Søren Johansen (University of Copenhagen and CREATES); Bent Nielsen (Nuffield College & Department of Economics, University of Oxford & Institute for New Economic Thinking at the Oxford Martin School.) |
| Abstract: | The Forward Search is an iterative algorithm concerned with detection of outliers and other unsuspected structures in data. This approach has been suggested, analysed and applied for regression models in the monograph Atkinson and Riani (2000). An asymptotic analysis of the Forward Search is made. The argument involves theory for a new class of weighted and marked empirical processes, quantile process theory, and a fixed point argument to describe the iterative element of the procedure. |
| Keywords: | Fixed point result, Forward Search, quantile process, weighted and marked empirical process |
| JEL: | C22 |
| Date: | 2013–10–02 |
| URL: | http://d.repec.org/n?u=RePEc:aah:create:2013-05&r=ecm |
| By: | Laffers, Lukas (Dept. of Economics, Norwegian School of Economics and Business Administration) |
| Abstract: | This paper provides a new simple and computationally tractable method for determining an identified set that can account for a broad set of economic models when economic variables are discrete. Using this method it is shown on a simple example how can imperfect instruments affect the size of the identified set when strict exogeneity is relaxed. It could be of great interest to know to what extent are the results driven by the exogeneity assumption which is often a subject of controversy. Moreover, flexibility gained from the new proposed method suggests that the determination of the identified set need not be application-specific anymore. This paper presents a unifying framework that approaches identification in an algorithmic way. |
| Keywords: | Identification; Models; Discrete Variables. |
| JEL: | C10 C21 C26 C61 |
| Date: | 2013–01–08 |
| URL: | http://d.repec.org/n?u=RePEc:hhs:nhheco:2013_001&r=ecm |
| By: | Marczak, Martyna; Gómez, Victor |
| Abstract: | This article proposes a new multivariate method to construct business cycle indicators. The method is based on a decomposition into trend-cycle and irregular. To derive the cycle, a multivariate band-pass filter is applied to the estimated trend-cycle. The whole procedure is fully model-based. Using a set of monthly and quarterly US time series, two monthly business cycle indicators are obtained for the US. They are represented by the smoothed cycles of real GDP and the industrial production index. Both indicators are able to reproduce previous recessions very well. Series contributing to the construction of both indicators are allowed to be leading, lagging or coincident relative to the business cycle. Their behavior is assessed by means of the phase angle and the mean phase angle after cycle estimation. The proposed multivariate method can serve as an attractive tool for policy making, in particular due to its good forecasting performance and quite simple setting. The model ensures reliable realtime forecasts even though it does not involve elaborate mechanisms that account for, e.g., changes in volatility. -- |
| Keywords: | business cycle,multivariate structural time series model,univariate band-pass filter,forecasts,phase angle |
| JEL: | E32 E37 C18 C32 |
| Date: | 2013 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:fziddp:642013&r=ecm |
| By: | Majid M. Al-Sadoon |
| Abstract: | This paper extends multivariate Granger causality to take into account the subspaces along which Granger causality occurs as well as long run Granger causality. The properties of these new notions of Granger causality, along with the requisite restrictions, are derived and extensively studied for a wide variety of time series processes including linear invertible process and VARMA. Using the proposed extensions, the paper demonstrates that: (i) mean reversion in L2 is an instance of long run Granger non-causality, (ii) cointegration is a special case of long run Granger non-causality along a subspace, (iii) controllability is a special case of Granger causality, and finally (iv) linear rational expectations entail (possibly testable) Granger causality restriction along subspaces. |
| Keywords: | Granger causality, long run Granger causality, L2 -mean-reversion, p-mixing, cointegration, VARMA, controllability, Kalman Decomposition, linear rational expectations |
| JEL: | C10 C32 C51 |
| Date: | 2013–01 |
| URL: | http://d.repec.org/n?u=RePEc:upf:upfgen:1356&r=ecm |
| By: | Stefano Grassi (Aarhus University and CREATES); Paolo Santucci de Magistris (Aarhus University and CREATES) |
| Abstract: | The persistent nature of equity volatility is investigated by means of a multi-factor stochastic volatility model with time varying parameters. The parameters are estimated by means of a sequential indirect inference procedure which adopts as auxiliary model a time-varying generalization of the HAR model for the realized volatility series. It emerges that during the recent financial crisis the relative weight of the daily component dominates over the monthly term. The estimates of the two factor stochastic volatility model suggest that the change in the dynamic structure of the realized volatility during the financial crisis is due to the increase in the volatility of the persistent volatility term. As a consequence of the dynamics in the stochastic volatility parameters, the shape and curvature of the volatility smile evolve trough time. |
| Keywords: | Time-Varying Parameters, On-line Kalman Filter, Simulation-based inference, Predictive Likelihood, Volatility Factors |
| JEL: | G01 C00 C11 C58 |
| Date: | 2013–02–18 |
| URL: | http://d.repec.org/n?u=RePEc:aah:create:2013-03&r=ecm |
| By: | Kevin Sheppard; Lily Liu; Andrew J. Patton |
| Abstract: | We study the accuracy of a wide variety of estimators of asset price variation constructed from high-frequency data (so-called "realized measures"), and compare them with a simple "realized variance" (RV) estimator. In total, we consider almost 400 different estimators, applied to 11 years of data on 31 different financial assets spanning five asset classes, including equities, equity indices, exchange rates and interest rates. We apply data-based ranking methods to the realized measures and to forecasts based on these measures. When 5-minute RV is taken as the benchmark realized measure, we find little evidence that it is outperformed by any of the other measures. When using inference methods that do not require specifying a benchmark, we find some evidence that more sophisticated realized measures significantly outperform 5-minute RV. In forecasting applications, we find that a low frequency "truncated" RV outperforms most other realized measures. Overall, we conclude that it is difficult to significantly beat 5-minute RV. |
| Keywords: | Realized variance, volatility forecasting, high frequency data |
| JEL: | C58 C22 C53 |
| Date: | 2013–02–12 |
| URL: | http://d.repec.org/n?u=RePEc:oxf:wpaper:645&r=ecm |
| By: | Hsiu-Hsin Ko (National University of Kaohsiung); Masao Ogaki |
| Abstract: | We use a residual-based bootstrap method to re-examine the finding of the Granger causality relationship from exchange rates to fundamentals in Engel and West (Exchange rate and fundamentals, Journal of Political Economy 2005, 113 (3), 485–517), in which the evidence for the relation is taken as evidence for the present-value model for exchange rates. The test results are against the previous findings. The Monte Carlo experiment results suggest that the causality test implemented in the previous study tends to spuriously reject null hypotheses. Thus, the existing evidence for the present value model for exchange rates is not robust. |
| Keywords: | Bootstrap, Granger causality, exchange rates, fundamentals |
| JEL: | F30 F31 C32 |
| Date: | 2013–02 |
| URL: | http://d.repec.org/n?u=RePEc:roc:rocher:577&r=ecm |