REGRESSION QUANTILES UNDER HETEROSCEDASTICITY AND MULTICOLLINEARITY: ANALYSIS OF TRAVEL AND TOURISM COMPETITIVENESS

• Authors: Kalina Jan , Vašaničová Petra , Litavcová Eva
• Languages of publication: EN

Abstracts

EN

In the linear regression, heteroscedasticity and multicollinearity can be characterized as intertwined problems, which often simultaneously appear in econometric models. The aim of this paper is to discuss various approaches to regression modelling for heteroscedastic multi collinear data. A real economic dataset from the World Economic Forum serves as an illustration of various individual methods and the paper provides a practical motivation for quantile regression and particularly for regularized regression quantiles. In the dataset, tourist service infrastructure across 141 countries is modelled as a response of 12 characteristics of the Travel and Tourism Competitiveness Index (TTCI). Regression quantiles and their lasso estimates turn out to be more suitable for the dataset compared to more traditional econometric tools.

Keywords

EN

LINEAR REGRESSION; MODEL SELECTION; ROBUSTNESS; REGRESSION QUANTILES; LASSO; TOURISM

• Publisher: Institute of Slovak and World Economics and Institute for Forecasting, Slovak Academy of Sciences
• Journal: Ekonomický časopis (Journal of Economics)
• Year: 2019
• Volume: 67
• Issue: 1
• Pages: 69 – 85

Contributors

author

Kalina Jan

• Institute of Computer Science of the CAS, Pod Vodárenskou věží 2, 182 07 Praha 8, Czech Republic

author

Vašaničová Petra

author

Litavcová Eva