Suchergebnisse

'Dieser Beitrag gibt eine Einführung in die verschiedenen Verfahren zur regressionsanalytischen Behandlung von Panel-Daten, in der auf Ableitungen und eine extensive Benutzung von Matrix-Algebra verzichtet wird. Allerdings kann die Darstellung nicht ganz ohne eine formale Schreibweise auskommen, wobei jedoch im letzten Abschnitt anhand eines konkreten Rechenbeispiels die benutzten Formeln näher erläutert werden. Auf diese Weise möchte der Beitrag einserseits aufzeigen, daß die Regressionsanalyse mit Panal-Daten im wesentlichen nur auf rechentechnisch einfach durchzuführende Transformationen der Daten hinausläuft und mit Hilfe der üblichen Statistik-Programmpakete (z. B. SPSS) durchgeführt werden kann, und andererseits einen leichten Zugang zur bestehenden Lehrbuchliteratur ermöglichen.' (Autorenreferat)

Panel data methods help understand relationships between variables (e.g., features of individuals, firms, states, taxpayers) located in specific environments (e.g., within jurisdictions, established groups, cohorts, geographic locations), over time. Because these environments have certain unique characteristics, the differences between the variables occurring because of the environments they are located in may lead to observable or unobservable heterogeneity in the data. To account for this variation, to correct for heterogeneity in statistical estimates, and to increase the efficiency of estimators, researchers use panel data methods. Because panel data models capture more variance, less collinearity among the variables, more degrees of freedom, and higher efficiency of the estimators, their estimates are more informative compared with the estimates calculated by time-series or cross-sectional ordinary least squares regression methods. This case introduces panel data regressions and reviews the two approaches to their estimation: fixed-effects regression methods and random-effects regression methods. Next, it explains the conditions appropriate for using each method and shows how to test for assumptions the random-effects regression method needs. This case is based on a peer-reviewed paper examining assessor consolidation reform in the state of Indiana in 2008 and its effects on cost savings. In this study, all observations were located within the counties and the data spanned between 2005 and 2014. While the original publication demonstrated the cost savings that resulted from the consolidation reform, the purpose of this case is to show how to use panel data to estimate economies of scale in the public sector.

The paper explores the effect of measurement errors on the estimation of a linear panel data model. The conventional fixed effects estimator, which ignores measurement errors, is biased. By correcting for the bias one can construct consistent and asymptotically normal estimators. In addition, we find estimates for the asymptotic variances of these estimators. The paper focuses on multiplicative errors, which are often deliberately added to the data in order to minimize their disclosure risk. They can be analyzed in a similar way as additive errors, but with some important and consequential differences.

This study aims to analyze how the demographic structure affects carbon dioxide (CO2) emissions in Top Emitters, namely China, the United States, the European Union (EU-28), India, Indonesia, Russia, Brazil, Japan, Canada, and Mexico. This study uses panel data from ten countries stated in Top Emitters for the period 2000-2014 sourced from the World Resource Institute, World Bank and UNESCO Institute for Statistics. This study uses the Panel Data Regression method with the best model chosen is the Random Effect Model (REM) and four demographic structure variables, namely the dependency ratio, sex ratio, higher education ratio, industrial employment ratio. The results of this study indicate that the dependency ratio, sex ratio, higher education ratio, industrial employment ratio have a significant effect on carbon dioxide (CO2) emissions in Top Emitters. The results of this study are expected to provide policies that can be implemented by the government.Keywords: Demographic Structure, Top Emitters, Panel Data Regression MethodJEL : I25, O15, Q5

This paper concerns the use of synthetic data for protecting the confidentiality of business data during statistical analysis. Synthetic data sets are traditionally constructed by replacing sensitive values in a confidential data set with draws from statistical models estimated on the confidential data set. Unfortunately, the process of generating effective statistical models can be a difficult and labour-intensive task. Recently, it has been proposed to use easily-implemented methods from machine learning instead of statistical model estimation in the data synthesis task. J. Drechsler and J.P. Reiter (2011) have conducted an evaluation of four such methods, and have found that regression trees could give rise to synthetic data sets which provide reliable analysis results as well as low disclosure risks. Their conclusion was based on simulations using a subset of the 2002 Uganda census public use file. It is an interesting question whether the same conclusion applies to other types of data with different characteristics, for example business data which have quite different characteristics from population census and survey data. In particular, business data generally have few variables that are mostly categorical, and often have highly skewed distributions with outliers.

In this paper we investigate the applicability of regression-tree-based methods for constructing synthetic business data. We give a detailed example comparing exploratory data analysis and linear regression results under two variants of a regression-tree-based synthetic data approach. We also include an evaluation of the analysis results with respect to the results of analysis of the original data. We further investigate the impact of different stopping criteria on performance.

While it is certainly true that any method designed to protect confidentiality introduces error, and may indeed give misleading conclusions, our analysis of the results for synthesisers based on CART models has provided some evidence that this error is not random but is due to the particular characteristics of business data. We conclude that more careful analysis needs to be done in applying these methods and end users certainly need aware of possible discrepancies.

We develop a non-dynamic panel smooth transition regression model with fixed individual effects. The model is useful for describing heterogenous panels, with regression coefficients that vary across individuals and over time. Heterogeneity is allowed for by assuming that these coefficients are continuous functions of an observable variable through a bounded function of this variable and fluctuate between a limited number (often two) of extreme regimes. The model can be viewed as a generalization of the threshold panel model of Hansen (1999). We extend the modelling strategy for univariate smooth transition regression models to the panel context. This comprises of model specification based on homogeneity tests, parameter estimation, and diagnostic checking, including tests for parameter constancy and no remaining nonlinearity. The new model is applied to describe firms' investment decisions in the presence of capital market imperfections.