A Simple Representation of the Bera-Jarque-Lee Test for Probit Models
Joachim Wilde
IWH Discussion Papers,
No. 13,
2007
Abstract
The inference in probit models relies on the assumption of normality. However, tests of this assumption are not implemented in standard econometric software. Therefore, the paper presents a simple representation of the Bera-Jarque-Lee test, that does not require any matrix algebra. Furthermore, the representation is used to compare the Bera-Jarque- Lee test with the RESET-type test proposed by Papke and Wooldridge (1996).
Read article
Prekäre Einkommenslagen in Deutschland: Ein Ost-West-Vergleich 1996 bis 2002
Herbert S. Buscher, Juliane Parys
IWH Discussion Papers,
No. 2,
2006
Abstract
The paper investigates the distribution of equivalence-weighted net household income for West and East Germany, covering the period from 1996 to 2002. The data set used is the annual cross section data set “Mikrozensus”. The main issues of the paper are twofold. First, we analyze standard measures of income distributions as well as measures of inequality. Second, we set up a Logit model to explain relative poorness in East and West Germany using Mikrozensus data to capture household characteristics. The main focus in this section deals with the question how different types of forms of living and the number of children will affect the risk of falling into precarious income situations. The results show that the risk of getting poor is higher for families with children as well as for single persons with children.
Read article
A note on GMM-estimation of probit models with endogenous regressors
Joachim Wilde
IWH Discussion Papers,
No. 4,
2005
Abstract
Dagenais (1999) and Lucchetti (2002) have demonstrated that the naive GMM estimator of Grogger (1990) for the probit model with an endogenous regressor is not consistent. This paper completes their discussion by explaining the reason for the inconsistency and presenting a natural solution. Furthermore, the resulting GMM estimator is analyzed in a Monte-Carlo simulation and compared with alternative estimators.
Read article