Publications of Kőrösi, G.

The determinants of foreign direct investment in transforming economies: A comment

We are deeply convinced that econometric and other statistical methods are essential tools of empirical economic analysis. However, they are useful only if applied properly. There is an overwhelming evidence that Zhen Quan Wang and Nigel J. Swain failed to do so in their referred article. All these problems of the empirical analysis suggest that the claims made by the paper are not supported by the data and the estimated models and can (and should) be considered as the results of pure speculation.

Aggregation and cointegration

The analysis of economic time series assumes specific economic behaviour of a representative agent. The data used in analysis is generated by aggregating observations of all individuals in a population. This is valid only if all members of a population have the same data generating process, but what happens if their behaviour is heterogeneous? This paper examines the properties of test statistics for cointegration when the aggregate data consists of heterogeneous individuals.

Comparative review of some econometric software packages

Reviews several econometric software packages. Aremos for estimation methods and tests; Pc-Give used for general instrumental variable estimation; IAS (Interactive Simulation system-Bonn) used for extensive policy simulations; LIMDEP(LIMited DEPendents variables) used for estimating and analysing models with limited dependent variables.

Practical econometrics

Textbook, suitable for undergraduate or graduate courses in econometrics up to the intermediate level, focusing on the practical use of the methods and on creating an understanding of the main concepts, ideas, and motivation. Considers preparations for econometric analysis, including a discussion of the database. Covers single equation models, discussing the basic model, stochastic regressors, time series analysis, dynamic specification, heteroskedastic disturbances, non-normal disturbances, multicollinearity, nonlinear models, structural stability, external (prior) information, joint tests of hypotheses, model selection, panel data based modeling, and analysis and forecast. Treats inference in simultaneous models, covering the classical simultaneous model and its estimation; hypothesis testing in linear simultaneous equation models; and analysis and prediction of simultaneous models. Addresses other selected topics, including investigating causal relationships by econometric methods; models with qualitative and limited dependent variables; and latent variables. Provides information on econometric software packages. Korosi is with the Hungarian Academy of Sciences. Matyas is at Monash University. Szekely is at Budapest University of Economics. Index.