C32 Econometric Methods: Multiple/Simultaneous, Equation models: Time‑Series Models
Authors:
RNDr. Jitka Pomenkova, Ph.D.

Institution:
Mendel University, Faculty of Business Economics, Department of Statistics and Operational Analysis, Brno, Czech Republic

Keywords:
time series analysis, cointegration, VAR models, Monte Carlo simulations

Background - hypothesis:
Effective application of economy policy instruments assume, from the economy policy maker point of view, knowledge of impact of these instruments to the real economy. Consequences identification of instruments application, causality between economy policy instruments and macroeconomic indicators is possible by use of logical assumption, supposition and economy theories. Even though economy theory gives important information about expected causality or relation between time series and their character, for its application in practice it is necessary to verify results by empirical analysis.

Objectives:
Presented article deals with different problems during empirical causality validation of the macroeconomic time series in post-transform economy.
The characteristic attribute of post-transform economy, from time series dependency analysis point of view, is sample size. If we consider institutional conditions in the middle and east European countries before Soviet union disintegration and their nowadays position in the process of European integration, we can analyse many changes in the structure of institutions given economies. These changes can be considered as structural changes.
Common use processes allude to several problems, for example small sample size, model parsimodity, variable identification, detection of structural change moment, etc.
Discussed paper deals with causality testing between chosen macroeconomic time series of post-transform economies from chosen methodological point of view. Further attention is paid to small sample size impact on analysis results and on the suggestion of possible solution approaches.

Data/Methods:
Sources of data: OECD database

Expected Results:
Many structural changes are identified in the macroeconomic time series of the post-transform economies. From this reason, many statistical hypothesis testing is limited by sample of size. Concurrently, different methods are more appropriated for the different causality problem identification.

Contact person
RNDr. Jitka Pomenkova, PhD.
Mendel University of Agriculture and Forestry
Faculty of Business Economics
Department of Statistics and Operational Analysis
Zemedelska 1
613 00 BRNO
Czech Republic
email: pomenka@mendelu.cz