| Abstract [eng] |
Sometimes there happen different crisis in economics, sometimes occur natural disasters or other phenomenons which substantially affect the behaviour of the economy. This impact reflects in a time series of economic indicators whose behaviour may go into the melting-pot after some crisis or natural disaster. Usually such a change is called a structural change. Very often structural change is hardly visible then this problem can be solved by applying special tests for detecting potential structural change. In this paper three known tests for identification of structural change are examined. The aim of this work is to compare the power of these tests and propose the new one which will have more desirable features. During the investigation of the tests that are designed for detecting structural changes it emerged that CUSUM test is not fitted for the case of epidemic structural change. Whereas the power of other tree tests that are designed for detecting epidemic structural changes differ narrowly, however the new test has the advantage against the other tests – it does not treat an outlier as a possible structural change. |
