Statystyczna analiza wyników testu koniunktury

Results of business activity tests are characterized by very strong structural variation. Complex, multidimensional structure of population of enterprises being considered is the main source of diversity. Dynamic approach introduces periodical variation in values of certain variables, intensity of observed processes, and quality of ex post predictions. For instance, in construction industry, one factor which introduces the largest variation is the criterion of ownership. Also for industrial enterprises and trading companies, the largest differences are observed between groups of public sector and private sector enterprises. For the purpose of analysis of business activity tests results, traditional statistical methods may be used. In this chapter we introduce point estimators and interwal estimators of balances and business activity indicators. We consider the problem of aggregation of balances and define distribution of aggregated balances and its main moments. We present a construction of a significance test with test statistics A which has asymptotically normal distribution. The test measures significance of differences of business activity indicators between selected structural groups. We list assumptions necessary to obtain probability distributions of appropriate variables and show the most important properties of these distributions. Applications of proposed statistical analysis methods are described using construction industry activity data for Poland and Italy. As an example of conclusions drawn for Poland, we found strong correlation between assessments of business activity and sociopolitical conditions, as well as unfair treatment of different groups of business enterprises by central administration. Analysis of seasonal variation of construction industry activity confirmed legitimacy of assumed business activity indicator. It is a leading indicator which precedes changes in economic situation by about 3 months. Results of business activity tests for Italian industry do not confirm existence of significant seasonal variation of business activity. Results aggregated over 30 years of conducting tests in Italy clearly indicate that in construction industry, unfavorable periods outnumber favorable ones.