Building panels from archives: the longitudinal representativity

• Published in: Metron (Rome) 2022-04, Vol.80 (1), p.121-138
• Main Authors: Biffignandi, Silvia, Zeli, Alessandro
• Format: Article
• Language: English
• Subjects: Mathematics and Statistics; Statistical Theory and Methods; Statistics
• ISSN: 0026-1424; 2281-695X
• DOI: 10.1007/s40300-021-00199-5

In this paper, starting from ASIA business register and from the related Events Register, the longitudinally persistency of core units of the abovementioned database is examined. The M&A events are considered. To evaluate the representativity of the longitudinal panel obtained using the proposed approach, the dataset is integrated with the financial reports database with reference to some relevant economic variables. Italian financial reports database includes annual reports transmitted (according to the law) to the local Chamber of Commerce (CC) by all limited firms. The annual reports are collected and checked by CC. The database includes around 700,000 firm financial reports every year. It represents a very rich and complete source of information, therefore in this study it is used as reference target population. The complete set of data is compared both to panel based on the M&A (Mergers and Acquisitions) events and to results related to a panel build up without reference to the M&A events. Comparative analyses are carried out for the representativeness of most important economic variables. Representativeness with respect to main economic variables is analyzed. Statistical analyses, broken down by specific categories, based mainly on R -indicators are performed.