Open AccessThe effects of company characteristics on financial reporting quality – the application of the machine learning technique
Author(s) -
Željana Aljinović Barač,
Mario Bilić
Publication year - 2021
Publication title -
ekonomski vjesnik
Language(s) - English
Resource type - Journals
eISSN - 1847-2206
pISSN - 0353-359X
DOI - 10.51680/ev.34.1.5
Subject(s) - profitability index , xbrl , listing (finance) , business , financial statement , accounting , voluntary disclosure , quality (philosophy) , accounting management , affect (linguistics) , audit , stock market , finance , transparency (behavior) , accounting information system , computer science , philosophy , epistemology , paleontology , linguistics , computer security , horse , biology
Purpose: The paperaims to determine the level of financial reporting quality (FRQ) in listedcompanies in Croatia, as an example of a macro-based accounting system with anunderdeveloped capital market, and identify company characteristics that affectit. The paper systematizes the existing key knowledge of FRQ. Furthermore, itcritically analyses the principles and direction of influence of variousqualitative and quantitative as well as financial and non-financialcharacteristics of a company. Methodology: The empirical analyses involve jointtesting of the machine learning technique (MLT) and the economichypotheses. M5 algorithm is applied to identify the factorsthat influence the quality of voluntary reporting as well as the direction andintensity of their influence. Results: The resultsshow that profitability, stock market listing duration (in years), and companysize positively affect the level and extent of FRQ through voluntary disclosureof information in the annual financial reports of Croatian listed companies. Inaddition, differences in FRQ between different areas of economic activity anddepending on the type of auditor were found. Conclusion: Croatiancompanies should adopt good reporting practices to meet the requirements of theglobal market and thus contribute to the improvement of the overalltransparency system. The same is expected from the relevant regulatoryauthorities who should encourage full disclosure. The paper provides severalscientific contributions: first, the spatial dimension of the research; second,the comprehensive literature review; and third, the MLT application in the researchon FRQ. An important practical implication of these findings is that they willhelp financial statement users in the economic decision-making process.