eThesis - School of Business electronic theses

» hse_ethesis_13096.pdf   size:818 KB (836743)

ABSTRACT This paper provides new insights into payment behavior prediction by using daily ledger data from a case company. The results indicate that a model based on ledger information is more accurate and efficient in future payment failure prediction than models traditionally based on financial and background variables. More specifically, the results implicate that a company benefits the most from credit risk rating services when its own ledger data are used as a part of the rating scoring model. In addition, the results have three implications. First, ledger information can be used in the future to create more accurate and up-to-date credit rating scores, which reduces the dependency over publicly available information. Second, the length of the client firm-creditor firm relationship seems to significantly reduce problems of informational asymmetry. Third, problems of asymmetric information, such as adverse selection, lemon’s problem, moral hazard and agency costs could be significantly reduced in the future using a shared information pool deriving from corporate ledgers.