Working Paper Cnr-Ceris, N° 04/2011 

An artificial neural network approach for assigning rating judgements to Italian Small Firms

Greta Falavigna
Cnr-Ceris researcher
Ph.D. in Economics and Technology Management

Cnr-Ceris
Via Real Collegio, 30
10024 Moncalieri (To)
Tel. 011-6824941
e-mail: This e-mail address is being protected from spam bots, you need JavaScript enabled to view it

ABSTRACT:Based on new regulations of Basel II Accord in 2004, banks and financial nstitutions have now the possibility to develop internal rating systems with the aim of correctly udging financial health status of firms. This study analyses the situation of Italian small firms
that are difficult to judge because their economic and financial data are often not available. The intend of this work is to propose a simulation framework to give a rating judgements to firms presenting poor financial information.
The model assigns a rating judgement that is a simulated counterpart of that done by Bureau van Dijk-K Finance (BvD). Assigning rating score to small firms with problem of poor availability of financial data is really problematic. Nevertheless, in Italy the majority of firms are small and there is not a law that requires to firms to deposit balance-sheet in a detailed form. For this reason the model proposed in this work is a three-layer framework that allows us to assign ating judgements to small enterprises using simple balance-sheet data.

KEYWORDS: rating judgements, artificial neural networks, feature selection
JEL CODES: C15; C45; G24

I thank Secondo Rolfo (Director Ceris-CNR), Giulio Calabrese (Ceris-CNR) and Alessandro Manello (Ceris-CNR) for percipient remarks about the present study, as well as Enrico Viarisio for the commendable research assistance. This study has been financed by Piedmont Region through the project “ICT Converging on Law: Next Generation Services for Citizens, Enterprises, Public Administration and Policymakers”, related to the announcement “Converging Technologies 2007”. The usual disclaimer applies.”
 

download