Implementation of reliable rating systems for small credit portfolio is hindered by non-observed default events in databases and short time series of data available. In this study we propose an approach to handle those two challenges while developing rating systems. We further extend the approach by estimating systematic risk, that is, co-movements of creditworthiness of debt securities' issuers over time. Based on financial information from the PSVaG's debt securities portfolio, we could show that including a systematic risk component significantly increase the model accuracy.