• credit risk
• default
• distance-to-default (DD)
• expected default frequency (EDF)
• forecasting power
• KMV model
• Merton model

The structural changes in the recent years have focused attention on an urgent need to develop an effective prediction model in order to avoid unexpected company default. Default risk is the possibility that a firm does not meet its debt on time. But defaults are unpredictable and costly and the best we can do it is to calculate the probability that these events will happen.
The purpose of this dissertation is to present the basic idea of Merton model and KMV model and also to determine if the DD indicator has any forecasting power in predicting firm’s default. We use real data to examine the probability of default of several firms over the period 2004 - 2012 in the EU and in the USA.
Our results indicate that the KMV model is useful to identify risk levels and thus discriminate firms. To a large extend is able to predict default even if the limited number of firms considered, the simultaneous calculation of unobservable variables, the simplified assumptions on which the model is based, the lack of transparency of the financial markets, do not increase the reliability of the forecast.
However, in spite of these limitation, the Merton approach (and consequently the KMV approach) remain a very fundamental approach to credit risk models.