A market data and news-based early warning system for credit risk
A market data and news-based early warning system for credit risk


Risk Guard identifies signs of financial instability in listed and non-listed corporates and financial institutions, industry sectors and countries up to one year in advance.
Using an early warning system is not only reasonable to prevent losses but is also often required by law.


Scope of application

Internal and external ratings, as used for fundamental risk assessment, cannot immediately reflect abrupt changes in macroeconomic conditions or in a company’s business situation. For this reason it makes sense to use additional instruments that are designed to identify increases in credit risk at a very early stage. The solution proposed by RSU is Risk Guard, a market data and news-based early warning system.

Risk Guard implements many models for predicting downgrades or defaults based on the following categories:

  • listed companies
  • non-listed companies
  • financial institutions
  • industry sectors
  • countries and regions

It enables decision makers to be proactive rather than reactive, for example by intensifying analysis, reviewing ratings or adjusting their portfolio or investment strategy.

Risk Guard can be used in all areas where increases in credit risk need to be identified as early as possible, for example

  • credit risk monitoring
  • risk management
  • portfolio management
  • asset management
  • treasury


Risk Guard computes risk scores for the selected entities on a daily basis. It predicts significant changes in creditworthiness and flags the obligors concerned so that analysts can take a closer look at them. According to the type of information available, different quantitative methods are used. Whereas the models used for assessing listed companies, countries, regions and industry sectors are multi-factor models based on (financial) market data, unlisted companies are analysed by means of a Merton-based model which combines industry-specific market data and financial statement information.

In addition, news-based models have been developed for both listed and non-listed corporates and financial institutions using advanced Machine Learning methods. German language news reports from all relevant sources are processed every day, which results in a risk score and possibly a warning signal.
RSU regularly validates all models and develops them further.


Risk Guard is an easy to set up system for identifying and administrating warning signals. Since it can also process signals from other, internal sources, it can even serve as a central early warning platform.

Critical cases are systematically evaluated. The results of this evaluation can be clearly tracked over time.
An entity is flagged as critical if any of the assigned models has produced a warning signal. Critical cases must be assigned to individually defined evaluation categories. These usually correspond to (subsequent) internal processes. Both the changes in risk scores and the corresponding evaluations remain transparent over time.

Risk Guard is a web-based application available with both German and English text; it only requires a VPN (or CredNet) connection. Like all applications provided by RSU, it is operated under the strict security standards customary in the banking industry. Protected transmission channels and secure data storage at a level 3-certified data processing centre ensure a high degree of data security.