The subject of the present award procedure is the provision of services to support the professionally, technically, mathematically and methodologically qualified consultants in order to achieve the project objective defined below.

LOT-0001
L-Bank - Europe-wide procurement of services for the development of a new risk classification procedure for corporate clients.
L-Bank is the promotional institution of the state of Baden-Württemberg. Its business activities are governed by the legal mandate to assist the country in the performance of its public tasks, in particular in the areas of structural, economic and social policies, while implementing support measures in accordance with European Union state aid rules. The subject of the present award procedure is the provision of services to support the professionally, technically, mathematically and methodologically qualified consultants in order to achieve the project objective defined below. The aim of the development project is to develop a mathematically and statistically developed probability of default rating (PD rating) for companies, which can be applied to small and medium-sized enterprises (SMEs) as well as to large enterprises. The result of the model is supposed to be a one-year PD. No algorithms based on artificial intelligence may be used. The model should take into account all necessary risk drivers. The following are some examples of possible risk drivers: - Balance sheet data (if applicable) - Market data - Qualitative factors - Environmental, social and governance (ESG) factors - Segmentations according to the size of the company, industry, country of residence, etc. Among other things, L-Bank's internal data history is available for development, which includes internal balance sheet data, qualitative factors and, for SMEs, a sufficiently large default history. Expanding the data to include external factors such as market data cannot be ruled out. In addition to the selection and weighting of model risk factors on the basis of standard univariate and multivariate methods, the model will be calibrated to the long-term default rate of the portfolio. In doing so, it is necessary to take into account the fact that the internal default rate differs from SMEs to large companies or possible other intersection dimensions of the portfolio. The model should be structured in such a way that scenario or sensitivity calculations can also be carried out downstream for longer periods of time. However, these calculations are not the subject of the service. As L-Bank uses the standard credit risk approach for regulatory purposes, it is not mandatory to meet the stricter requirements of the internal rating-based approach (IRB) approach. The development will therefore focus on the suitability of the rating procedure for L-Bank's internal management purposes. The data preparation required for the modelling and the modelling itself should be documented in a comprehensible form. The effects of the introduction of the new PD rating for companies will be examined at the PD and rating level according to different intersection dimensions of the portfolio. As a prerequisite for these impact analyses, a prototype of the new rating procedure (in MS Excel) is to be developed. The above-mentioned project objective results in the following scope of work: New development of a risk classification method for corporate customers - Data specification and provision: Raw data is required for development, which must first be extracted from various data sources and compiled. In this context, a rough specification of the necessary data is required in advance, so that the data can be provided by L-Bank before the start of the project. - Data preparation: The main goal is to convert the raw data into a form that is suitable for the development of a rating model: - A definition of a key date grid and enrichment of the key date grid with information valid on the respective key date, such as master data, quantitative and qualitative factors, segmentation characteristics or historical rating results. - Selection of relevant (credit risk-exposed) customers - Derivation of a 1-year default indicator - Possibly enrichment of internal data - Model development: - Creation of quantity structures - Coordination of a risk factor longlist - Development of a quantitative module (incl. standard univariate and multivariate analyses) - Development of a qualitative module (incl. standard univariate and multivariate analyses) - If necessary, consideration of Segmentation effects due to different industries, size, country of residence, etc. - Calibration - Integration of downstream rating components - Documentation: During the above-mentioned work steps, documentation must be created continuously, which must be finalized before the project is completed. - Prototype: A prototype is developed in MS Excel that is able to calculate the PDs and ratings according to a newly developed method over the portfolio on a key date. The basis for the implementation of the prototype is the documentation provided by the previous model development. - Impact analysis: The impact analysis is based on the most recent cut-off date of the database on which the model is based. With the help of the prototype, the portfolio is evaluated at this deadline and impact analyses are prepared at the PD level. This includes, in particular, migration analyses with regard to the change in the rating result from the current corporate customer rating to the newly developed rating procedure. - Communication with the client and documentation of the work results should be in German. The service is intended to be provided for a period of approximately 12 months. The services will start after the procurement procedure has been completed.