LB-Rating is a system designed for preparing, editing, validating and managing internal ratings in accordance with the Basel III/IV framework. It provides standardized and objective credit ratings for various types of obligors as well as for specialized lending.
Currently, LB-Rating comprises twelve modules. Clients only acquire licenses for the modules they need for their specific business. This offers them flexibility and a cost-efficient way of meeting their needs.
Moreover, a client’s own rating systems can be easily integrated in LB-Rating.
A rating evaluates the creditworthiness of a borrower. RSU’s rating models also assign an explicit probability of default to each customer. The factors that influence the probability of default differ, for example, between corporate customers, sovereigns, and banks. For this reason, different models have been developed:
AIRCRAFT FINANCING
The module calculates the probability of default, loss given default and expected loss for aircraft financing exposures. The results are aggregated and shown in annualized form and as a time profile. Transactions ranging from the acquisition of individual planes to that of entire fleets can be assessed.
All standard forms of financing, such as loans to airlines, finance leases, operating leases and warehouse facilities, can be treated adequately in dedicated sub-modules.
Special characteristics
The module is designed to cover the contractual arrangements of current financing structures. Participations in syndicated loans can also be processed due to the specific loss calculations performed for each tranche. The positive effects of favourable loan-to-value ratios (LTVs), efficient operating lessors or servicers (in the case of warehouse facilities), and of residual value and payment guarantees are considered in the default simulation in an adequate manner.
Loss-minimizing instruments such as ECAs (export guarantees), guarantees (payments guarantees, residual value guarantees, deficiency guarantees) and collateral (liquidity reserves, maintenance reserves, movable and immovable assets) can be included in the calculation. The loss amounts are calculated for each tranche, taking into account their individual collateralization and the seniority of the claim in an adequate manner. Cross-collateralized and cross-defaulted loans can be rated without difficulties.
The Aircraft Financing module in practice
In 2007 the module was revised and updated to reflect current trends in aircraft financing. It has been routinely reviewed every year since 2009. The module has good statistical accuracy due to a comparatively large database of almost 12,600 pooled ratings.
In late 2009, the module received supervisory approval for use under the IRBA for single-airline transactions, its main area of application.
BANKS
The Banks module can be used to rate any institution worldwide that does typical banking business and is accountable to a supervisory authority, irrespective of its legal form. This includes, among others, the following types of banks: financial supermarket, universal bank, investment bank, mortgage bank, savings and loans associations/building society, savings bank, cooperative bank, regional bank, bank holding, state-owned or government-related bank, supranational bank. A banking licence is not a prerequisite.
Special characteristics
The module is based on a scorecard approach. In addition to the quantitative and qualitative ratios, a banking system rating is also included in the calculation. In addition, the ownership situation, any potential government support, or membership in an institutional protection scheme are taken into account.
In order to ensure a global scope of application, the module distinguishes different geographic segments. Transfer risk is taken into account for foreign institutions.
Since the Banks module shares a number of methodological functions with the Insurance Companies module, there is potential for synergies if both modules are used.
The Banks module in practice
The module has been routinely reviewed every year since 2005. It has a very high statistical accuracy due to a large database containing almost 100,000 ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
CORPORATES
The Corporates module is designed for rating companies worldwide. The module is divided into several submodules to reflect geographic differences.
Special characteristics
The rating algorithm is based on a scorecard approach, where quantitative and qualitative factors are combined using appropriate weights. For stock exchange-listed companies, the module additionally applies an option pricing based model.
In addition to the quantitative ratios, qualitative factors are also taken into account. Furthermore, the analyst can enter warning signals or store information on guarantee/support structures.
In order to ensure a global scope of application, the module distinguishes different geographic segments. Transfer risk is taken into account for foreign companies.
The Corporates module in practice
The module has been routinely reviewed every year since 2005 and was revised in 2012. It has very high statistical accuracy due to a large database containing almost 260,000 ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
FUNDS
The ratings provided by the Funds module can be relevant for two types of exposures:
debt: loans to funds, repurchase agreements, derivative deals with funds, etc., i.e. exposures involving counterparty risk; credit institutions can use the ratings to calculate regulatory capital requirements
equity: these ratings can be used for internal risk management purposes.
Special characteristics
The module is based on a scorecard approach. It consists of the sub-modules “mutual funds” and “hedge funds”. The mutual funds sub-module covers equity funds, bond funds, money market funds, commodity funds and balanced funds. Some of the quantitative factors are calculated based on a complex portfolio model, which takes the country and asset allocation of a fund into account. In the hedge funds sub-module the fund strategy is an important factor. Different qualitative criteria which have a shown high discriminatory power are an integral part of the rating. Exceptional situations can be taken into account through overrides and warning signals.
The Funds module in practice
A large number of observations were used to develop the mutual funds sub-module. The hedge funds sub-module was developed from both own observations and a large external database. Based on these, it was possible to define very robust parameter values. The module has been routinely reviewed every year since 2009. In late 2009, it received supervisory approval for use under the IRBA.
INSURANCE COMPANIES
The Insurance Companies rating system is designed for rating companies that are commonly classified as insurers or generate more than 50% of their gross operating income from insurance business. All types of insurance companies and conglomerates focusing on insurance business can be assessed.
Special characteristics
The module is based on a scorecard approach. In addition to the quantitative and qualitative aspects, the ownership situation is taken into account. A distinction is made between the different accounting standards applied in the industry.
The module is divided into four sub-modules to reflect the heterogeneity of the insurance industry: life and health insurances form the “Life” segment, “Non-Life” covers property insurance and reinsurance, whereas “Reinsurance” is only used for this particular type of insurer. Very diversified insurance companies are assigned to the “Composite” segment.
Transfer risk is taken into account for foreign companies.
The Insurance Companies module in practice
The module has been routinely reviewed every year since 2006. It has very high statistical accuracy due to a large database containing more than 18,000 ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
IINTERNATIONAL COMMERCIAL REAL ESTATE
The International Commercial Real Estate module is designed for rating commercial real estate projects. These may relate to individual properties or to real estate portfolios.
Special characteristics
The module is divided into several sub-modules, each of which corresponds to a specific financing purpose (investor, developer, etc.). A further segmentation is based on the main use categories, namely office, retail, residential, industrial, and operator (especially hotels). Detailed aspects of the transaction can be entered, including different tranches and unscheduled repayment clauses.
The module is simulation-based. 10,000 scenarios are created to reflect the general economic development, using Monte Carlo simulations. In a next step, a number of regional property markets are analysed and included in the simulation. These differentiated scenario estimates form the basis for the simulation of the future cash flows. Debt service and cash flows are compared and the causal structure determining the probability of default is reproduced.
The International Commercial Real Estate module in practice
The module has been routinely reviewed every year since 2005. The database contains more than 27,000 ratings.
In early 2007, the module received supervisory approval for use under the IRBA. Credit institutions can use it to determine regulatory capital requirements for locations outside Germany.
INTERNATIONAL REGIONS AND MUNICIPALITIES
The module is used to assess the economic ability and willingness of an international local government body outside Germany to meet its payment obligations in full and punctually. The module covers regions and municipalities below the level of federal government (country), which perform public sector responsibilities for private households and companies within their administrative areas and can levy taxes and other charges.
Special characteristics
The model is based on a scorecard approach and uses information about the country’s political and economic environment in addition to qualitative and quantitative data. Particular risks are taken into account using a special warning system.
In order to be able to map the different local administrative units in an adequate manner, a difference is made between municipalities and regions based on the NUTS /LAU classification defined by Eurostat. Different rating algorithms are applied depending on the segment and, in addition, for all regions and municipalities transfer risk is included in the rating as an input.
The International Regions and Municipalities module in practice
The module has been constantly updated since 2005. It has a good statistical prediction accuracy thanks to a comparatively large database of almost 7,900 ratings due to pooling.
In early 2007, the German financial supervisory authority granted the approval to use the module for the IRB approach.
LEASING
Leasing Companies
The module is intended for rating leasing companies that apply German accounting standards (HGB). It performs a net asset value calculation to take the specific characteristics of these companies into account. The rating model is based on a scorecard approach.
SPC Real Estate Leasing
The module, which uses both scorecard and simulation elements, is designed for assessing real estate leasing projects. The residual value of the property is estimated by simulation. Transfer risk is included for offshore transactions.
The Leasing module in practice
The module has been routinely reviewed every year since 2005. It has high statistical accuracy due to a large database of more than 17,500 ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
LEVERAGED FINANCE
The module uses a scorecard-based approach and is suitable for rating traditional LBOs (leveraged buyouts), MBOs (management buyouts) and corporate-to-corporate transactions. It focuses on acquisitions involving substantial borrowing. The debt service is paid from the (future) cash flows of the target company.
Special characteristics
The rating model includes quantitative factors that are computed based on the cash flows of the target company as well as the equity and debt structure. Qualitative aspects, such as contractual arrangements, business plan, sector trend, etc. are also examined. Unusual circumstances can be taken into account through overrides and warning signals.
The module calculates the probability of default for the debt used for the acquisition. Both senior and subordinated debt are included in the calculation. The module is not limited geographically and can therefore be used for transactions worldwide
The Leveraged Finance module in practice
The Leveraged Finance module was developed based on a large number of real cases, which resulted in robust parameter settings. It has been reviewed annually since 2009, in which year it received supervisory approval for use under the IRBA.
PROJECT FINANCE
Usually, this module will be used to rate complex facilities such as power plants, industrial plants or infrastructure projects. The borrower is normally a special purpose vehicle. The module can be used universally and not only calculates the probability of default but also the loss given default and expected loss.
Special characteristics
Given the diversity of project finance transactions, nine main segments (e.g. transport, energy) and about 50 sub-segments (e.g. road and rail, power plants with PPA or Merchant arrangements) have been defined. Transactions can be specified in great detail. Transfer risk is taken into account for international projects.
The Project Finance module is a simulation-based rating system, which determines the expected cash flow of a financed project. For this purpose, 50,000 scenarios are generated using Monte Carlo simulations to describe future macroeconomic developments. These differentiated scenario estimates are taken as a basis for the simulation of the future cash flows. Segment-specific volatilities are included in the simulations. Debt service and future cash flows are compared. The result is used to model the causal structure determining the probability of default and the loss given default.
The Project Finance module in practice
The module has been routinely reviewed every year since 2005. It has high statistical accuracy due to a large database of more than 39,000 pooled ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
SHIP FINANCING
The module calculates the probability of default, loss given default and expected loss of companies that were specifically established to finance single ships or fleets. It can model the different types of ship finance worldwide. An ordinary ship mortgage loan can be entered, as can an additional construction finance agreement or an equity bridge loan.
Special characteristics
The module is segmented in six main and more than 30 sub-categories so that it can be used for all types of ships.
The model is based on Monte Carlo simulations which generate 10,000 scenarios on the general economic development. In a next step, the global ship market is analyzed and included in the simulation. These differentiated scenario estimates are taken as a basis for a projection of the future cash flows for the financed object. The causal structure for the probability of default and the recovery rate is modeled by comparing the debt service and the future cash flows.
The Ship Financing module in practice
The module has been constantly updated since 2005. It has a very high statistical prediction accuracy thanks to a very large database of more than 51,600 ratings due to pooling.
In early 2007, the German financial supervisory authority granted the approval to use the module for the IRB approach.
SOVEREIGNS AND TRANSFER RISK
The probability of default of sovereigns is calculated for obligations in both domestic and foreign currency. Empirically, defaults in local currency are much rarer than defaults in foreign currency, since they are, to a certain extent, influenced by different factors. For this reason the module includes two separate submodules to treat both probabilities.
In addition, the Sovereigns and Transfer Risk module calculates the risk that a foreign company may be prevented from meeting its payment obligations in foreign currency by government-imposed restrictions. This rating can be referenced by other modules.
Special characteristics
The module is based on a scorecard approach. In order to be able to model all sovereigns in an adequate manner, a segmentation is applied based on a country’s level of development. This classification is updated annually. Different models are used for each segment. In addition, extensive special rules are implemented for euro countries, off-shore countries, etc.
The Sovereigns and Transfer Risk module in practice
The module has been routinely reviewed every year since 2005. It has very high statistical accuracy due to a large database containing more than 10,500 ratings.
In early 2007, the module received supervisory approval for use under the IRBA.
Migration matrices and PD profiles
RSU provides one-year migration matrices and multiannual PD profiles for each of its twelve rating systems. Information about rating transitions is crucial for banks in various areas of risk management. Once IFRS 9 takes effect, modelling long-term rating migrations will become even more important.
As market leader for internal rating systems for wholesale banking, RSU has a large enough data pool at its disposal to be able to derive rating migrations from the internal models. As a result, the cyclical properties of the rating systems reflect in the migration matrices and PD profiles, which is essential for the institutions that use them.
A team of experts develops the models at the heart of LB-Rating
When developing rating systems, models for estimating probabilities of default and loss ratios are created based on historical information and solid expertise. However, empirically determined functional relationships may change or become less stable over time. For this reason, rating systems must, by law, be reviewed on a regular basis.
RSU’s methodology department validates the rating systems every year in consultation with the institutions that contribute to RSU’s data pool. Reviews are performed according to a defined validation policy using professional information and statistical computing technology.
LB Rating provides accurate estimates of probabilities of default
The rating indicates a probability of default (PD). The main measure of the quality of a rating system is what we refer to as “discriminatory power”, i. e. the system’s ability to distinguish between high-risk and low-risk obligors. Some rating systems also determine the Loss Given Default ratio (LGD), which is another area where accuracy is crucial.
Essentially, RSU’s methodological work focuses on the validation of the PD and LGD estimates computed. This involves, in particular, considering the defaults actually observed. To obtain accurate estimates it is important to identify distinctive rating criteria and incorporate them in the models.
Precise estimates and a broad range of rating systems due to a large data pool
The rating records are anonymised and used to test and further develop the systems. The resulting data pool is usually much larger than the amount of data available to a single institution, allowing more accurate analysis and estimates. It also includes sufficient data to cover special segments, which would pose a major statistical challenge to any single institution.
In addition to the advantages of pooled data in terms of methodological quality, clients benefit from economies of scale as they share the development and operating costs. At the same time, ratings are kept strictly confidential and data protection is ensured.
LB-Rating contains a total of 12 modules, covering a comprehensive spectrum of obligors. This spectrum is continuously enlarged and refined to ensure that each segment suits its purpose.
Easy internet-based access
LB-Rating is a completely web-based application, which can be accessed using Internet Explorer. Once it has been individually configured and activated, it requires no additional local installation.
Intelligent concepts and high-quality, security-focused technologies
The LB-Rating Framework contains all rating modules and provides general services and various interfaces for the connection of local systems. In particular, due to its modular structure, it can be expanded without difficulty to incorporate a client’s own rating systems. LB-Rating was implemented using a modern Java architecture (Java EE) as well as IBM products and thus complies with a very common and well established industry standard. As a result, it can be easily integrated into any client’s IT infrastructure and provides a high degree of overall reliability.
The application is based on a thin-client concept, i.e. all necessary data is provided by the server to the extent possible. Installation and maintenance services are carried out centrally on the server without affecting the user. Changes made on the server, e.g. new releases or security updates, take effect at the same time for all clients. Communication with the system is by secure and encrypted SSL data transfer through dedicated networks.
Quality-assured, error-avoiding processes
Methodological parameters such as score weights or calibration settings are stored separately and can be changed at short notice without modifying the software.
Technological modifications are performed twice a year at fixed dates. All related processes are based on ITIL, thus ensuring secure and high-quality IT workflows.
Changes made are completely transparent and verifiable
All changes made to LB-Rating and the related documentation are recorded centrally and can therefore be verified at a later date. This allows to address any functional, legal or technical issues in full detail.
Rating-Flex is the ideal solution for transferring existing rating systems to an audit-proof IT platform. It allows to incorporate a client’s own rating algorithms into LB-Rating.
Incorporated rating systems benefit from the complete functionality of LB-Rating.
The layout and handling are the same as for standard LB-Rating modules: users working with RSU modules and Ratings-Flex won’t notice any difference. Dedicated help functionalities including context-sensitive help texts can also be implemented.
As Rating-Flex systems are proprietary, their results are not included in RSU’s data pool and are not analysed by RSU. Each client is responsible for the quality of its own algorithms.
