Quanteam Research Papers

The Basel II Risk Parameters. Estimation, Validation, and Stress Testing

The estimation and validation of the Basel II risk parameters PD (default probability), LGD (loss given default), and EAD (exposure at default) is an important problem in banking practice. These parameters are used on the one hand as inputs to credit portfolio models, on the other to compute regulatory capital according to the new Basel rules. The book covers the state-of-the-art in designing and validating rating systems and default probability estimations. Furthermore, it presents techniques to estimate LGD and EAD. A chapter on stress testing of the Basel II risk parameters concludes the monograph.


Measuring the discriminative power of rating systems
Assessing the discriminative power of rating systems is an important question to banks and to regulators. In this article we analyze the Cumulative Accuracy Profile (CAP) and the Receiver Operating Characteristic (ROC), which are both commonly used in practice. We give a test-theoretic interpretation for the concavity of the CAP and the ROC curve and demonstrate how this observation can be used for more efficiently exploiting the informational contents of accounting ratios. Furthermore, we show that two popular summary statistics of these concepts, namely the Accuracy Ratio and the area under the ROC curve, contain the same information and we analyze the statistical properties of these measures. We show in detail how to identify accounting ratios with high discriminative power, how to calculate confidence intervals for the area below the ROC curve, and how to test if two rating models validated on the same data set are different. All concepts are illustrated by applications to real data.

An abridged version of the paper has been published under the title »Testing Rating Accuracy« in Risk, January 2003.

Download PDF

Empirical Comparison of Different Methods
for Default Probability Estimation
In the third Consultative Document (CP3) of the New Basel Capital Accord default probabilities are key input variables in the determination of regulatory capital. This raises the question how default probabilities should be estimated. We compare empirically several approaches of default probability estimation for small and medium size companies using about 10 years of data. It turns out that the inclusion of time series information into the estimation procedure improves the quality of default probability estimates considerably.

Download PDF

Quanteam Technical Solution Papers

ACE: Developing complex Client/Server-Architectures
with the comfort of Java
With ACE, the C++ Open-Source-Library, complex client/ server-architectures and shared applications can be easily and quickly developed. It provides C++ not only with a uniform class structure for over 20 platforms but also with extremely complex patterns in the manner of mini-frameworks. Although Java, due to the Internet boom, has already pushed through, it gains further acceptance thanks to the large developmental surrounding like Eclipse and the simple integration of diverse APIs over projects, which have been predominantly implemented in C++. We demonstrate in this paper how, by applying ACE, a large part of the complexity of Multi-Threading, IPC, network programming and asynchronous event-handling, i.e. essential »ingredients« of client/ server-solutions, can be capsulated and how the application developer can be freed from the basic developmental tasks.

Download PDF (available only in German)

Benchmarking C++ against Java for Numerical Derivatives Pricing
Numerical methods play an important role in modern investment banking especially in the field of derivatives pricing. The implementation of pricing algorithms is mostly done in C++ because of its high performance. However, the development of code in C++ has to be done very carefully since C++ is a rather complex language with a lot of potential pitfalls. An alternative would be the use of special classes that try to avoid those problems by hiding some individual aspects for potential errors, e.g. like safe pointers from the STL. Another alternative would be to switch to a safer programming language like Java. In this article we compare the performance of three different implementations of a simple derivatives pricing problem, the pricing of a European call option in the Black-Scholes framework using an implicit finite difference method. We compare those implementations, two in C++ and one in Java, with respect to computational time and memory requirement.

Download PDF    Download Sources

The OO pattern and a cluster application in harmonious concert
Patterns are something wonderful. They provide structure and promote reusable solution approaches for the continuously changing software development environment. However, the use of patterns is not necessarily obvious for every type of development. This article examines a case in which the classic »gang-of-four-patterns« was used in a cluster project.

Download PDF