arXiv subjects

Ferenc Illés

Publications and source records attributed to Ferenc Illés.

3 recordsLinked to original sources

arXiv2025

Through-the-Cycle PD Estimation Under Incomplete Data -- A Single Risk Factor Approach

Banks are required to use long-term default probabilities (PDs) of their portfolios when calculating credit risk capital under internal ratings-based (IRB) models. However, the calibration models and historical data typically reflect prevailing market conditions. According to Basel recommendations, averaging annual PDs over a full economic cycle should yield the long-term PD. In practice, the available data are often temporally incomplete - even for high-risk portfolios. In this paper, we present a method for the simultaneous calibration of long-term PDs across all sub-portfolios, based on the single risk factor model embedded in the Basel framework. The method is suitable even for smaller, budget-constrained institutions, as it relies exclusively on the bank's own default data. A complete dataset is not required - not even for any individual sub-portfolio - as the only prerequisite is the presence of overlapping data before and after the missing values, a mild condition that is typically met in practical situations.

q-fin.RM↗

arXiv2024

Linearly representable games and pseudo-polynomial calculation of the Shapley value

We introduce the notion of linearly representable games. Broadly speaking, these are TU games that can be described by as many parameters as the number of players, like weighted voting games, airport games, or bankruptcy games. We show that the Shapley value calculation is pseudo-polynomial for linearly representable games. This is a generalization of many classical and recent results in the literature. Our method naturally turns into a strictly polynomial algorithm when the parameters are polynomial in the number of players.

cs.GT↗

arXiv2022

Estimation of the Shapley value by ergodic sampling

The idea of approximating the Shapley value of an n-person game by Monte Carlo simulation was first suggested by Mann and Shapley (1960) and they also introduced four different heuristical methods to reduce the estimation error. Since 1960, several statistical methods have been developed to reduce the standard deviation of the estimate. In this paper, we develop an algorithm that uses a pair of negatively correlated samples to reduce the variance of the estimate. Although the observations generated are not independent, the sample is ergodic (obeys the strong law of large numbers), hence the name "ergodic sampling". Unlike Shapley and Mann, we do not use heuristics, the algorithm uses a small sample to learn the best ergodic transformation for a given game. We illustrate the algorithm on eight games with different characteristics to test the performance and understand how the proposed algorithm works. The experiments show that this method has at least as low variance as an independent sample, and in five test games, it significantly improves the quality of the estimation, up to 75 percent.

cs.GT↗