Table of Contents
1 Review of Regulatory Definition
2 Math Expression of Regulatory Definition
3 Scenarios of MoC - observed, mis-identified defaults
3.1 Uncertainty in D1 - 1 year
3.2 Uncertainty in Di - multiple years
4 Scenarios of MoC - unobserved, defaults
4.1 Uncertainty in Di and Ni - 1 year
4.2 Uncertainty in Di - multiple years
5 Scenarios of MoC - general case
6 Numerical Example
7 MoC as a measure: impact of uncertainty in long-run DR
Research Objectives and Themes
This paper develops a quantitative framework to measure the Margin of Conservatism (MoC) for long-run average default rates, specifically addressing regulatory deficiency categories A, B, and D without relying on predefined distribution assumptions or arbitrary confidence intervals.
- Mathematical quantification of uncertainty based on relative data deficiencies.
- Integration of regulatory guidelines (EBA and ECB TRIM) into a consistent MoC model.
- Methodological extension to both observed and unobserved default scenarios.
- Demonstration of model application through empirical and fictitious portfolio examples.
- Establishment of a flexible, transparent, and intuitive framework for risk parameter adjustment.
Excerpt from the Book
3 Scenarios of MoC - observed, mis-identified defaults
Define Di and Ni as the number of defaulter and non-defaulters observed in the ith year with certainty, in other words, there is no data quality issue or other uncertainty. Hence the DRlong-run is the long-run average PD with no uncertainty in the underlying observations.
We further define the Margin of Conservatism for 1-year Default Rate (MoC1-year,i DR) and the Margin of Conservatism for long-run average Default Rate (MoClong-run DR) as following:
MoCi DR = DR~1-year i - DR1-year i (3)
MoClong-run DR = DR~long-run - DRlong-run (4)
3.1 Uncertainty in D1 - 1 year
Let d1 be the number of obligors that was observed in N1 whose default status was not correctly flagged due to mis-aligned counting of 90-Day delinquency period, or monthly snapshot windows, or reasons such as record mis-match between systems, etc. Therefore, the firm could have missed d1 defaults in the calculation of year-1 default rate.
The uncertainty adjusted 1-year default rate, DR~1-year i is:
DR~1-year i = Di + di / Ni (5)
The uncertainty adjusted long-run default rate, DR~long-run is:
DR~long-run = D1 / N1 + ... + Di+di / Ni + ... / Y , for i ∈ [1, Y] (6)
Summary of Chapters
1 Review of Regulatory Definition: Reviews the EBA and ECB requirements regarding the identification and classification of data deficiencies in PD and LGD estimation.
2 Math Expression of Regulatory Definition: Establishes the formal mathematical notation for one-year and long-run average default rates as the basis for further calculations.
3 Scenarios of MoC - observed, mis-identified defaults: Introduces the concept of Margin of Conservatism for scenarios where default events are observed but misidentified or miscounted.
4 Scenarios of MoC - unobserved, defaults: Extends the MoC methodology to scenarios involving unobserved defaults and incomplete observation periods.
5 Scenarios of MoC - general case: Provides a generalized mathematical proposition for calculating MoC based on relative uncertainty parameters for defaults and observations.
6 Numerical Example: Illustrates the practical application of the proposed formulas using a fictitious banking portfolio undergoing system migration.
7 MoC as a measure: impact of uncertainty in long-run DR: Analyzes the sensitivity of the MoC model to unit record changes and percentage changes in data uncertainty.
Keywords
Advanced IRB, Long-run Default Rate, Central Default Tendency, Risk Weighted Assets, RWA, Margin of Conservatism, MoC, PD Conservative Add-on, Data Deficiency, Regulatory Compliance, Default Probability, Quantitative Finance, Risk Management, Model Uncertainty, EBA Consultation Paper
Frequently Asked Questions
What is the core focus of this research paper?
The paper focuses on quantifying the Margin of Conservatism (MoC) for long-run default rates to address data deficiencies categorized by regulators, avoiding non-empirical distribution assumptions.
Which categories of deficiencies does the proposed MoC cover?
The model specifically addresses category A (estimation errors due to data), category B (diminished representativeness), and category D (other uncertainties).
What is the primary objective of this work?
The goal is to provide a robust, transparent, and empirically-based method to adjust risk parameters for uncertainty, ensuring compliance with EBA and ECB requirements.
Which scientific methodology is applied here?
The author uses a relative uncertainty framework that calculates adjustments based on countable deficiencies in records and observations, rather than relying on qualitative judgment or tail properties.
What topics are discussed in the main body?
The body covers mathematical definitions of default rates, scenarios for observed and unobserved default errors, a generalized proposition for relative uncertainty, and numerical examples.
Which keywords best characterize the paper?
Key terms include Advanced IRB, Margin of Conservatism (MoC), PD Conservative Add-on, Long-run Default Rate, and Regulatory Compliance.
How does this paper treat category C deficiencies?
Category C deficiencies, which are primarily model-related, are excluded from this specific approach and are noted to be addressed in a separate, dedicated paper.
What is the significance of the "Unit Record Change" analysis?
It allows risk managers to understand the precise impact on the MoC when a single data record is changed, providing a granular view of risk sensitivity.
- Quote paper
- Yang Liu (Author), 2018, Margin of Conservatism for Long-run Average Default Rate, Munich, GRIN Verlag, https://www.grin.com/document/443862
