Hostname: page-component-7b7567c8c-2jgqx Total loading time: 0 Render date: 2026-11-19T20:19:44.397Z Has data issue: false hasContentIssue false

A scaling model for severity of operational losses using generalized additive models for location scale and shape (GAMLSS)

Published online by Cambridge University Press:  30 October 2012

Amandha Ganegoda*
Affiliation:
School of Actuarial Studies, Australian School of Business, Sydney, Australia
John Evans
Affiliation:
School of Actuarial Studies, Australian School of Business, Sydney, Australia
*
*Correspondence to: Amandha Ganegoda, School of Actuarial Studies, Australian School of Business, UNSW, Sydney NSW 2052, Australia. E-mail: amandha@unswalumni.com
Rights & Permissions [Opens in a new window]

Abstract

In this paper, we investigate the problem of how to combine operational losses collected from various banks of different sizes and loss reporting thresholds in order to estimate the distribution of operational loss severities for a bank of a given size. We model the severity of operational losses by using the extreme value theory to account for the reporting bias of the external data, and a regression analysis based on the GAMLSS framework to model the scaling properties of operational losses. In contrast to previous studies on the scaling problem, our analysis gives particular emphasis to the scaling properties of the tail of the loss distribution. Contrary to existing knowledge, we find that the size of a bank is an important determinant of the severity of operational losses and that the tail index of the distribution is negatively correlated with the size of the bank. The results indicate that for very large banks, distribution of the operational loss severity can be extremely heavy tailed (i.e. tail index less than 1), a finding which have significant implications for capital calculation as well as for risk management. Furthermore, we also demonstrate that the capital estimates provided by our model is consistent with the industry standards and the model can be used by individual banks to simulate data to complement their internal data.

Information

Type
Papers
Copyright
Copyright © Institute and Faculty of Actuaries 2012
Figure 0

Table 1Mean and the 99.9th percentile loss for the hypothetical Bank I and Bank II.

Figure 1

Figure 1The distribution of total Losses for two risk types with same mean.

Figure 2

Figure 2The density of two different Pareto risks.

Figure 3

Table 2The Distribution of the Selected Sample by Total Assets.

Figure 4

Table 3The Summary Statistics by Business Line.

Figure 5

Table 4The Summary Statistics by Event Type.

Figure 6

Figure 3Mean Excess Plot of Operational Losses.

Figure 7

Table 5Estimation Results b, τ, and a for Various Thresholds.

Figure 8

Table 6The List of Fitted Distributions.

Figure 9

Table 7The Competing Base Models.

Figure 10

Figure 4The Step-Wise Selection Strategy of Explanatory Variables.

Figure 11

Table 8Estimated Parameter Values for Model G.

Figure 12

Figure 5Normalized Quantile Residual Plot of the Model G.

Figure 13

Table 9The Robustness Analysis.

Figure 14

Figure 6QQ Plot of Simulated Losses vs. Observed Losses for US Banking Industry. Note: quantiles for the observed losses have been computed using the weights estimated in subsection 4.2.

Figure 15

Table 10Summary Statistics of the Simulation for the US Banking Industry.

Figure 16

Figure 7QQ plot of Simulated Losses vs. Observed Losses for Bank A. Note: quantiles for the observed losses have been computed using the weights estimated in subsection 4.2.

Figure 17

Figure 8QQ plot of Simulated Losses vs. Observed Losses for Bank B. Note: quantiles for the observed losses have been computed using the weights estimated in subsection 4.2.

Figure 18

Table 11Summary Statistics of the Simulation for Bank A and B.

Figure 19

Table 12Re-estimated Parameter Values of Model G after Removing Bank A.

Figure 20

Table 13Re-estimated Parameter Values of Model G after Removing Bank B.

Figure 21

Table 14Estimated Average Loss ($ mill).

Figure 22

Table 15Estimated Loss at the 99th Percentile ($ mill).

Figure 23

Table 16The increase in the 99th percentile loss when the size of the bank doubles.

Figure 24

Table 17Estimated Tail Indices.

Figure 25

Figure 9The Lorenz curves of the loss distributions at a bank with 300 bill assets. Note: an equity ratio 6.7% has been assumed to compute the Lorenz curve.

Figure 26

Figure 10The Lorenz curves of the loss distributions at a bank with 600 bill assets. Note: an equity ratio 6.7% has been assumed to compute the Lorenz curve.

Figure 27

Table 18Cumulative % of individual losses as the % of total losses.

Figure 28

Table 19Cumulative % of individual losses as the % of total losses.

Figure 29

Table 20Estimated Average Loss ($ mill).

Figure 30

Table 21Estimated Loss at the 99th Percentile ($ mill).

Figure 31

Table 22Estimates of ${\rm{\hat{\lambda }}}j$ for a bank with $100 bill assets.

Figure 32

Table 23Capital Estimates (in $ mill) for Hypothetical Banks.

Figure 33

Table 24Capital Estimates (in $ mill) for Hypothetical Banks after implementing a 1-in-10,000 year probability limit.

Figure 34

Figure 11Backtesting Results for Bank A.

Figure 35

Figure 12Backtesting Results for Bank B.