SAS Documentation
SASĀ® Solution for Stress Testing
Reference manual - version 08.2021
Loading...
Searching...
No Matches
irmst_riskagg_hicopula.sas File Reference

hicopula aggregation of risk. More...

Go to the source code of this file.

Detailed Description

hicopula aggregation of risk.

Parameters
[in]ds_in_configuration_tableInput configuration table. It provides information about risk sources to be aggregated and information about the corresponding scenarios.
[in]ds_in_copula_tableInput configuration table of the COPULA functions. It provides information about the copula functions to be used the aggregation.
[in]ds_in_marginal_tableInput configuration table of the marginal distributions. It provides information about the marginal distributions of the hierarchy leaves.
[in]ds_in_corr_tablesList the correlation matrices that have been loaded. These matrices need to match the matrices required by the configuration table DS_IN_COPULA_TABLE.
[in]working_libnameWorking LIBNAME. By default, it is set to work. This option is typically used to facilitate debugging.
[in]in_alphaProbability tail level to evaluate the Value-at-Risk
[in]in_seedSeed
[in]in_ndrawsNumber of simulation draws to be sample from the COPULAs.
[in]high_agg_level_varHigher level of aggregation variable.
[in]low_agg_level_varLower level of aggregation variable.
[out]ds_out_risk_analysis_tableOutput table. It contains results about the hicopula aggregation..

Given an input configuration table (DS_IN_CONFIGURATION_TABLE), a COPULA configuration table (DS_IN_COPULA_TABLE) and marginal distribution configuration table (ds_in_marginal_table), this macro will aggregate the risk sources defined in the configuration table (DS_IN_CONFIGURATION_TABLE) using hierarchical COPULA aggregation. Particular attention is needed when specifying the COPULA configuration table. This table may require COPULA functions which depends on correlation matrices. In order to successfully run this aggregation, the matrices need to be loaded.

Notice that the aggregation can be performed with respect to different scenarios: the configuration table (DS_IN_CONFIGURATION_TABLE) and the (DS_IN_CORRELATION_TABLE) need to contain a properly specified scenario variable (<SCENARIO_ID>).

The structure of the DS_IN_CONFIGURATION_TABLE table is as follows:

Variable Type Description
<SCENARIO_ID> CHARACTER/NUMERIC Scenario identifier variable.
<NAME> CHARACTER Name of the risk source.
<RISK_MEASURE> CHARACTER Type of risk measure: VaR, EC, etc....
<SUBDIVISION< CHARACTER It identifies the business subdivision.
<RISK_SOURCE< CHARACTER It identifies the risk source.
<COUPULA< CHARACTER It specifies the COPULA function.
<MARGINAL< CHARACTER It specifiess the marginal distribution.
<VAR< CHARACTER It specifies the variable on which the Risk Source Measure is measured.
<GROUP< CHARACTER It identifies the group for the COPULA aggregation.
<REFNAME< CHARACTER It identifies the reference name for COPULA aggregation.

The structure of the DS_IN_COPULA_TABLE table is as follows:

Variable Type Description
<SCENARIO_ID; CHARACTER It identifies the scenario.
<RISK_MEASURE; CHARACTER It identifies the type of risk measurement (VaR, EC, etc..).
<NAME; CHARACTER It identifies the copula function.
<PARAMETER; CHARACTER Parameter type: type, cov, corr, theta, df, ..
<CVALUE; CHARACTER Character value for parameter "type".
<NVALUE; NUMERIC Numeric value for parameters "args" and "parm".

It is important to notice that when the <PARAMETER; is either "cov" or "corr", the corresponding value for <CVALUE; contains the name of the correlation matrix needed for the copula aggregation. The structure of these correlation matrix table is as follows:

Variable Type Description
<SCENARIO_ID; CHARACTER It identifies the scenario.
<RISK_MEASURE; CHARACTER It identifies the type of risk measurement (VaR, EC, etc..).
<NAME; CHARACTER It identifies the risk source .
<TYPE; CHARACTER type of dependence matrix: corr
<name 1; NUMERIC Risk for Name 1
<...; NUMERIC Risk for Name 2
<name k; NUMERIC Risk for Name 3

The structure of the DS_IN_MARGINAL_TABLE table is as follows:

Variable Type Description
<SCENARIO_ID; CHARACTER It identifies the scenario.
<RISK_MEASURE; CHARACTER It identifies the type of risk measurement (VaR, EC, etc..).
<NAME; CHARACTER It identifies the marginal distribution.
<PARAMETER; CHARACTER Parameter type: type, args, parm.
<CVALUE; CHARACTER Character value for parameter "type".
<NVALUE; NUMERIC Numeric value for parameters "args" and "parm".

Author
SAS Institute INC.
Date
2020

Definition in file irmst_riskagg_hicopula.sas.