Migration notes 3.1

This page explains the changes required to migrate to the stated version of the FRTB Accelerator.

Migrate to 3.1

Upgrading from version 3.0 - see FRTB Accelerator 3.1 Release Notes.

Breaking Changes

• In the DRC Trade Level file, the obligor and tranche fields have been replaced and will need to be updated in existing CSV files will need to be updated in order to be loaded into 3.1. See below for details.

• The new Equity Bucket Descriptions file is required, and the new EquityBucketDesc store must be populated, for the Equity calculations. As a consequence of the direct bucketing changes, the market cap and economy are now determined from this store.

• The direct bucketing changes impose a new order on file loading.

  The SA bucket files must now be loaded before (or at the same time as) the sensitivities files.

SA DRC Data Model Changes

• In the DRC Trade Level file:
  • The obligor and tranche columns are combined into a single Underlying column.
  • An optional risk-factor column has been added. If not provided, it will be generated from the Underlying and maturity date (and seniority for DRC non-Sec).
• In the datastore:
  • The DRCBase store includes all the fields needed to calculate the Gross JTD and a reference to the risk-factor.
  • The RiskFactorDescription store now includes the maturity, seniority (DRC non-Sec), and a reference to either the obligor (DRC non-Sec) or tranche (DRC Sec non-CTP).
  • The Obligor and Tranche stores continue to describe the obligors and tranches.

RRAO Data Model Changes

• In the Trade Attributes file, the optional RRAO category field has been added.
  • If not provided, it will be populated in the ETL. However, it will not be populated in a way that is amenable to changing per jurisdiction.
• In the ETL, a fact is added to the cube (a row in the TradeBase store) if:
  • The RRAO field is “Y” (unchanged), OR
  • The RRAO category is provided (new)
• In the datastore, the RRAO store has been added. Indexed by category, it contains the RRAO description (flags and type).
  • These fields have been moved out of the trade attributes store.
• In the calculations, a filter has been added to require the RRAO flag to be “Y” (previously this was implicit because facts were only added when the flag was Y).
• In the SA cube, the RRAO flag hierarchy has been added.

Direct Bucket Mappings

• In the CSV files:
  • An optional bucket column has been added to the Delta, Vega, and Curvature sensitivities files and to the DRC Trade Level files.
    • If not provided, it will be determined from the underlying description fields.
  • Some of the underlying descriptions fields are now optional and can be populated from the bucket.
  • Note: Either the bucket or the underlying description fields must be provided.
  • Note: A new Equity Bucket Description file has been added to populate the underlying description fields based on the equity bucket.
• In the ETL:
  • The bucket mapping and description files must be loaded before the sensitivities files.
    • If they are loaded in the same command (e.g. using the “ALLSA” topic alias), then the DLC will load bucket files first.
    • By default, the bucket files are loaded at the same time as the sensitivities files (using the “ALLSA” topic alias). This is unchanged from pervious versions.
  • If the bucket column is not populated, it will be looked up from the underlying description fields using the bucket mappings.
  • If the underlying descriptions fields are not populated, they will be looked up from the bucket using the bucket descriptions.
• In the datastore:
  • The stores mapping the underlying descriptions to buckets are no longer part of the cube schema (they are now “isolated” stores). They are used in the ETL, but do not become part of the cube.
  • The bucket column has been added to the UnderlyingDescription store.
  • The EquityBucketDesc store has been added.
  • The mapping stores for FX and GIRR buckets have been removed.

Support for Subsidiaries

The netting sets and imported values functionality directly affect the following measures, which are the ones requested in the CRR2 reporting template.

• “risk-class risk-measure Risk Charge scenario”
• “risk-class Delta Sensitivities direction”
• “drc-risk-class Default Risk Charge”
• “drc-risk-class Gross JTD direction”
• “RRAO Exotic”
• “RRAO Other”

Where:

• risk-class is one of: GIRR, CSR non-Sec, CSR Sec non-CTP, CSR Sec CTP, Equity, Commodity, FX
• risk-measure is one of: Delta, Vega, Curvature
• scenario is one of: blank (for medium correlation scenario), High, Low
• direction is one of: Long, Short
• drc-risk-class is one of: DRC non-Sec, DRC Sec non-CTP

Additionally, the SA measures that follow on from the above list are also impacted, up to the “Portfolio Risk Charge”.

A new measure variation with the suffix “(reported)” has been added that includes the impact of the netting sets and imported values. For example, “Portfolio Risk Charge (reported)”. The netting sets and imported values do not currently impact the unmodified measures; for example, “Portfolio Risk Charge” is not changed.

Netting Sets

A new measure variation has been added with the suffix “(netted)”. The netting sets only impact the measures with this suffix (and with the “(reported)” suffix).

• The new (optional) Legal Entity Attributes file (and corresponding LegalEntityAttributes store) maps legal entities to netting sets.
• A new Netting Set hierarchy allows for filtering by netting set.
• The capital charge calculations are performed with the netting sets as the leaf level, so that:
  • The capital charges are calculated without netting sensitivities across netting sets.
  • After the capital charges are calculated (independently) for each netting set, they are summed.

As with the “(reported)” suffix, the same set of measures are directly and indirectly impacted.

Imported Values

A new measure variation has been added with the suffix “(imported)”. The imported values only impacts measures with this suffix (and with the “(reported)” suffix).

• The new (optional) Legal Entity Imports file (and corresponding LegalEntityImports store) contains the imported values.
• The imported values are identified by legal entity, parameter set, and measure name (and as-of date).
• The measure name must be one from the list above.

Static Data Overrides

• New files and stores have been added that parallel the stores for the underlying, obligor, tranche, and RRAO descriptions.
• The new files and stores have the same contents as the existing stores, plus a “parameter set” column.
• During the calculations, if the parameter set is not “BCBS”, then these stores are queried for potential overrides.
  • If an override is found, then the sensitivities are shifted to the new location in the cube (corresponding to the overridden data).
  • The override lookup queries will walk up the parameter set tree structure looking for overrides. For example, if you have added a variant of the CRR2 parameter set, then the CRR2 overrides will still be used for this variant.

Performance note: When drilling down on the calculations, additional data may be loaded and subsequently discarded. This impacts drilling down into levels that may be overridden. For example, if you add the filter rating=AAA to your query, the data and overrides will still be loaded for all ratings. Then, after the overrides have been applied, the non-AAA ratings will be discarded. This means that applying the filter rating=AAA will not necessarily be any faster than if the filter was not applied.

Spring Bean Configuration

Some of the functionality for customizing the Datastore Helper and for adding copper measures to cubes can now be registered using Spring beans.

Datastore Configuration

Beans of type Consumer<IDatastoreConfigurator> will automatically be called to provide extra datastore configuration. This is an alternative to adding methods to DatastoreCustomisationConfig.

For example:

@Bean
public Consumer<IDatastoreConfigurator> myDatastoreConfig() {
    return configurator -> {
        // Code for additional datastore configuration here
    };
}

Copper Configuration

Beans of type Consumer<ICopperContext> will automatically be called to provide additional copper configuration to a cube (based on qualifier). They will be called during the withCalculations phase of the cube description builder.

For example, to add some copper configuration to both the SA and IMA cubes:

@Qualifier("saCopperConfig")
@Bean
public Consumer<ICopperContext> saMyCopperConfig() {
    return this::myCopperConfig;
}

@Qualifier("imaCopperConfig")
@Bean
public Consumer<ICopperContext> imaMyCopperConfig() {
    return this::myCopperConfig;
}

protected void myCopperConfig(ICopperContext context) {
    // Code for additional copper configuration
}

Input file formats

New Files

Modified

ETL Changes

The new Bucket column in the Delta, Vega, and Curvature files is optional. If not provided it will be populated using the previously loaded bucket files.

Additionally, a lot of the fields in the SBM underlying description have become optional, as long as the bucket is specified. For CSR and Equities, these fields can be populated from previously loaded bucket description files.

These bucket description files are now also used to populate these fields when loading CRIF files.

Configuration files

Modified

frtb.properties

frtb-config.properties

frtb-data-load.properties

Datastores

Added

Modified

Removed

Cube schema

Added

Measures

Added

Drillthrough Columns

The following drillthrough columns have been added to properly display the vectors used for the SA sensitivities

Context values

No change.

DLC Topic Aliases

The following topics have been added to aliases:

Additionally, the “SAOverrides” alias has been added to the “ALLSA” alias.