Glossary

This page provides a definition of the key terms used within FRTB SA QIS. Additionally, see the Market Risk terminology provided in [MAR10].

BCBS

The Basel Committee on Banking Supervision (BCBS) is the primary global standard setter for the prudential regulation of banks and provides a forum for regular cooperation on banking supervisory matters.

Book

Groups of trades that follow a particular trading or banking strategy. Every trade is mapped to a single book. A book will always belong to a desk. Books can be nested into larger books (usually with a parent / child hierarchy). Books can span across legal entities. There can be tens of thousands of books in a large international bank.

Bucket MAR10.11

Buckets provide a means for grouping together risk positions by common characteristics. Buckets are prescribed for each risk class. For example:

• GIRR: Bucketing is performed using the currency of the instrument.
• Equities: Buckets are numbered from 1 to 11 and sensitivities are bucketed based on the type of equity (small cap/ large cap, sector)

Capital Charge

A capital charge (also known as regulatory capital or capital adequacy) is the amount of capital a bank or other financial institution has to hold as required by its financial regulator. This is usually expressed as a capital adequacy ratio of equity that must be held as a percentage of risk-weighted assets.

Compliance

FRTB SA QIS is designed and built to support the FRTB Standard Approach and Internal Models Approach as per the BCBS 457specification.

Correlation

Correlation is a measure of the degree to which two securities move in relation to one another.

Correlations are prescribed for each risk class and applied during intra-bucket aggregation (for example, for Risk Position), defined as Rho (ρ).

Inter-bucket aggregation is used to compute the Risk Charge and applies correlations Gamma ( γ).

Cube

OLAP on Big Data is a powerful concept that involves pre-aggregation of massive volumes of data into multidimensional cubes and then querying them to get faster results.

Within FRTB SA QIS, it is possible to have pre-defined cubes (for each of the FRTB SA QIS Components) with all Measures set in advance or to let users dynamically select Measures and instantiate them on-the-fly in the cubes (ActiveMeasures).

Curvature

Curvature risk captures the additional risk due to movement in the Delta when the price changes.

Data Model

An FRTB Data Model is provided to support the SA and IMA approaches. This is the default Data Model for all data required, and incorporates the following features:

• Trades with economic and organisational attributes

• Trades, Positions and Sensitivities per Risk Factor and other granular data for SA

• Trade Historical P&Ls, Historical Simulations and other granular data for IMA

• Classifications, Books, Desks, Risk Categories, etc.

• Correlations and other constants imposed by regulators

Datastore Schema

A Datastore Schema arranges data in stores that can reference each other. The schema contains three separate lists:

• Store descriptions (one for each store in the Schema)

• Reference descriptions that describe how stores reference each other

• Partitioning directives

Default Risk Charge

• This captures the ‘jump-to-default’ risk for credit instruments (for example Credit Default Swap - CDS).

• If the underlying issuer defaults, the buyer of the CDS will receive payment from the seller. The Default Risk Charge is another input into the overall Aggregated Capital Requirement (ACR) formula.

Delta

Delta is one of four major risk measures used by option traders. Delta measures the degree to which an option is exposed to shifts in the price of the underlying asset (i.e. stock) or commodity (i.e. futures contract). Values range from 1.0 to –1.0 (or 100 to –100, depending on the convention employed).

Desk [MAR12]

The concept of desk is a key element of the FRTB regulation. Banks need to identify them and have evidence to support their choice. Constraints on trader alignment with desks and management structures are quite strict. Desks are usually organised as a combination of geographical, business lines and traded instruments criteria. There may be hundreds of desks in a large international bank. FRTB stipulates that desks are classified as either trading or banking desks. A trader can only belong to one desk.

Dimension

A dimension is a named grouping of one or more hierarchies. Its purpose is mainly to provide customers with a means to categorise these separate hierarchies and regard them as a single entity, analogous to storing them together in a directory or folder.

Drill-down

This is the shifting of aggregation level from one member to a set of its child members in a hierarchy. For example, a value aggregated for the “USA” member in a Geographical hierarchy could be drilled-down to show the set of values aggregated at the State level.

Drill-through

A drill-through is a presentation of the set of data objects that underlie a particular aggregator location.

Epoch

An epoch marks the commit of a transaction on a Datastore timeline. Essentially, it is identified by a unique sequence number and a timestamp. An epoch can also receive a label if it needs to be precisely identified (e.g. “Revaluation Run 5 - 11AM”).

FRTB

The Fundamental Review of the Trading Book (FRTB) is a Basel Committee on Banking Supervision initiative to overhaul trading book capital rules.

Hidden Measure

A hidden measure is an aggregated measure that is used for calculation purposes, but should not be visible within a presentation

Hierarchy

A hierarchy provides a means for classifying data objects according to some property of these objects. For example, the Customer Location property of an Order entity could be classified on a Geographical hierarchy. A hierarchy allows classification at a number of different levels. (Note: For releases prior to ActivePivot 5, a ‘dimension’ was equivalent to what has now become a ‘hierarchy’. In ActivePivot 5, a dimension is defined as a set of hierarchies that have been grouped together and allotted a (dimension) name - see above).

IMA

The FRTB Internal Models Approach.

A capital charge calculated for a bank using the output of that bank’s internal risk measurement model.

Legal Entity

Banks’ activities are conducted through groups of legal entities or branches thereof. A trade will always involve one legal entity Desks often span across legal entities (i.e. global desks). There may be tens of legal entities in a large international bank group

Level

A Level in a hierarchy allows classification of a property at a particular level of abstraction.

MDX

MultiDimensional Expressions (MDX) is a query language for OLAP databases.

Measure

A measure is property of a data object that can be aggregated within a hypercube.

Member

A member of a level in a hierarchy is a particular attribute value that can be used to classify a property. For example, the Country level within a Geographical hierarchy would have members “UK”, “France”, “USA”, and so on.

Parameterisation

Parameterisation is the process of defining or choosing parameters

Query

A Datastore query retrieves entities from a Datastore that meets a specified set of conditions. A typical query includes the following: An entity kind to which the query applies. Optional filters based on the entities’ property values, keys, and ancestors.

Residual Risk Add-on

The SA rules state that an instrument with an underlying(that is not covered by Delta, Vega or Curvature) must have Residual Risk calculated. An example of an exotic underlying would be a weather derivative. In this case, there are no risk factors that measure/stress weather. The Residual Risk Add-on is the simple sum of gross notional amounts of the instruments bearing residual risks, multiplied by a risk weight of 1.0% for instruments with an exotic underlying and a risk weight of 0.1% for instruments bearing other residual risks.

Risk Charge(SBM only)

• Computed as an aggregation of risk positions (as defined below) across buckets within a risk class.

• This aggregation includes the application of the prescribed correlations.

• Final risk charge (Delta separated from Vega) is the aggregate across all buckets and risk classes. E.g. Delta GIRR (all vertices) plus Delta (Equity), etc.

Risk Classes

The risk classes used for FRTB SBM calculations are as follows:

• Commodity: Commodity risk

• (SA only) CSR Sec CTP: CSR Securitisation (Correlation trading portfolio). Includes securitisation of underlying asset. A correlation trading portfolio consists of securitisation positions and nth-to-default credit derivatives.

• (SA only) CSR Sec non-CTP: CSR Securitisation (non-Correlation trading portfolio). Includes securitisation of underlying assets.

• (IMA only) CSR: Credit spread risk

• Equity: Equity risk

• FX: Foreign exchange risk

• GIRR: General interest rate risk

Risk Factor

Each instrument or trade may be “sensitive” to multiple risk factors. For FRTB SBM, each risk factor is mapped to one of the prescribed risk classes defined above.

For FRTB SA QIS, we expect clients to send the sensitivities to underlying risk factors at the most granular level possible, i.e. at Trade level for OTCs and Position level for fungible instruments.

The native Atoti Server capabilities provides the required aggregation, netting and multi-dimension analysis.

Risk Measure

For the purposes of FRTB SA QIS (SA), a risk measure is one of the following:

• Delta: Based on the sensitivities of a bank’s trading book to regulatory Delta risk factors

• Vega: Based on the sensitivities of a bank’s trading book to regulatory Vega risk factors

• Curvature: This captures any incremental risk not captured by the Delta risk of an instrument with optionality. Curvature risk is based on two stress scenarios, involving an upward shock and a downward shock to a given risk factor. The worst loss of the two scenarios is the risk position to be used as an input into the aggregation formula which delivers the capital charge.

Risk Position

For Delta and Vega risks, it is a sensitivity to a risk factor. For Curvature risk, it is the worst loss of two stress scenarios.

Within the Solution, the sequence for calculating the risk position is as follows:

• A risk position is computed for Delta and Vega (within each risk class) by first netting sensitivities within risk class and buckets as prescribed [MAR21.4](1)

• Net sensitivities are then multiplied by a prescribed risk weight [MAR21.4](3 )

• Finally, the weighted sensitivities are aggregated between sensitivities within the same bucket, using a prescribed formula and correlations.

• For Curvature, risk positions are calculated by first computing a curvature risk charge based on the worst loss, after deducting delta risk position, from upward and downward shocks of each risk factor and applying prescribed risk weights.

• Curvature risk exposures are then aggregated within the same bucket, using prescribed formulae and correlations [MAR21.5].

Schema

A formal description of the structure of a database: the names of the tables, the names of the columns of each table, and the data type and other attributes of each column.

Solution

Atoti Solutions are projects that contain business logic, implementation best practices and software code to enable a faster time-to-market and help clients confidently address use cases such as regulations.

They are built on:

• Atoti Server
• ActiveMonitor (optional)
• Atoti UI (optional)

Vega

Vega is the measurement of an option’s price sensitivity to changes in the volatility of the underlying asset. Vega represents the amount that an option contract’s price changes in reaction to a 1% change in the implied volatility of the underlying asset.

Vertices

The term vertex refers to a tenor or expiry/maturity point along which a risk factor sensitivity is mapped or projected.

Examples of what the vertices represent:

• For GIRR Delta: points along a risk free yield curve.

• For CSR Delta: points along a credit spread curve.

• For Commodity Delta:time to maturity for a traded commodity. It is worth noting there are NO vertices for FX or Equity Delta risk classes.

• For Vega: option expiry dates.

In all cases, the vertices are prescribed. For sensitivities that are not exactly mapped, linear interpolation to prescribed points has been implemented.

Other interpolation methods may be implemented as needed.