Coder Social home page Coder Social logo

jmptrader / opensimm Goto Github PK

View Code? Open in Web Editor NEW

This project forked from opengamma/opensimm

0.0 0.0 0.0 369 KB

Reference implementation of the ISDA-proposed Standard Initial Margin Model (SIMM) for non-cleared derivatives

Home Page: http://www.opengamma.com/opengamma-platform/opensimm

License: Apache License 2.0

Java 100.00%

opensimm's Introduction

OpenGamma

OpenSIMM by OpenGamma

Build Status License

This repository contains a standalone reference implementation of the ISDA-proposed Standard Initial Margin Model (SIMM) for non-cleared derivatives. The aim of OpenSIMM is to support standardization and adoption of the methodology which will be used industry-wide as the basis for the exchange of bilateral initial margin.

As the SIMM model has yet to be finalized, OpenGamma will continue to work with the industry on the standard as it evolves. We have implemented the initial HVaR-based approach in this release, but we are working with dealers on implementing the current version of the model, which uses a different methodology.

OpenSIMM is released as open source software under the Apache v2.0 license.

For more information please refer to our website, or contact us at [email protected]

If you would like the full implementation whitepaper, please contact us at [email protected]

Releases

This repository contains version 1.0 as available in Maven Central. OpenSIMM requires Java SE 8 or later.

<dependency>
    <groupId>com.opengamma</groupId>
    <artifactId>opensimm</artifactId>
    <version>1.0</version>
</dependency>

Building OpenSIMM

Prerequisites:

The source code can be cloned using git from GitHub:

git clone https://github.com/OpenGamma/OpenSIMM.git

Simply run this command to compile and install the source code locally:

mvn install

The project contains only one external dependency - TestNG which is required for running the unit tests. Apart from this, the project is entirely stand-alone.

Running OpenSIMM

The calculation can be run from the command line.

First run:

mvn package

This will build a jar file: opensimm-1.0.jar

The jar file can then be executed:

java -jar target/opensimm-1.0.jar src/test/resources/simm-sample/simm.properties

which should give the output similar to the following:

----------- ----------
Asset Class Var
----------- ----------
CO            564.3703
IR            447.5351
CR             33.3300
EQ            740.7143
----------- ----------
Total       1,785.9496
----------- ----------

SIMM Calculator

The primary class in the system is the SimmCalculator class. The following sections describe how to calculate a VaR measure for asset classes using SIMM.

The calculator can either be used with raw data provided programatically, or by loading csv files.

Data Types

In order to perform the calculation a number of different types of data need to be supplied:

  • Risk Factors - risk factors describe categories that a portfolio can be split into. A risk factor has:

    • a unique name,
    • an asset class - one of IR (Interest Rate), CR (Credit), EQ (Equity) or CO (Commodity)
    • the risk type - one of SENSITIVITY or EXPOSURE
    • the shock type - either AB (Absolute) or RE (Relative). If relative then a shift can also be supplied.

  • Risk Factor Base Levels - the base levels describe the starting value for each risk factor

  • Risk Factor Shocks - the shocks are a set of perturbations to be applied to the base levels. These are then used in performing the VaR calculation.

  • Base Currency - the currency that all output will be in

  • FX Rates - the set of current FX rates for all currencies in the portfolio

  • FX Rate Shocks - the set of shocks to be applied to the FX Rates

  • Derivative Portfolio - the portfolio to be evaluated broken down into exposures against each Risk Factor.

  • Initial Margin Portfolio - any initial margin broken down into exposures against each Risk Factor.

  • Variation Margin Portfolio - any variation margin broken down into exposures against each Risk Factor.

Executing Programmatically

To execute the calculation (for instance in a test):

// Create risk factor names
RiskFactor IBM = StandardRiskFactor.of("IBM");
RiskFactor USD_IRSL3M_2Y = StandardRiskFactor.of("USD-IRSL3M-2Y");

// create risk factor properties and put into map
Map<RiskFactor, RiskFactorProperties> riskFactorProperties = new HashMap<>();
riskFactorProperties.put(USD_IRSL3M_2Y, RiskFactorProperties.relativeShock(IR, SENSITIVITY, 0.04));
riskFactorProperties.put(IBM, RiskFactorProperties.absoluteShock(CR, SENSITIVITY));

// create risk factor base levels
Map<RiskFactor, Double> baseLevels = new HashMap<>();
baseLevels.put(USD_IRSL3M_2Y, 0.01);
baseLevels.put(IBM, 0.0120);

// create fx matrix for all currencies to be used
FxMatrix fxMatrix = FxMatrix.builder()
    .addRate(EUR, USD, 1.40)
    .addRate(GBP, USD, 1.60)
    .build();

// create risk factor shocks
Map<RiskFactor, List<Double>> riskFactorShocks = new HashMap<>();
riskFactorShocks.put(USD_IRSL3M_2Y,
    // Normally many, many more
    Arrays.asList(1.0025, 1.0025, 0.9975, 0.9975, 1.0000, 1.0000, 1.0000, 1.0025));
riskFactorShocks.put(IBM, Arrays.asList(0.0001, -0.0005, -0.0050, 0.0002, -0.0006, -0.0051));

// create FX shock
Map<Pair<Currency, Currency>, List<Double>> fxShocks = new HashMap<>();
// Shock length must match that for IR risk factors - USD_IRSL3M_2Y in this case
fxShocks.put(Pair.of(EUR, USD),
    Arrays.asList(1.0000, 1.0010, 0.9975, 0.9950, 1.0002, 1.0100, 0.9950, 0.9970));
fxShocks.put(Pair.of(GBP, USD),
    Arrays.asList(0.9985, 1.0020, 1.0000, 0.9950, 1.0003, 1.0100, 0.9940, 0.9960));

// Now setup the calculator with all the data we have so far
SimmCalculator calculator = SimmCalculator.builder()
    .varLevel(0.9)
    .baseCurrency(EUR)
    .riskFactors(riskFactorProperties)
    .riskFactorLevels(baseLevels)
    .fxMatrix(fxMatrix)
    .riskFactorShocks(riskFactorShocks)
    .fxShocks(fxShocks)
    .build();

// Build the portfolio to be assessed
List<PortfolioExposure> portfolio = Arrays.asList(
    PortfolioExposure.of(USD_IRSL3M_2Y, 250_000, USD),
    PortfolioExposure.of(IBM, 100_000, GBP),
    // Currency positions can be held too
    RawDelta.of(FxRiskFactor.of(GBP), 125_000, GBP));

// It is also possible to provide current initial and variation
// margining positions in the same format - see version of
// calculator.varByAssetClass taking three arguments

// Calculate the values for the portfolio
Map<AssetClass, Double> result = calculator.varByAssetClass(portfolio);

Executing from data files

It is possible to load the required data from csv files rather than programmatically creating the data as above. A set of loaders are provided to facilitate this.

// getFile() method just returns a valid java.util.File,
// its implementation is unimportant
Map<RiskFactor, RiskFactorProperties> riskFactorProperties =
    RiskFactorDefinitionsLoader.of(getFile("risk-factor-definitions")).load();

Map<RiskFactor, Double> riskFactorLevels =
    RiskFactorBaseLevelsLoader.of(getFile("risk-factor-base-levels")).load();

FxMatrix fxMatrix = FxRateLoader.of(getFile("fx-rates")).load();

Map<RiskFactor, List<Double>> riskFactorShocks =
    RiskFactorShocksLoader.of(getFile("risk-factor-shocks")).load();

Map<Pair<Currency, Currency>, List<Double>> fxShocks =
    FxShocksLoader.of(getFile("fx-rate-shocks")).load();

SimmCalculator calculator = SimmCalculator.builder()
    .varLevel(0.9)
    .baseCurrency(EUR)
    .riskFactors(riskFactorProperties)
    .riskFactorLevels(riskFactorLevels)
    .fxMatrix(fxMatrix)
    .riskFactorShocks(riskFactorShocks)
    .fxShocks(fxShocks)
    .build();

Set<RiskFactor> riskFactors = riskFactors.keySet();
List<RawDelta> derivativesPortfolio =
    PortfolioLoader.of(getFile("portfolio-derivatives"), riskFactors).load();
List<RawDelta> initialMargin =
    PortfolioLoader.of(getFile("portfolio-initial-margin"), riskFactors).load();
List<RawDelta> variationMargin =
    PortfolioLoader.of(getFile("portfolio-variation-margin"), riskFactors).load();

Map<AssetClass, Double> var =
    calculator.varByAssetClass(derivativesPortfolio, initialMargin, variationMargin);

File Formats

The file formats are straightforward - just simple CSV files. There are examples in test/resources/simm-sample.

Risk Factor Definitions

Contains the definitions of the fundamentals of the risk factors.

RiskFactorName, AssetClass, RiskType,    ShockType, Shift
USD-IRSL3M-2Y,  IR,         SENSITIVITY, RE,        0.04
IBM,            CR,         SENSITIVITY, AB
SP500,          EQ,         EXPOSURE,    RE
  • RiskFactorName - the risk factor name, can be any String but must be unique in the file
  • AssetClass - the asset class the risk factor belongs to - one of IR, CR, EQ or CO
  • RiskType - the risk type of the risk factor - one of SENSITIVITY or EXPOSURE
  • ShockType - the type of shocks - one of RELATIVE or ABSOLUTE
  • Shift - if the ShockType was RELATIVE, a shift can be supplied here

Risk Factor Base Levels

Contains the initial levels for the risk factors.

RiskFactorName, BaseLevel
USD-IRSL3M-2Y,  0.01
IBM,            0.012
SP500,          1000
  • RiskFactorName - the risk factor name, must be defined in the Risk Factor Definitions
  • BaseLevel - the initial level for the risk factor

Risk Factor Shocks

Contains the shocks to be applied to the risk factor levels.

RiskFactorName, Shocks
USD-IRSL3M-2Y, 1.0025, 1.0025, 0.9975, 0.9975, 1, 1, 1.0025, 0.9975, 1, 1, 1, 1, 1, 1, 1.0025, 0.9975, 1, 1, 1, 1, 1.0025, 0.9975, 1, 1, 1
IBM, 0.0001, -0.0005, -0.005, 0.0002, -0.0006, -0.0051, 0.0003, -0.0007, -0.0052, 0, 0.0004, -0.0001, 0.0011, 0.0008, 0.0012
SP500, 1.011, 0.995, 0.997, 1.001, 1.0002, 1.01, 0.9955, 0.9975, 1.0012, 1.0003, 1.0101, 0.9951
  • RiskFactorName - the risk factor name, must be defined in the Risk Factor Definitions
  • Shocks - list of shocks to be applied to the base levels. All shocks for a particular asset class must be the same length. Additionally, the shocks for IR asset class must be the same length as the FX shocks in the FX Shocks file.

FX Rates

Contains the base FX rates.

CurrencyPair, Rate
EUR/USD,      1.4
GBP/USD,      1.6
  • CurrencyPair - the currency pair the rate is for, format is CCY1/CCY2
  • Rate - the base FX rate for the currency pair - the EUR/USD line in the sample above is read as "1 EUR = 1.4 USD"

FX Shocks

Contains the shocks to be applied to the FX rates.

CurrencyPair, Shocks
EUR/USD, 1, 1.001, 0.9975, 0.995, 1.0002, 1.01, 0.995, 0.997, 1.001, 1.0002, 1.01, 0.995, 0.997, 1.001, 1.0002, 1.01, 0.995, 0.997, 1.001, 1.0002, 1.01, 0.995, 0.997, 1.001, 1.0002
GBP/USD, 0.9985, 1.002, 1, 0.995, 1.0003, 1.01, 0.994, 0.996, 1.001, 1.0002, 1.01, 1.005, 0.9999, 0.9999, 1.0011, 1.005, 0.996, 0.998, 1.0025, 1.01, 0.995, 0.997, 1.001, 1.0002, 1.0001
  • CurrencyPair - the currency pair the shocks are for, must be defined in the FX Rates file.
  • Shocks - list of shocks to be applied to the base FX rates. All shocks must be the same length, and must be the same length as the shocks for the IR asset class in the Risk Factor Shocks file.

Portfolio File

Contains the exposures of a portfolio to the various risk factors. The same file format is used for the portfolio, plus any offsetting initial or variation margin files.

RiskFactorName, Amount,   Currency
EUR-OIS-2Y,     100000,   EUR
EUR-OIS-5Y,     -20000,   EUR
USD-IRSL3M-2Y,  20000,    EUR
IBM,            30300,    EUR
SP500,          100000,   USD
  • RiskFactorName - the risk factor name, must be defined in the Risk Factor Definitions
  • Amount - the amount of the exposure to the risk factor
  • Currency - the currency for the Amount. Must be defined in the FX Rates file.

Properties File

It is possible to define a properties file holding the details of the data files, which can simplify the loading process.

base-currency=EUR
var-level=0.9
risk-factor-definitions=src/test/resources/simm-sample/risk-factor-definitions.csv
risk-factor-base-levels=src/test/resources/simm-sample/risk-factor-base-levels.csv
risk-factor-shocks=src/test/resources/simm-sample/risk-factor-shocks.csv
fx-rates=src/test/resources/simm-sample/fx-rates.csv
fx-rate-shocks=src/test/resources/simm-sample/fx-rate-shocks.csv
portfolio-derivatives=src/test/resources/simm-sample/portfolio-derivatives.csv
portfolio-initial-margin=src/test/resources/simm-sample/portfolio-initial-margin.csv
portfolio-variation-margin=src/test/resources/simm-sample/portfolio-variation-margin.csv
  • base-currency - the base currency for the calculation. All results will be in the base currency.
  • var-level - Optional field allowing the VaR confidence level to be set. Not generally required as standard default value of 99% is set in the code. However, for small data sets a reduced value may be required.
  • risk-factor-definitions - path to the risk factor definitions file
  • risk-factor-base-levels - path to the risk factor base levels file
  • risk-factor-shocks - path to the risk factor shock file
  • fx-rates - path to the FX rates file
  • fx-rate-shocks - path to the FX rate shocks file
  • portfolio-derivatives - path to the portfolio file
  • portfolio-initial-margin - Optional field allowing the path to a portfolio file for any initial margin to be set
  • portfolio-variation-margin - Optional field allowing the path to a portfolio file for any variation margin to be set

opensimm's People

Contributors

jodastephen avatar jonathansenior avatar kirkwylie avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.