GARP_FRM-package: Functions to implement the topics presented in 'Financial...
In GARPFRM: Global Association of Risk Professionals: Financial Risk Manager

Description Details Author(s) References See Also

This package provides a framework to implement the topics and ideas presented in 'Financial Risk Manager (FRM) Part 1: Quantitative Analysis' (2012) and 'Financial Risk Manager (FRM) Part 1: Foundations of Risk Management' (2012).

Package:	GARPFRM
Type:	Package
Version:	0.1.0
Date:	2014-04-03
License:	GPL

The purpose of this package is to implement the concepts and methods presented in the Global Association of Risk Professionals (GARP) Financial Risk Manager (FRM) Part I series of books. Developing the GARPFRM package is a collaborative project between the University of Washington Computational Finance & Risk Management Program and the Global Association of Risk Professionals to develop R packages that facilitate the learning of risk management concepts. The GARPFRM package provides a framework to implement the topics presented in the Financial Risk Manager (FRM) Part 1 books. The user should be able to follow along in the books and using the GARPFRM package and supporting packages.

TODO: Add stuff about GARP

TODO: Add stuff about UW-CF&RM

The current version of the package covers the majority of chapters in 'Financial Risk Manager (FRM) Part 1: Quantitative Analysis' (2012) and 'Financial Risk Manager (FRM) Part 1: Foundations of Risk Management' (2012). Due to the nature of econometrics involving time series data, nearly all functions in this package require data as an xts object and GARPFRM uses the xts package for working with time series data.

Delineating Efficient Portfolios

This section covers the risk and return characteristics of combinations of securities and what is known as the efficient frontier. We first consider the case of a portfolio consisting of two assets and the impact that correlation has on the shape of the efficient frontier. We also learn how to compute the optimal portfolio, optimal in the sense that the portfolio has the maximium return for a given risk. These ideas are then generalized for a portfolio consisting of any number of assets. Key functions in this section are portSDTwoAsset, portReturnTwoAsset, efficientFrontierTwoAsset, efficientFrontierTwoAsset, and efficientFrontier. See vignette("DelineatingEfficientPortfolios") for several examples and explanations demonstrating these functions.
Capital Asset Pricing Model

The CAPM section discusses estimating alpha and beta as well as their statistical significance. Visually interest functions are chartSML and plot. Of primary applicability are the functions getStatistics and hypTest used to retrieve alpha, beta and their relevant statistic as well as estimating their significance. getStatistics, getAlphas, getBetas, plot, hypTest, and chartSML. See vignette("CAPM_TF") for an implementation and explanation of these functions.
Performance Measures

The Performance Measures section to evaluate performance in terms of return and risk, i.e. risk adjusted performance measures, to more easily compare assets with different levels of risk. The functions provided in this section are from the PerformanceAnalytics package. The key functions in this section are TreynorRatio, SharpeRatio, CAPM.jensenAlpha, CAPM.jensenAlpha, TrackingError, InformationRatio, and SortinoRatio. See vignette("PerformanceMeasures") for several examples and explanations demonstrating these functions. See here for an interactive web application.
Quantitative Analysis

Chapters 1-8 in Quantitative Analysis cover probability, statistics, and linear regression. Functions to implement these topics are included in the base R distribution. See vignette("QuantitativeAnalysisBasics") for several examples and explanations.
Monte Carlo Methods

The Monte Carlo Methods chapter presents simulation methods used in derivatives pricing and risk management. The implementation in this package focuses on Monte Carlo simulations with one random variable (i.e. one source of risk) and bootstrap resampling. Key functions are monteCarlo, bootFUN, bootMean, bootSD, bootStdDev, bootSimpleVolatility, bootCor, bootCov, bootVaR, and bootES. See the demos demo("monte_carlo") and demo("bootstrap"). See vignette("MonteCarloMethods") for several examples and explanations. Additionally, an interactive web application is available here for generating asset price paths via Monte Carlo simulation.
Estimating Volatilities and Correlation

The main emphasis of this section is estimating volatilities and correlations using an EWMA model and a GARCH model. We provide an EWMA function to estimate volatilities, covariances, and correlations. We also provide a forecast function to forecast volatility from an EWMA model. To implement univariate GARCH models, we make use of the rugarch package. The uvGARCH function is a wrapper around ugarchspec and ugarchfit to specify and fit a GARCH model. We provide a forecast function to forecast volatility from a GARCH model. The forecast function is wrapper around ugarchforecast.

See vignette("EstimatingVolatilitiesCorrelation") for several examples and explanations of the EWMA and GARCH models. We also provide interactive web applications to implement EWMA and GARCH models.
Quantifying Volatility in VaR Models

This section discusses Value at Risk (VaR) models and different methods to estimate VaR levels. Among the methods discussed are EWMA, GARCH, historical simulation, and the VarCov approach. See vignette("QuantifyingVolatilityVaRModels") for several examples and explanations. Additionally, an interactive web application to implement VaR backtests is available here.