ARX ANALYTICS & ADVISORY

Leveraging our core analytical expertise, we are continuously researching and developing a range of financial analytical products for financial institutions such as banks, asset management firms, insurance firms etc. The processes and methodologies which ARX develops are robust and effective and are based on extensive research and where necessary, discussion with academics.

ARX MARKET RISK MODULE

Market Risk Module which will enable institutions to effectively evaluate and manage risk across interest rate, commodity, equity & currency exposures through calculation of measures such as Value-at-Risk (VaR), Conditional Value-at-Risk (CVaR) and Component VaR and covers a wide range of instruments types.

An industry best practice for estimating the market risk of trading operations involves projecting profit and-loss distributions of portfolios of financial instruments over short time horizons and then summarizing that information into single number, such as value at risk (VaR). Easy to understand and conceptually straightforward, VaR has long been an industry standard for estimating market risk.

Further, the New Basel II Accord refers to “stress testing” numerous times and in relation to each of its three pillars. The recent developments in financial sector have underscored the importance of suitable instruments which not only assess the vulnerability in the financial system but also identify the specific risk to financial institutions.

Stress testing supplements VaR estimates and provides banks with a robust estimate of Market Risk. ARX Market Risk Module enables banks to define scenarios and assess the impact of such scenarios on the profit or
loss of the portfolio.

These scenarios may be based on historical events such as distress economic condition or choosing one off scenarios and embedding them into a valuation model. Scenarios comprise shocks, to risk factors, specified as an absolute, a percentage, standard deviation or log standard deviation shift in the values. Thus our model is stress tested by extreme scenarios that can occur from time to time but are rare according to the probability distribution assumed for risk factors.

With Clients we are totally transparent in the technology and methodology used in building and validating the individual models which integrate into the VaR module.

Features

A. Drill down risk at any level

ARX Market Risk Module enables institutions to estimate the risk for multiple, user-defined portfolios. Portfolios are defined based on a combination of one or more dimensions such as Counterparty, Line of Business, Asset Class and Instrument Type Etc. This enables assessment of risk at any required level of granularity in the organization.

B. Customized portfolio risk analysis

ARX Market Risk Module offers the flexibility of calculating the risk of the portfolios based on user defined parameters such as time horizon, decay factor, confidence level and supports seasonality adjustments, which are crucial in case of commodity exposures.

C. Regulatory Requirement (Basel II)

Instrument valuation and risk measure estimation is based on industry standard methodologies and statistical techniques. ARX Market Risk Module provides for Analytic, Historical Simulation and Monte Carlo Simulation Methods for estimation of risk measures.

Risk Calculation Process

A. Pricing of Instruments

ARX Market Risk Module has integrated pricing routines that enable valuation of a wide range of instruments. Instruments are priced based on their underlying risk factor. The pricing functions also estimate the Greeks such as Delta, Gamma, and Rho etc. of the option instruments.

B. Risk Factor Modeling & Simulation

Underlying risk factors for each instrument are modeled using stochastic processes such as Black, Ho-Lee & Hull-White, Etc. The parameters of each of these processes are estimated using standard statistical techniques

C. Risk Measure Estimation

ARX Market Risk Module enables institutions to estimate the risk measures such as VaR, CVaR, Undiversified VaR and Component VaR for multiple portfolios of market traded instruments based on user specified parameters. While VaR measures the worst expected loss due to the changes in the market risk factors at a user-specified confidence, CVaR is the average of losses exceeding the VaR value and is a coherent measure of risk.

Model Validation

Under validation, ARX Market Risk Module the actual profit & loss are compared with the estimated VaR values and calculating
back testing measures based user defined look back period. In addition to reporting the number of exceptions, the application
utilizes a Kupiec Test which verifies whether the frequency of exceptions is greater than the frequency of expected exceptions
as determined by the VaR model. Other back test measures computed as part of validation of market risk model
include loss exception deviation, average loss duration and loss duration deviation.

Stress Testing

The New Basel II Accord refers to “stress testing” numerous times and in relation to each of its three pillars. The recent developments in financial sector have underscored the importance of suitable instruments which not only assess the vulnerability in the financial system but also identify the specific risk to financial institutions.

Loan Delinquency Modeling

ARX has developed a model using Markov chain principle for predicting transition probabilities of various loan delinquency buckets. Since event of default for any obligor occurs after going to various delinquencies, therefore it is essential to monitor the migration of loans into various delinquencies.

A Markov chain is a random process with the property that the next state depends only on the current state. Markov chains are useful as tools for statistical as well as stochastic modeling in almost all fields of modern applied mathematics.

Markov chains are used in variety of different phenomena, including asset prices and market crashes. Dynamic macroeconomics heavily uses Markov chains. An example is using Markov chains to exogenously model prices of equity (stock) in a general equilibrium setting. Markov chains are being used to estimate probability of loan default and probability of loan delinquency.

Portfolio Optimization

ARX is developing various optimization techniques and models for more sophisticated analysis of the portfolio. These optimization models play an important role in financial decisions.

Financial applications have a long history of including optimization, starting with Markowitz’s origin of the quadratic optimization model for determining an efficient portfolio to minimize variance for a given return.

Portfolio optimization continues to be an active area with most applications focused on linear and quadratic optimization. General nonlinear optimization arises in this area, however, as models begin to acknowledge and capture the nonlinearity, asymmetry, and non-normality associated with returns in practice. In addition, complex financial products often involve a variety of nonlinear relationships that lead to nonlinear optimization in parameter estimation, tracking, and hedging. Credit instruments and their risk management also introduce nonlinearities that are difficult to include in linear or quadratic models e.g. use of Optimization to estimate and reduce delinquencies and charge-offs in loan portfolio.

ARX is in the process of developing various models using optimization techniques in the following areas:

• Portfolio Selection
• Risk Management

Incorporating uncertainty:

• Stochastic optimization
• Robust optimization

Extensions of the models:

• Nonlinear optimization
• Large-scale problems