Showing posts with label VaR. Show all posts
Showing posts with label VaR. Show all posts

Tuesday, August 9, 2011

VaR and Stress Testing

Stress Testing is the study of the effects on the valuation of a portfolio of a specified set of changes in risk factors resulting from exceptional but plausible events. This is a procedure to create stress tests, through stress tests that can be varied in nature. And said historical scenarios, that is to say based on past experience, can rub shoulders with hypothetical scenarios said, that is to say based on events deemed possible in the future knowing that any changes of macroeconomic factors , sociological or political. To stress tests can also be added sensitivity tests, the best known is to increase or degrade instantly settings of a risk / multiple degrees or in percentage terms.

Scenario building and testing should follow an iterative and collaborative approach. Indeed, the convergence of views of all stakeholders, whether financial (risk managers ...) or not (economists, business ...), is needed to reach a consensus and identify all the factors (economic, environmental, commercial ...) that would influence the risk parameters. The analysis then produced by stress testing, primarily intended for risk managers and the Directorate General of the institution, and should be a detailed analysis to understand the exposure of the bank. This team approach then facilitates the production process of risk management indicators (dashboards) and ownership of findings by the top management.

Sunday, August 7, 2011

Market Risk Assessment -2

Using a single number, indicate the VaR of a portfolio exposure to market risk and the likelihood of loss under the conditions proposed. The risk is also measured in monetary units, the same as those in which balances are established. Finally, among its other benefits, VaR:

* To evaluate the performance and correct any risk-based.
* Promote the information and transparency to the extent that it is a measure expressed in non-technical terms and may be subject to periodic reports.
* To determine the allocation of funds to invest and set quantitative limits for risk managers.

Three calculation methods are generally used to estimate the distribution of losses. They have in common to estimate the potential change in portfolio value from the data of the past, however, differ on the following:

* The historical method: observation of the historical behavior of the position to estimate the VaR;
* The parametric method: decomposition of the position of instruments based on different risk factors (equity indices, rates of different maturities, exchange rates ...) and then estimate the probability distribution of risk factors;
* The Monte Carlo simulation Monte Carlo market factors from an a priori probability distribution and estimate the VaR, as the historical method, from the sample generated.

Market Risk Assessment -1

Quantification of risk is a major concern of financial players because it allows them to answer questions like "How can we lose with our portfolio in normal market conditions or abnormal for a time horizon given? ". VaR (Value at Risk) and Stress Testing are the two most common methods of quantification of risk and are usually combined.

Although the VaR can be used as a baseline measurement for all types of risk and the overall level of society, its most common use for market risk.
VaR is a probabilistic measure of the loss of a portfolio point of a given composition as a result of future changes in risk factors. It is defined by the maximum probable loss at a confidence level of x% (for a time horizon of one day / one week, etc.).. Var corresponds to the loss that will not be exceeded in more than (100-x)% of cases, when a position given structure is maintained for a period [0, T].

If Vt is the value of the position t, the VaR is given by:
Pr {Vo - VT} ≤ ≥ VaR (100 - x) / 100

For example, consider a portfolio for which the VaR of € 100 million with a confidence level of 99% over a period of one week, meaning that, under normal market conditions, there is a probability of 1 % to record a loss of more than € 100 million for the week of detention (period over which the change in value of the portfolio is measured).