Market Risk
The number and types of traded securities, energy, commodities, stocks, indices, interest rates, currencies and options that can be entered in the system is unlimited across markets, nodes and hubs. You define, name and point to data sources for these markets, nodes or hubs and have full configuration control on them.
Financial and Physical transactions are handled with ease. Spot, index, futures, forwards, swaps, options, fixed price, balance of, volumetric, physical, load following shape, exotic deals, swaptions. For physical, retail and structured transactions, you can enter any complex cash flow based on any number of indices modeling charges, fees, incentives, penalties, transportation costs, etc.
No portfolio structure is pre-set unlike traditional risk systems which only allow for a determined structure often organized in a few commodity or division books. Irrespective on how your transactions were captured, the Risk Server has a dynamic portfolio builder that allows you to rapidly configure any desired multi-layered embedded portfolio structure for them.
Valuation is done recursively across the entire portfolio structure, using the same sets of simulated price and volume sample paths. You can then view risk results marginally at given portfolio levels, using drill-down and roll-up up capability for aggregation. Since valuation is based on correlated sample paths of all basic random processes (prices and volumes), correlation across portfolio layers is implicit.
|
|
The system models stochastically and simulates the FX (and corresponding interest rates) of all transactions and evaluates your portfolio in any desired currency on the fly. While Settlement is done in the native currency of the float of a deal, when performing MTM and Risk, you can switch, for consolidated view, to any currency you desire, and save the results for reference. For example, you can have your entire portfolio evaluated at the same time in US $, and in Euro, etc.
When computing risk some of the features you can set or control are:
- Holding Period can be a few days for Value at Risk or many years for Cash Flow at Risk.
- Number of Monte Carlo simulations used to form a probability distribution.
- Percentile or Risk level (95% to 99%) at which risk statistics are to be computed and displayed.
Click image for a larger view.
For market Risk, the following risk metrics are computed by Monte Carlo simulation:
- MTM and P&L
- Value at Risk (realized, unrealized or both added together)
- Cash Flow at Risk and Earnings at Risk (at any income statement level)
- Liquidity at Risk (stress testing market and portfolio levels)
| Expected Loss | Tail Risk |
Mean or Expected Value |
Tail Gain |
|
The average below the red risk tail of the distribution. This is the average amount you could lose if the worst case occurs. This stat is called CVaR (Conditional VaR) as it is the average value conditional on the value being below the red line. For skewed prices such as electricity it can be significantly lower than the risk percentile. So CVaR is often more meaningful than VaR. |
This is the worst case scenario or the lower percentile of the distribution, shown in red. If risk level is 95% it means that 5% of the time (or 1 time out of 20) value will fall below the red line. A useful number is Value at Risk computed by the difference of green and red lines: VaR = Mean - Tail Risk |
This is the expected future value of your portfolio. It is also the mean value of the distribution at all moments in time. Note the green line is also the forecasted settled values of your portfolio at all times. Risk engine centers sample paths around values predicted by settlement module. So risk and MTM are always consistent. |
This is the up side or the higher percentile of the distribution, shown in light blue. It represents the upside or best case scenario. (Similar to Tail Risk, except in reverse) If risk level is 95% it means that 5% of the time (or 1 time out of 20) value will be above the light blue line. |
| Risk Plot Aggregated through Time | Risk Plot by Monthly Time Buckets |
|
|
The Monte Carlo engine performs comprehensive stress testing of market price levels and volatility as well as portfolio positions. Unlike most risk systems, we do not just shock the forward curves and leave volatility and correlations intact. All facets of the entire Monte Carlo engine may be adjusted and resimulated at run time.
You can use time-bucket valuation to see your risk in weekly, monthly, quarterly, or yearly blocks. You can thus optimize and/or hedge your portfolio. E.g., monthly blocks allow you to hedge risk via forwards or futures. This additional dimension allows you to compute and display risk for a bucket across portfolio levels or vice versa. You can also integrate risk values Forward as cash flow accumulates, or Backward as positions unwind.


