Gas Appraiser
The Gas Storage Appraiser is a sophisticated spread option calculator designed to both determine the daily optimal injection and withdraw schedule based on daily quoted forward prices and the unique injection and withdraw limits of the storage asset and to provide a total expected value over the life of the asset. The daily optimal injection and withdraw schedule is used for mark-to-market accounting of the storage asset as optimal positions are compared to actual traded positions. The total asset value is used term value at risk and asset appraisal applications.

Underlying the IA Gas Storage model is the RealVaR optimization algorithms that maximize asset value subject to the injection and withdraw constraints and costs. The model calculates the daily injection and withdraw schedules or MTM schedules by maximizing the forward price spread subject to storage constraints. Although the optimizer solves for monthly spreads, it is able to move through the intra-month ratchets (volumetric related changes in injection/withdraw rates). Its output shows the quantities for each month – both injection and withdraw – and the expected profits.

To value the storage asset we simulate the changes in the forward curves through the life of the asset and rerun the optimizer for each forward draw. We currently use a form of a regime switching GBM model for the forward curve simulation although both mean reverting and two factor models are available if requested. The model pulls unique volatilities and correlations for each month and runs these through a traditional GBM model to get successive realizations of forward prices.

It is important that each realization of a price be precisely related to the other prices comprising the forward curve as the spreads are the main determinants of value. We use a Cholesky decomposition matrix for the random numbers correlating both seasonality and term structure. Both the volatility and correlations matrices measure seasonality and term structure (decay).

The intra-month trading optimizer compares the existing balance-of-month (spot) price to the prompt forward price. There is a probability that the spread could improve and a risk tolerance that is used to determine whether or not the trader should execute the trade or wait for the possibility of a better spread. If the total of the expected value of waiting (including the time value) is greater than the intrinsic (transact now) value and if the VaR of waiting is less than the risk tolerance, then the model will wait for the next random draw. As time goes on both the time value decreases and the volume decreases due to a monthly injection/withdraw rate constraint (maximum per day volume limit) increasing the chance that the trade will be made.

Both intrinsic and extrinsic value are measured, the sum of these are equal to the total asset value over a finite time horizon.  

The model uses the IA proprietary RealVaR optimization algorithms to calculate price spreads subject to physical constraints. A price spread is defined as the difference between two prices. Spread-related price risk occurs when a firm's profitability is driven, not by the actual price level of the commodity, but by the price spread that exists between a firm's commodity purchases and its sales revenues.

The value of most energy assets is driven by the effective management of spread-related price risk. Using Real Option analytics, RealVaR Appraiser values the future arbitrage possibilities of energy assets as price spreads change over time.

Three types of price spreads are traded in the energy and electricity industries.

1. Locational spreads
2. Calendar (time) spreads
3. Product spreads   

Locational spreads drive the value of transportation assets, time spreads drive the value of storage assets, and product spreads drive the value of production assets including:

A locational spread is a difference in price that is caused by geographic location.
The commodity and the delivery period are the same.

A calendar (time) spread is a difference in price caused by two different delivery periods.
The commodity and the geographic location are the same.

A product spread is a difference in price caused by two different products. 

The geographic location and the delivery period are the same.
The crack spread and the spark spread are additional examples of product spreads.

The spark spread, which represents the margin for an electric generation plant, is the difference between electricity prices and the market price of the spark fuel in equivalent terms.

The “spark” fuel can be coal, oil, or natural gas.