Emerging Practices for Capital Adequacy © Copyright 2003, CCRO. All rights reserved. 18 Figure 9: Using Monte Carlo Simulation to Assess Market Risk Economic Capital Because of these complexities, devising closed-form analytical solutions for energy markets is difficult. Simulation offers flexibility in handling these features in the underlying price processes, whether through using current market price information, financial and economic time series data, or probabilistic models. Monte Carlo simulation accounts for these intricacies but does so through a much more complex analysis. Simulation does have its drawbacks, primarily because of the sheer number of iterative simulations and the proper treatment of the multidimensional correlation matrix.6 The greater the number of simulations, the more precisely the simulation converges to the underlying price probability distribution. A typical complex simulation may require between 5,000 and 10,000 iterations ― a number that takes considerable time and computing power to complete, especially for large, complex multi-commodity portfolios (across geographic regions and covering many delivery periods). As more complexities are introduced into estimating price behavior and, ultimately, modeling the exposures of the asset/business, the more suited a simulation solution becomes for defining this process. The user must constantly evaluate the need to model such complexities for their respective asset/businesses in order to determine the appropriate solution for quantifying price behavior and the impact on the company’s financial performance. Table 3 summarizes analytical and simulation solutions. 6 The correlation matrix can be tricky to manipulate because of its dimension, which is governed by the number of risk factors modeled (commodities, time intervals, and delivery locations). Market Risk Economic Capital Probabilistic NOPLAT 1 Iterations (5 - 10,000) Period Probabilistic NOPLAT 2 Probabilistic NOPLAT 3 Probabilistic NOPLAT 4 Probabilistic NOPLAT 5… Iterations (5 - 10,000) NPV