Asset-Based Trading Financial Engineering tools for asset-backed trading and hedging strategies. By Carlos Blanco & Michael Pearce 79 HERE, WE DISCUSS the role of risk and trading analytics to support asset-based optimization and trading strategies building upon our prior contributions which showed how dynamic risk simulation tools can be applied to complex decision problems involving multiple sources of risk and state variables as well as calculate risk, value and performance metrics. 1 Optimization of Physical Assets & Trading Strategies Traditionally, physical assets such as storage facilities, pipelines, transmission networks, refineries, and power plants were operated by engineers rather than trading groups, resulting in sub-optimal performance from a profit maximization perspective. Many energy and commodity firms have established asset-backed trading groups whose objective is to enhance the risk adjusted profitability of its physical assets based on observable market spreads, their potential variability, as well as the specific asset operating constraints. It is critical for asset-based trading groups to articulate and communicate the value and risk of those strategies with asset operators as well as senior management. For different reasons, poor communication and lack of understanding between engineers that operate and maintain the assets and the financial engineers in charge of modelling the performance of those assets is the norm rather than the exception. For example, the first generation models used by risk groups and traders were too simplistic and did not accurately capture the market and asset dynamics and constraints, leading to poor or unrealistic decisions from an operational perspective. Figure 1 shows the key building blocks for the design of decision support analytics for asset-based strategies. The valuation, risk metrics and hedging ratios calculated from model that do not fit the criteria outlined above are not just likely to be inaccurate, but also likely to lead to sub-optimal decisions that would impact the asset-base strategys profitability. Fortunately, one of the more relevant developments in recent years in financial engineering applied to energy risk modelling is the pricing, modelling and hedging of physical assets and asset-based strategies for power and gas portfolios. Risk Management For Asset-Based Trading Strategies Risk management for asset-based trading strategies involves identifying and quantifying multiple risk dimensions. To capture those risk dimensions, a dynamic simulation framework with three critical components is needed: Ability to handle multiple risk factors (e.g. commodities, credit events, operational issues), multiple instruments (e.g. physical contracts, derivatives), as well as the ability to capture events taking place at multiple steps in time. Many energy and commodity firms have established asset- backed trading groups Figure 1: Building Blocks of Asset-Backed Trading Support Analytics Source: www.nquantx.com Realistic Multi-Step Spot, Forward Curve & Spread Scenario Asset & Trading Constraints Explicitly Captured & Modelled Integration Of Market Liquidity, Hedging & Transaction Costs Backtesting Performance Of Asset-Based Trading Strategy Dynamic risk simulation involves modelling the variability of one or more metrics (e.g., cash flow, earnings, mark-to-market, liquidity, etc.) based on realistic potential changes in a set of key state variables, as well as the firms response to those changes (e.g., operating, hedging and trading strategies). The analysis consists on leaping forward at various points in time into the future. New technologies exist which streamline the storage and consistent definition of data across the enterprise, and risk management systems should take full advantage of them to achieve consistency of risk information across different business units. For example, contract, market and counterparty information that often resides in different systems used by market and credit risk groups needs to be integrated in order to perform Earnings at Risk (EaR) and Cash Flow at Risk (CFaR) analysis (Figure 2). Another problem is that the risk management groups in many firms are still organized in silos, and often focused on measuring individual risks, or risks for a component of the portfolio. For example the financial portfolio to hedge or monetize the value of the asset is just a component of the profitability of the asset, and analysing it in isolation is not sufficient. Spreadsheets vs CTRM Systems: Combining the best of both worlds Decision-support tools for asset-based trading strategies need to strike a balance between flexibility and integration into existing systems. Those tools need to be capable of taking multiple inputs coming from different systems, and also perform a large number of computations in simulation-based environments. The first column in Table 1 illustrates the main problems with the current set up of decision support analytics at many firms, which includes a mix of individual spreadsheets design to solve specific issues, stand-alone solutions developed by quantitative teams that are not fully integrated or vetted (e.g. Matlab-based solutions) and tasks performed by larger systems. The second column suggests solutions to each of those problems. On one side, E/CTRM systems are usually designed to excel at performing other tasks such as scheduling, nomination or accounting for physical trades. However, their limited offering in terms of trade- support analytics and asset optimization have the limitation of being non-intuitive and inflexible to perform customized analysis. A flexible environment such as the one provided by spreadsheets as well as programming languages September 2011 80 New technologies exist which streamline the storage and consistent definition of data across the enterprise ASSET BACKED TRADING Performance Measurement For Asset-Based Strategies One of the areas where firms can reap tangible benefits from deploying risk analytics for asset- based strategies is in the process of setting up performance benchmarks. If a trading group relies on the firms assets to meet their profit targets, but traders are not charged an explicit premium based on the optionality embedded on those assets, they are clearly operating at an advantage over other units. As a result, their raw performance is often stellar, but there may be reasons besides the skill or luck of the traders. Setting up the baseline profitability based on mechanical low risk trading strategies can level the playing field, and also assist in determining which traders are truly adding value. For example, storage assets can be dynamically hedged by optimizing the intrinsic value of the asset over time. That intrinsic value could be the baseline profit to measure the contribution of the traders, not the actual performance. Figure 2: Application of Decision-Support Analytics For Asset-Backed Trading Strategies Source: www.nquantx.com Valuation Hedging Market Risk Budget Risk Cash Flow Risk Credit Exposure Operational Risk Volumetric Risk Performance Measurement September 2011 ASSET BACKED TRADING 81 for statistical and mathematical computing (such as R and Matlab) can enable analysts and traders to monitor the impact of alternative hedging strategies or perform sensitivity analysis to change in market conditions or evaluate operational events. Stand-alone spreadsheets play a critical role and can provide quick answers not available through other systems, but their use also introduce other issues that need to be evaluated. For example, traders and analyst still spend a significant amount of their time maintaining and updating those spreadsheets. That process is highly prone to human error, and therefore introduces significant model risk for their firms. In addition, the more sophisticated spreadsheets are not easily audited and often fall outside the control of the risk, IT and audit groups. For example, if a trader moves to a competitor, poorly documented and tested spreadsheets are likely to be abandoned, which results in a loss of know-how for the firm. Instead of relying on inflexible CTRM systems or stand-alone spreadsheets, a better solution would be one that would allow traders and risk managers to continue using spreadsheets as their front- end tool, but with integrated data tools as well as analytic support engines running underneath and performing the intensive computation tasks. The right modular framework design would also increase the communication with other legacy systems. A flexible analytic engine for asset-based strategies should allow for the segregation of analytic processes to enable modular design and increased scalability. For instance, the separation of instrument-level valuation analytics from the market scenario generation engine increases the ease of editing or adding new valuation or Monte Carlo routines in the future.
The right modular framework design
would also increase the communication with other legacy systems systems Table 1: State of Asset-Based Trading Support Analytics Asset-Based Trading Support Analytics (ETRM + Spreadsheets) NQuantX Asset-based Trading Solutions (Integrated analytics with Excel interface) Multiple spreadsheets individually developed and maintained Excel interface to risk libraries High model risk Tested and supported code and applications. Source control Manual workflow/ intensive Automated Workflow Disconnected and/or unsupported trading analytics function Dedicated analytics group committed to continuous enhancements Reliance on individual expertise. Personnel risk Best of breed. Access to financial engineering support group Source: NQuantX, LLC Carlos Blanco is Co-Founder and Managing Director of NQuantX, LLC, a nancial engineering rm that develops customized software to design and implement hedging programs and trading strategies, as well as valuation and risk measurement of energy derivatives, long term contracts and physical assets. He also conducts several courses on energy derivatives hedging, pricing and risk management, as well as credit and counterparty risk management for the Oxford-Princeton Program. He is a lecturer on risk management at the University of California, Berkeley. E: carlos@nquantx.com Michael Pierce is Co-Founder and Director of Financial Engineering at NQuantX LLC. Some specic areas of development include energy forward curve calibration and modelling, development of a hybrid electricity model used in Monte Carlo real asset valuation of generating assets and load serving agreements, correlation matrix calibration and multi-regional temperature modelling for weather derivatives www. nquantx.com NQuantX Global Commodity Workbench was designed to meet needs of firms with asset-based strategies such as valuation, hedging and risk quantification by offering software solutions and advisory services. The end result is that traders can continue working with spreadsheets, but in a controlled environment and where it is easier and faster to deploy new solutions. Footnotes 1. See www.commodities-now.com for more information.