Decision Support – ‘Quick and Dirty’ without the ‘dirty’
In the wake of the credit crisis, the energy trading community had to provide quick answers to questions such as ‘what is our exposure to Bear / Lehman?’, ‘what does this change in the FX rate mean to our PL?’, ‘how does the cost of storage compare to shipping this quantity across the Atlantic?’. In the years that have followed, trading groups have recognized that ETRM systems that are well implemented can be the source of the ‘quick’ answers that support the decision making process. The functionality that is being asked of ETRM providers is evolving towards a compilation of the ‘facts’ in the databases, with the application of business rules, scenarios, simulations and statistics to provide outputs that can aid in the ‘quick’ collaboration and communication of decision imperatives, thus enabling problem solving.
A Decision Matrix
The ETRM, as the system of record, holds valuable history of transactions as well as their lifecycle impacts and outcomes. Engaging these historical ‘facts’ through the use of business rules and scenarios can enable users seeking to test their hypotheses to build the basis and get buy-in on their points of view. Different roles such as Marketers/Traders, scheduler optimizers, treasury, risk managers etc. have varying needs for the information represented by the transaction ‘facts’ and this creates a challenge for a uniform functional set that could address the common need set and be built upon to meet the specialized needs for each decision set or problem to be solved. Getting a matrix of decision processes/procedures becomes a building block for developing such capabilities in the ETRM. One such matrix can be to break out by decision level and functional dimensions to support each decision level. For example a typical trading organization would have needs in some or all of the following decision levels:
Along each of the levels the decision imperatives are broken down and looked at from the dimensions of:
A tool that could support such a decision matrix would present functions tailored to each of the decision points along the position lifecycles that could be engaged easily, rapidly and accurately – with full transparency to drill down into the components of the models used – to build the confidence from the users to rely on it. As an example consider a particular crude that is available for a particular refining asset. Based on the history of the transfers from the source of the crude to the asset, the ‘facts’ related to the costs involved, the yields possible of the given crude, the market curves for the period during which the crude would be obtained and when it would be processed, would make up the scenarios available to such a tool. Applying a set of business rules that looks at the worst case of costs expressed as a percentage of the market price and the best case for costs to transport and refine the crude and rules to value the yield of such a crude, gives a set of price points to for that crude. Laying out the cost and value characteristics for each of these scenarios in a visual context that is easy to understand and easy to manipulate can highlight value propositions around this possible crude movement. Hedges could be overlaid on the scenario to seek opportunity and mitigate risks. At the same time running through statistical analysis to check against VaR or Credit risk scenarios can help build the business case for the hypothesis that this crude represents a valuable business opportunity.
Accuracy and Transparency lead to confidence
As in the case of any model, user confidence is the key to successful adoption. The ETRM can be the basis of combining the historical ‘facts’ and projections based on current market conditions and forward curves or projections based on statistical simulations or distributions of possible outcomes to build greater confidence on the decision making process. Taking the example above, if costs were expressed as a percentage of the market price of the crude being considered, scenarios could be run for the worst case where all costs were at their highest proportion, the best case with all costs at their least proportion and a median case where the median proportion of the historical costs are used allows a collaborative understanding of the possible outcomes. This could lead to the decision on the price level that is targeted with a shared understanding of what the final PL performance may be. To build the confidence in the decision process requires that the model be accurate under all scenarios and with transparency into the assumptions used, the calculations as appropriate and the steps used to implement the business rules. Providing clear visualization of the components of the outcomes and the ability to trace the steps and business rules used is most important. It is this transparency feature of spreadsheets that has led to their overwhelming use in validating points of views and models. By embracing this capability while applying better business logic and greater reliability from the ‘facts’ database, ETRM systems can deliver a superior function set than spreadsheets have been able to this far.
Implementation – From ‘gut feel’ to profits
A point of view from the trading floor steps through the execution, liquidation, settlement, audit phases and then becomes part of the history of ‘facts’ upon which the next set of ‘what ifs’ are tested. Building a framework on which a multitude of diverse hypotheses can be tested and differing views on assumptions and facts can easily and quickly be evaluated, is the challenge for the ETRM vendor. Layering a collaboration platform such as Google’s recently introduced Google Wave with clear visualization metaphors that enable a rapid grasp of the dimensions of the business case and enable problem solving processes is the logical next step. As ETRM implementation mature and greater discipline is applied into the setup of scenarios and the supporting ‘facts’ such as prices, cost rates, forward curves, exchange rates, credit risk parameters, shocks etc. – the database of transactions becomes more valuable as it grows and learns the history of events and their impacts on the final outcomes. Keeping the framework for the decision support flexible and generic so that users can configure their unique scenarios and points of view while implementing transparency and accuracy that users can relate to is the Holy Grail for a solution that intends to get broad user adoption.
In conclusion, the convergence of business drivers such as the questions needing answers from trading floors and technological capabilities such as data warehousing and visualization is making the conditions ripe for ETRM providers to step up in their ability to support different levels of decision making. Implementers and systems integrators need to think through the needs for dynamic and statistical simulation in addition to static reporting to create user friendly and highly configurable, collaborative solutions to enable the next generation of tools that help users make better decisions. If the foundations of a realistic set of position lifecycle events, their impacts, business rules and a flexible decision matrix is properly implemented in ETRM systems, the improvements to the quality of decision making and problem solving will be a significant value driver within the organizations that use and learn from them.