Using limits in risk management – some limitations

Blowing the limits on using limits – What risk management practices and policies could do for energy traders if a ‘market catastrophe’ crisis hits the oil markets.

The recent cycle of financial market upheaval with major investment banks succumbing to derivative structures and the credit crunch begged the question – what did their risk management systems show them about the impending situation before it hit them. These institutions – among them Merrill Lynch, Bear Sterns and Lehman – were the most sophisticated users of risk management methodologies and systems and yet their reaction to worsening situations were not quick enough to prevent their demise. Energy derivatives also trade around illiquid markets and with high volatility and are exposed to underlying impacts of currencies and interest rates with the saving grace that the underlying asset can be much better managed than sub prime mortgages or derivatives thereof. This article explores some of the risk management best practices that have been developed for the crude and refined products markets and their relevance in the possible situation of ‘market catastrophe’.

The world of limits

Commodities trading risk management practices have focused on placing limitations on maximum fixed price or basis exposure and using limits to identify and track opportunities and threats. ETRM systems commonly support position limits, P&L swing limits, VaR limits, Credit limits, PFE (potential future exposure) or Credit VaR limits and have capabilities to notify on breaches of these limits. However there are many cases where the business processes to manage these limits are not updated to the complexity of today’s marketplace. For instance a semi annual or annual review of P&L swing limits does not take into account 5 and 10 $/barrel moves in a single day. Similarly setting static limits on outright or spread positions may be restrictive on a trading floor’s ability to squeeze out margin when opportunity presents itself or to be creative in how the market is being approached. Time is lost in investigating if breaches are indicating heightened risk or the ‘limitations’ of the limits regime. When the market can move from $147 to $91 a barrel in 2 months and it takes 25 days to move a cargo across the Atlantic, does that indicate a need to refine limits on a monthly basis? Or when a counterparty’s credit rating drops, or the rating of the bank providing the securitization for a counterparty drops – does that require a change in the limits on them or closer scrutiny till the exposure is liquidated? Making limits dynamic or even transient based on rules may offer some level of comfort on a trading floors ability to rapidly react to market changes without giving up the controls to regulate the operations of wet transactions.

The Quality factor

There is increased debate on whether statistical models that rely largely on volatility to represent the encapsulated risk due to a variety of market factors are adequate in the current market scenarios. One school of thought believes that all political, delivery and price risk are effectively captured with the single measure of volatility and that the distributions spread over the time to liquidation present an effective way to capture the risk. But instinct suggests there is a quality difference in the risk represented by a $1 Million VaR on a position on for example the BFOE market vis a vis the Nigerian crude market. Country risk factors can accommodate some of the underlying differences in political risk and the differences in liquidity and ETRM systems need to offer more capabilities for the risk manager to easily and independently manage such factors into each calculation set. The quality of risk in a counterparty’s credit rating and the rating of the bank offering credit security for a movement also needs to be factored into the PFE or CVaR models to bring added credibility to the number presented. Just as prices and actuals are reacted to in real time for position and P&L updates for risk management – the increased market complexity may be calling for new measures including real time currency and credit ratings into the statistical model of risk. Today the question often becomes – is a single measure such as VaR or CVaR sufficient to be used as limits on trading activity and what is the appropriate horizon to which the risk tests need to be applied.

Binary events such as the bankruptcy of a counterparty or a system wide problem such as the credit crunch also do not lend themselves to easy modeling in the ETRM context. Again the quality factor to be applied to results ex ante to such events may create limits that severely restrict the ability to leverage opportunity in the market place. Historical data are not indicators of the outcomes in such cases and this presents another challenge in dealing with such possible scenarios. While much has been learnt in the unwinding of Enron’s positions after their bankruptcy – not all counterparties were treated alike or are alike and there is no standard methodology/measure that would predict the impacts of such outcomes. Is the management strategy then to ‘hope that you have title to or possession of the barrels when the storm hits’ – comments from a risk manager.

Dealing with the breach

FEMA took a lot of negative publicity after Katrina for their reaction to the catastrophe. Even though risk managers are tasked with predicting and preparing for the worst, very often they are the ones required to clean up and manage the aftermath of events. As in the case of Katrina, effective procedures, communications and systems support can make the difference between a good or a poor response. Most risk managers and trading desks have agreed upon the principles for the basis of position and P&L swing limits and often a breach is an indicator for taking hedge corrections into strategies to come back within the established limits. Sometimes the limits are moved around to accommodate market conditions and in other cases there are processes for delegation of authorities (DoA) to deal with short term changes. A good ETRM implementation would provide actionable intelligence in what caused the spike that resulted in the breach by a combination of detailed P&L attribution and splitting out the components of the model used to calculate the measure. In many cases corrective action strategies are not defined for the mitigation of VaR limit breaches or PFE breaches and the attribution of the components to the various factors not presented in an intuitive manner which can cause delays in reaction to market changes. Change in time itself can lend to a change in the universe of available options to deal with a situation and so the speed of system support can become a differentiator as victims of Katrina or Enron can attest to.

Using a dynamic set of limits which react to changes in market price levels, changes to volatilities and other factors such as currency rates, interest rates or credit ratings would probably provide an acceptable solution. It would take great confidence in the ability to understand and attribute accurately the pieces causing the change in the underlying calculations to trust a system to prompt a risk manager to modify limits. Having strong processes around the use of DoA functions could be another control. The ETRM would need to behave like a hub collecting updates for the factors above and providing predictive estimations for the measures as well as estimations of modified limits that would adhere to the risk philosophy and levels that the organization could adopt. Managing the quality factors and their impacts would also need careful consideration into a rules engine that could support such functions. A convergence of control and flexibility with a clarity around what needs to be done for possible outcomes would enable the broad adoption of such moving limits in dealing with the complexities that a ‘market catastrophe’ would bring.


The present market structure has led to a lively debate at both the political and academic levels on how best to approach such situations, whether they are predictable and how to deal with the aftermath. We are already facing regulatory limits in the markets through Dodd Frank, MIFID/EMIR etc. and limits on credit extended to businesses. For the impacts of current events on the energy trading world – it will be interesting to see how risk managers proactively engage limits around their trading practices to better deal with the uncertainty of the near future.