Structured energy contracts and VaR

23 Apr, 2012

It is common practice for energy companies to measure and control the risk around energy contracts using a Value at Risk (VaR) metric.  This is relatively straightforward if the contract exposures can be marked to market against liquid forward curves.  However more complex structured contracts which are illiquid or have embedded optionality, such as power tolling or gas swing contracts, present some interesting challenges for effective VaR implementation. 

A home for complex contracts?

Structured contracts are common in energy companies, both as a means of managing more complex asset flexibility with counterparties and as a means of internal exposure transfer from a company’s Asset to its Trading books.  Common examples of structured contracts include:

  • Gas swing contracts – intertemporal optionality
  • Gas storage capacity contracts – intertemporal optionality
  • Indexed gas contracts – non-linear cross commodity exposures
  • Interconnector capacity contracts – interlocational optionality
  • Power tolling capacity contracts – cross-commodity spread optionality

These contracts are typically managed by the trading function within an energy company, where they are optimised and hedged in liquid forward markets.  As a result, risk on structured contracts is typically governed using the VaR metrics that govern other trading activity.  However the fundamental assumption behind VaR, that exposures can be liquidated within a holding period, does not sit comfortably with structured contracts that are illiquid and exhibit complex exposures.

In a previous article we explored the benefits and limitations of VaR as compared to ‘Earnings at Risk’ (EaR) based methodologies.  In many ways, complex and illiquid contracts that are part of a company’s structural portfolio (e.g. long term gas supply agreements), are better governed under an EaR than a VaR framework, given MtM and holding period issues.  But there are often practical considerations around reporting the risk on these contracts in a manner that is consistent with the other exposure activity within a trading function.  If a VaR approach is taken, there are some important challenges to overcome in implementing it successfully. 

Adapting VaR for complex contracts

One of the key issues with applying VaR to complex contracts is dealing with market liquidity within the holding period.  A two-day holding period may be a reasonable assumption for liquidation of relatively liquid forward and futures contracts, but complex tolling or gas swing contract exposures may take a lot longer to close out.  There are two approaches commonly adopted to address this problem:

  1. Extension of the holding period to a horizon within which an Origination function can assist with the liquidation of more complex exposures
  2. Liquidity discounts that reflect an inability to fully liquidate the exposure, e.g. assets are measured at the intrinsic value which can be hedged in liquid markets

 There are issues with both approaches. 

Extending the holding period:

  • To be able to effectively measure VaR on a portfolio of contracts a consistent holding period is required, but liquidity will tend to vary depending on contract type (e.g. an annual storage capacity deal may be closed out in 2 days while a 5 year gas swing contract takes 2 weeks)
  • Liquidity is fickle and will also vary with market conditions; it has a nasty tendency to dry up when it is most required, i.e. in the ‘worst case’ situation that is the focus of VaR
  • Extending the holding period becomes a hypothetical exercise if a company does not intend to liquidate a structured contract in a ‘worst case’ scenario, undermining the application of the VaR metric.

Liquidity discounts:

  • Discounting the value of a contract to reflect for liquidity issues in closing out its exposures is a subjective exercise that needs to be agreed by trading and risk control functions
  • The most objective method to derive a liquidity discount is to value the contract at the intrinsic value that can be hedged in the forward market, but this is likely to be a clumsy solution if the contract extends beyond the liquid curve horizon or has significant value associated with embedded optionality
  • Alternative methods tend to rely on a ‘mark to model’ approach which is subjective, lacks transparency and does not capture the impact of changes in market conditions on liquidity.

Defining an effective ‘mark to market’ (MtM) methodology for structured contracts is also a key challenge, but one we have dedicated a separate article to.  MtM issues carry over for VaR calculation (which is underpinned by MtM) and a consistent approach to MtM is as important as holding period issues.

An inconvenient solution

The risk management treatment of structured contracts is hampered by the problem that they often do not have a natural home in either a company’s Asset or Trading books.  The convenient solution when implementing VaR, is to ignore the holding period issues described above and apply VaR metrics in a mechanical fashion consistent with other less complex exposures.  But the price for this convenience is the undermining of VaR as a risk measure. The inconvenient but more effective solution may be to separate out structural portfolio positions for EaR treatment and more complex traded contract exposures for a tailored VaR treatment that reflects liquidity characteristics.