Energy risk management: the demise of Value at Risk?

19 Mar, 2012

‘Value at Risk is like an airbag that works all the time, except when you have a car accident’. – David Einhorn, hedge fund manager.

There has been a backlash against probabilistic ‘at risk’ measurement methodologies since the onset of the financial crisis.  The use of Value at Risk (VaR) has been at the core of some of the disastrous risk management failings suffered by banks over the last five years.  As a result, some well informed critics have launched a crusade against the methodology, focused on the false confidence VaR provides about extreme risks which it does little to capture.  But what role does VaR play as a risk management tool in European energy companies?

Energy companies are less VaR focused

There are strong ties between the commodity risk management methodologies of banks and energy companies.  The mark to market (MtM) principles that govern exposure valuation at banks, also underpin the risk management approach of energy companies.  And VaR, a measure of potential deviation in MtM value, has been adopted as the standard measure to manage the risk associated with energy trading functions. 

But for energy companies with portfolios of physical assets, VaR is typically accompanied by a broader risk measure focused on the potential deviation in portfolio ‘earnings’ over a reporting horizon.  This ‘earnings’ at risk (EaR) measure goes by different names depending on the earnings measure chosen, profit at risk (PaR) and cashflow at risk (CFaR) being two common alternatives.  But all these methodologies share a common goal: to measure the risk associated with structural asset exposures that can not easily be closed out, for example from power plants, gas contracts and customer portfolios.

Despite a similar name, there are some fundamental differences between the VaR and EaR approaches, summarised in Table 1.

Table 1: Key differences between Value at Risk and Earnings at Risk

VaR is not enough…

The principles behind VaR are an attractive proposition for a risk manager.  Quantifying risk by focusing on forward prices against which exposures can be liquidated is a powerful framework from which to govern risk.  But the effective application of VaR in energy companies faces serious limitations when it is applied to trading exposures that relate to the hedging of structural asset positions. 

Measuring VaR on forward gas trades associated with the hedging of gas storage or interconnector capacity has little meaning in isolation of the exposures of the asset itself.  And calculating an effective VaR on a complex flexible asset like storage capacity is very difficult.  Even if VaR is implemented on an approximation of the asset’s price exposure (delta), this falls a long way short of capturing asset exposure to price volatility (vega).  Life becomes even more complex when there are liquidity constraints that mean interaction between portfolio asset exposures need to be considered.

EaR as a risk measure is not as clean as VaR because it is not underpinned by an ability to constrain losses via liquidating exposures in the forward market.  EaR measurement is more diagnostic, focused on understanding the complex interaction between physical assets and hedges under different market outcomes.  But EaR does inform senior management as to how hedging actions can impact portfolio risk.  For a portfolio of flexible physical assets, an effectively implemented EaR framework may be a much more powerful risk management tool than a VaR based approach. 

For example utilities typically manage the risk around their retail load by contracting with a rolling forward hedge profile several seasons ahead of delivery.  Hedging may be internal (with their own generation assets) or external in the forward market.  Understanding portfolio risk on a scenario basis is difficult.  But an effective EaR framework  provides an insight into the interaction between customer contract prices, hedge profiles, movements in underlying commodity markets and load uncertainty.  This information is valuable for risk managers but also for commercial decision making on customer pricing and load hedging.

Pragmatic use of ‘at risk’ measures

Energy exposure risk measurement does not come down to a choice between EaR and VaR.  An effective framework should utilise both measures, with a focus on EaR for structural asset exposures and VaR for speculative and proprietary trading exposures.  However no degree of sophistication of ‘at risk’ calculation is a substitute for developing complementary risk measures.  An effectively implemented stress test framework is an invaluable tool for understanding the impacts of extreme risks.

The failings of VaR in the financial crisis were primarily the result of a ‘fundamentalist’ approach to the application of the VaR methodology, specifically:

  • The mechanistic application of VaR to support key risk management decisions e.g. limit setting and the allocation of risk capital
  • Failure to recognise the weaknesses and limitations of VaR, particularly in measuring extreme and highly correlated risks
  • And a blind belief in VaR as the only risk measure required

The damage suffered by banks from the misuse of VaR was magnified by the scale of leverage employed to ‘juice’ trading returns, a factor less relevant to energy companies which tend to take a relatively conservative approach to leverage.  While the financial crisis has highlighted some of the pitfalls and weaknesses of ‘at risk’ methodologies, they are still a key component of an effective energy risk management toolkit.  As important as the €m quantification of risk, is the understanding gained from analysing the interaction between portfolio exposures and risk factors.

Relevant Articles:

Preparing your business for a eurozone breakup?

Robust analysis of energy asset value and risk

Stochastic pricing models: Science, Art or Voodoo?