AS IF

Swing Contract (Neal Horrell) 

Contracts for the sale or purchase of energy, which provide flexibility as to quantity delivered – a swing feature or swing option – are common in the electric power, oil and gas industries. Typically, a swing option appears along with limitations on daily and cumulative amounts delivered. Energy contracts incorporated the swing feature long before the techniques to value options were developed, thus these embedded options were, as a general rule, mispriced, if priced at all.

Swing contracts are known by a variety of other names, such as base-load factor contracts, flexible nomination contracts, and take-or-pay contracts. They consist of a series of interrelated agreements to purchase an energy commodity over a prespecified period of time at a prespecified price, but at an unspecified rate between a minimum and a maximum. These contracts can also contain a number of complicating provisions that introduce path dependence: minimum and maximum load factors (cumulative or average amounts over some calendar period), take or pay (TOP) provisions, and ratchets (maximum quantity changes from day to day). Thus, unlike the typical option or swap that is “all or nothing”, these instruments have the possibility of partial exercise.

The swing contract arose from two facts about the energy industry: (1) the quantity of energy required is unknown in advance, and (2) there exists a separation between the energy producer, transmission, and end-user. The swing contract arose as a risk sharing and allocating mechanism. This type of contract originally allowed the energy producer to generate sufficient guaranteed revenue to meet debt service obligations and provide collateral (in the form of this contract) to finance development. Typically, it also allowed a purchaser to “make-up” prepaid gas.

Terms and Provisions

• periodic delivery within a prespecified time period
  The range of the contract can have either a specific maturity (which is the current practice) or be in effect for the life of the well.
• nominated amount 
  The nominated amount is the rate of delivery per period, which can either remain constant throughout the length of a swing period or vary based upon the last amount taken.
• strike price
  The strike price can be either fixed at the inception of the contract or linked to a prespecified reference (plus or minus a spread). If the strike is specified in advance, this is termed full protection. If the strike floats and is set at a future date, it is called partial protection. The latter would typically be set during bid week or relative to some index.
• swing right
  The swing amount can either be fixed to specific increments – e.g. 5 million cubic feet (MMCF) – or can be left open. The number of swings can also be limited.
• swing types 
  - A daily swing right allows the holder of the contract to change the amount of commodity that is delivered to it on the following delivery date. This will typically not affect the nominated amount. 
  - Nominated (persistent) swing right allows the holder of the contract to change the nominated amount of commodity that is delivered to it, starting the following delivery date, until expiration of the contract or until a new swing right is exercised. 
• maximum and minimum cumulative or average amounts 
• penalty function
  Should the total amount taken vary outside of a minimum to a maximum range, then compensation is paid to the energy supplier. This could either depend upon the then current spot price (the buyback) or be a fixed unit penalty. The buyback rule (or function) is of the form excess * (spot - strike) or deficit * (strike-spot). These functions are not necessarily constant but could be ratcheted. The fixed unit penalty is per unit amount of excess or deficient delivery.

Example

A typical swing contract may have the following form. Producer A agrees to sell to gas pipeline company B 100 MCF per day at a fixed price for a one-month period. B has the right the day before to alter the amount it purchases by 10 MCF from the previous day’s level (the swing). However; B’s purchases cannot be less than 50 MCF nor greater than 150 MCF. In addition, B must purchase 3000 MCF over the month. 

The decisions rest entirely with company B. It should choose the purchases to maximize the value of the contract. 

Pricing

The valuation of these instruments is quite complicated. This is not only because the options have American style elements, but also they have path dependence.

There are several approaches to valuation of these complex contracts. These include tree approaches, Monte Carlo simulations, and quasi-analytic approaches (closed form solutions). We are proposing the use of dynamic stochastic programming as the modeling methodology.

Electricity and VaR (Neal Horrell) 

Visions of extreme weather flooded American newspapers and airwaves this summer.  Viewers saw withered crops, ruined lives and dead and dying barnyard beasts.  The death toll from the excessively hot weather in the Midwest was nearly 200 people.  Even worse, on July 30^th of this year, Cinergy was forced to default on forward wholesale electricity contracts, at a loss of $73 million.  James Rogers (President and CEO) described the losses to the Wall Street Journal.  He stated that they were due to a 1% weather event.  Additionally he observed that a business is not planned around an event that has a small chance.  It is almost possible to hear the frustration in his words. 

Unfortunately, the Cinergy experience will not be unique among the regulated utilities, who must learn to play the game according to new rules. Under the regulated system, a utility was guaranteed a specific return on capital, subject to some limitations.  This approach actually encouraged the industry to take risks.  If costs rose, they would obtain rate relief.  However, if the company hedged, regulators disallow rate increases if the hedge worked.  Worse, hedging losses would pass to the shareholders.  However, the move to an unregulated environment has to a certain extent removed this possibility while at the same time it has retained the obligation of the utility to supply energy.   In effect, these companies had an option to appeal to the state public utilities commission and recoup any extreme events over time.

Regulated utilities cannot afford to ignore organizational issues. The use of the Value at Risk (VaR) has become the de facto standard for risk management within the electric utility industry. Unfortunately, there was a  failure in the communication of the limitations of VaR to some senior management.  

Additionally, there are three major methodological issues in the application of VaR and its variants. 

1. The assumed distribution of asset returns can be wrong.  This will misestimate the probability of an event.  The traditional implementation of this approach assumed normally distributed asset returns. This is clearly not true in the electricity markets. The traditional approach underestimates the likelihood of an extreme event. 

2. The VaR methodology assumes a stationary distribution. This may be true, but it is clear that there is a wide variation in temperature between summers.  These extreme summers demonstrate a different degree day distribution.  Electricity demand and prices usually show a major peak in the summer and a minor peak in the winter.  This requires a time varying parameter implementation. 

3. A low probability event cannot be safely ignored.  A one- percent event happens 3-4 times per year on average.  This event also had a huge negative payoff.  The combination creates tail events with a large expected value.  The issues can be seen in a simple game.  The game is free to play.  It pays one dollar 99% of the time and costs $99 the rest of the time.  This is a fair game, yet most people would not pay it.  This is because of the high loss associated with the low probability state.  This is analogous to the extreme (low probability) event that Cinergy experienced.

The events of the summer demonstrate an undue reliance on VaR present the risk manager with a myopic view of the world.  The most prudent approach is to use VaR to provide an indication as to the daily typical operating characteristics of the firm.  However, this does not give the true picture of the situation as it ignores the extreme events.  Thus VaR is also combined with scenario analysis.  Those systems combining the two typically avoid betting the ranch.  This combination can have Jim Rogers riding high in the saddle again.