10.0 Introduction

You may remember earlier in the course, when we were discussing transmission congestion and Locational Marginal Prices, that in the presence of transmission congestion the RTO winds up with extra money left over after paying the generators the LMP at their location for all energy produced. We had called this extra money the "congestion rent" or the "merchandizing surplus" and you may still be wondering what the RTO does with that extra money.

In this lesson, we'll see how the congestion rent is tied up with one of the key ways that suppliers and consumers can hedge volatility in LMPs. These financial instruments, known as Financial Transmission Rights or FTRs, are the main topic of this lesson.

Learning Outcomes

By the end of this lesson, you should be able to:

• Identify the difference between temporal and locational risk in an electricity market.
• Define a Financial Transmission Right and a Contract for Differences in electricity markets.
• Calculate the spark spread
• Calculate the value of a Financial Transmission Right
• Construct a perfect hedge using Financial Transmission Rights and Contracts for Differences

Lesson Roadmap

Questions?

If you have questions, please feel free to post them to the Questions about EBF 483 discussion forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.

10.1 Risk in Electricity Markets: Temporal, Locational and the Spark Spread

Have a look at one or more of the following animated maps of Locational Marginal Prices in US RTOs:

• MISO LMP Contour Maps [1]
• JCM Contour Map [2]
• ISO New England Real-Time Maps and Charts [3]

It won't take you too long to realize that LMPs are highly volatile. They are calculated at thousands of different locations and change almost constantly. You may also notice that the difference between LMPs at various locations is also volatile. Sometimes the price at one node is higher than at another node, and sometimes it is lower.

We generally define two dimensions of risk in electricity markets:

• Temporal risk pertains to volatility in the LMP at a specific location over time.
• Locational or "basis" risk pertains to volatility in the LMP across space (between two or more locations).

These two types of risk may need to be managed through various hedging instruments (which we'll get to very soon) but they also may represent arbitrage opportunities. Two of the most common ways of exercising arbitrage in electricity markets are through "virtual bidding" (arbitraging the difference between the clearing price in the day-ahead and real-time electricity market) and through the "spark spread" (the difference between fuel and electricity prices).

Virtual bidding offers a mechanism for electricity market participants to take advantage of differences between day-ahead and real-time prices at a specific location. It involves buying or selling some quantity of electricity in the day-ahead market and then taking an offsetting position in the real-time market. Large financial institutions like investment banks and hedge funds engage in a lot of virtual bidding, but other types of market participants like generating companies and electric utilities also engage in virtual bidding.

The mechanics of virtual bidding are very simple. A market participant first takes a short or long position in the day-ahead market. A short position is known as a "dec" and a long position is known as an "inc." If that market participant's inc or dec clears the day-ahead market (in other words, if that participant would get dispatched if it represented an actual physical need to buy or sell electricity), then the market participant must take an offsetting position in the real-time market. So, a day-ahead dec would be paired with a real-time inc, and a day-ahead inc would be paired with a real-time dec. The quantities offset one another and in the end, the market participant does not have to buy or sell any actual electricity. But the market participant is paid the LMP for the inc and pays the LMP for the dec.

For example, a market participant submits a 1 MWh inc in the day-ahead market, believing that the day-ahead price will be greater than the real-time price. We'll say that the inc clears the market and the day-ahead LMP is $25/MWh. This same market participant would submit a dec to the real-time market, and we'll say that the dec clears the real-time market and the real-time price is $20/MWh. What has basically happened is that this market participant has sold 1 MWh of energy at $25/MWh and bought that same MWh for $20, netting $5 in profit.

The second mechanism for exercising arbitrage is through the "spark spread," which is the difference between the electricity price and the cost of fuel to generate electricity. The arbitrage opportunity that the spark spread represents is typically the opportunity to buy fuel and sell electricity. Sparks spreads in financial markets are typically defined as the difference between the LMP and a natural gas generator with certain characteristics (like a heat rate of 10,000 and variable O&M costs of $2.50 per MWh).

As an example, let's say that the electricity price is $100/MWh and the cost of fuel is $5 per million BTU. The marginal cost of a gas-fired generator at this fuel price with a heat rate of 10 million BTU/MWh and variable O&M costs of $2.50/MWh would be 10*5 + 2.50 = $52.50/MWh. The spark spread would thus be $100/MWh - $52.50/MWh = $47.50/MWh.

The figure below shows some historical LMPs in PJM as compared to our hypothetical gas generation marginal cost of $52.50/MWh. During some hours, the spark spread is negative, indicating that it would not be profitable to buy fuel and sell electricity. During other hours the spark spread is positive, indicating that it would be profitable to buy fuel and sell electricity.

Figure 10.1: The spark spread is positive when the electricity price exceeds the cost of fuel for a representative power plant, and is negative when the cost of fuel exceeds the electricity price.

Source: This image by S. Blumsack © Penn State University is licensed under CC BY-NC-SA 4.0 [4] 

10.2 Financial Transmission Rights and Contracts for Differences

Now we can turn to the question of how locational and temporal risk can be hedged in electricity markets, and what the RTO does with the congestion revenue that it collects. In summary:

• Market participants can hedge locational price risk using Financial Transmission Rights, which are instruments that offset differences between LMPs at defined locations.
• The RTO uses this congestion revenue to pay the holders of Financial Transmission Rights.
• Financial Transmission Rights are auctioned off by the RTO, and the proceeds from the auction are paid to holders of "Auction Revenue Rights," which are typically transmission companies or electric utilities.
• Finally, market participants can hedge temporal price risk using Contracts for Differences. These financial instruments are not bought and sold through any RTO market, but involve bilateral agreements between individual market participants.

10.2.1 Financial Transmission Rights

Financial Transmission Rights (FTRs) are financial instruments that entitle the holder to the difference between LMPs at two defined locations. The parameters for an FTR are:

• A "source" node, which we will call node a
• A "sink" node, which we will call node b
• A quantity, in MW, which we will call M.

Figure 10.2: A Financial Transmission Right can be defined between any two nodes in an electricity network.

Source: This image by S. Blumsack © Penn State University is licensed under CC BY-NC-SA 4.0 [4] 

The holder of an M megawatt FTR from node a to node b is entitled to receive:

M*(LMP_b - LMP_a)

For each hour during the duration of the FTR.

FTRs are typically auctioned off quarterly by the RTO and may have different durations (one-month FTRs versus quarterly FTRs, for example). Most FTRs are structured as obligations, which means that the holder of the FTR gets the difference between the LMPs at node a and node b, no matter whether that difference is positive or negative. If LMP_b > LMP_a then the holder of the FTR is paid money by the RTO. If LMP_b < LMP_a then the holder of the FTR must pay the RTO.

Some FTRs may be structured as options that renew every hour, in which case during a given hour the FTR holder would choose to exercise the option only if LMP_b > LMP_a, i.e. If the payoff would be positive. The payoff from an M-megawatt FTR option from node a to node b would thus be:

Max(0, LMP_b - LMP_a)

FTRs also obey superposition, just like power flows. An M-megawatt FTR defined from a to b and an M-megawatt FTR from b to a will cancel each other out financially (as long as both FTRs are structured as obligations). An M-megawatt FTR from a to b and an M-2 megawatt FTR from b to a would have identical value as a 2 megawatt FTR from a to b.

As financial instruments, FTRs are very similar to swaps. A swap is an agreement to exchange the closing price of two different financial assets. In this case, the "swap" is between two nodes in the power network, not between two different financial assets.

Let's work through a few examples of valuing FTRs. We will use the three-node network in the figure below, where the LMP at node A is $14/MWh, the LMP at node B is $9/MWh and the LMP at node C is $14.50/MWh.

Figure 10.3: A three node network used to illustrate the valuation of Financial Transmission rights. The LMP at node A is $14/MWh, the LMP at node B is $9/MWh and the LMP at node C is $14.50/MWh.

Source: This image by S. Blumsack © Penn State University is licensed under CC BY-NC-SA 4.0 [4] 

• Example #1: A 5 MW FTR from node A to node C would have a value of 5*($14.50 - $14) = $2.50.
• Example #2: A 5 MW FTR from node A to node B would have a value of 5*($9 - $14) = -$25.
• Example #3: A 5 MW FTR from node A to node C plus a 10 MW FTR from node A to node C would have a value of (5+10)*($14.50 - $14) = $7.50.

Now it is your turn. Show that a 5 MW FTR from node A to node C plus a 2 MW FTR from node C to node A would have a total combined value of $1.50.

10.2.2 Auction Revenue Rights

In the presence of transmission congestion, the RTO collects some congestion revenue. We don't want the RTO to keep this money and earn a profit, because it might then have some incentive to create transmission congestion rather than alleviating transmission congestion.

The RTO typically gives the congestion revenue back to the holders of FTRs. If the RTO gave away the FTRs for free, then this would be an easy story to tell. But the RTO auctions off the FTRs rather than giving them away. So it then has to find a way to get rid of the FTR auction revenue.

What happens in the US is that RTOs give the FTR auction revenue away to the owners of transmission lines (because they are the ones who put capital into the transmission system and they should be rewarded for doing so) or to electric utilities (because their ratepayers in many cases paid for the transmission lines before these markets even existed, so they should get the benefits). Each of these entities holds an "Auction Revenue Right" (ARR), which is just a share of the total FTR auction revenue.

The figure below shows how the money flows between the holders of FTRs and ARRs. It can be summarized as follows:

Figure 10.4: The flow of money between participants in the market for Financial Transmission Rights and Auction Revenue Rights

Source: This image by S. Blumsack © Penn State University is licensed under CC BY-NC-SA 4.0 [4] 

• The RTO collects congestion revenue, which is must give away.
• The RTO auctions off FTRs, and collects revenue from that auction.
• The congestion revenue is then given away to the holders of FTRs
• The FTR auction revenue is then given away to the holders of ARRs.

In the end, this system is like the holders of FTRs buying those FTRs from the holders of ARRs, except that the RTO acts as a counter party to all of the transactions.

10.2.3 Contracts for Differences

In conventional financial market analysis, a contract for differences (CFD) is an agreement to exchange the opening and closing prices of some financial asset. In electricity markets, a CFD is a bilateral agreement in which one party gets a fixed price for electric energy (the strike price) plus an adjustment to cover the difference between the strike price and the spot price. This adjustment may be a positive or negative number.

CFDs are different than FTRs in two ways. First, a CFD is usually defined at a specific location, not between a pair of locations. Thus, CFDs are a tool principally for hedging temporal price risk - the variation in the LMP over time at a specific location. Second, CFDs are not traded through RTO markets. They are bilateral contracts between individual market participants.

CFDs may be defined as "one-way" or "two-way" contracts. A one-way CFD can have a couple of different payment mechanisms. First, a one-way CFD can be structured so that if the spot price exceeds the strike price, the seller pays the buyer the difference. Otherwise, there are no side payments. Second, a one-way CFD can be structured so that if the strike price exceeds the spot price, the buyer pays the seller the difference. Otherwise, there are no side payments.

A two-way CFD is just the sum of two one-way CFD and is basically a forward contract for electric energy. In a two-way CFD, the seller pays the buyer if the spot price exceeds the strike price; and the buyer pays the seller if the strike price exceeds the spot price.

Here is an example. Let's say that a generation company signs a 100 MWh one-way CFD with an electricity consumer. The strike price is $50/MWh, and the CFD is defined at the location of the consumer.

Let's first say that the LMP at the location of the consumer is $75/MWh. In this case, the generator would earn $50*100 = $5,000 in revenue from the CFD, but would then need to pay the consumer 100*($75-$50) = $2,500 under the terms of the CFD. So the generator's net CFD revenue would be $2,500.

Now let's say that the LMP at the location of the consumer is $40/MWh. In this case, there are no side payments and the generator's CFD revenues are $5,000.

10.3 Hedging with FTRs and CFDs

Thus far, we have seen that temporal risk can be hedged with Contracts for Differences. A one-way CFD can basically put a ceiling on the price of electricity. A two-way CFD is essentially identical to a forward contract for electricity at a fixed price. Locational risk can be hedged with Financial Transmission Rights.

In this section, we will see how a combination of CFDs and FTRs can be used to create a "perfect hedge" that shifts all temporal and locational risk. The end result of this perfect hedge is like a fixed-price contract at the strike price of the CFD, as long as the quantities of the CFD and FTR are equal to the amount of power being transferred from the source node to the sink node.

The table below outlines the perfect hedging model. We'll assume that there is a supplier located at node a, and a consumer located at node b. The supplier produces Q MWh in the real-time market and the consumer uses Q MWh. We will let F denote the size of a two-way CFD defined at the customer's node, and M denote the size of the FTR held by the supplier. The FTR is defined such that node a is the source and node b is the sink.

Let's walk through the rows of the table:

• The first row shows the spot market revenues and costs for the supplier and consumer.
• The second row shows the payments under the CFD, assuming the strike price is equal to p. Note that if the LMP at node b exceeds p, the generator pays the consumer the difference. If the LMP at node b is smaller than p, the consumer pays the generator.
• The third row shows the sum of payments to the supplier and payments by the consumer from the real-time energy market and the CFD.
• The fourth row shows the FTR payment. This is zero for the consumer because the supplier is assumed to hold the FTR.
• The fifth row shows the total payments for the supplier and consumer if the FTR is the same size as the CFD. Note that because of the CFD defined at node b, and because of the FTR, any payments involving the LMP at node b cancel each other out. This is because the payment stream for the supplier from the FTR and CFD move in opposite directions.
• The sixth row shows the total payments if the FTR, CFD and amount of physical production/consumption at nodes a and b are identical. In this case, all LMP terms cancel out. The supplier is paid the LMP at node a through the real-time market and pays the LMP at node a through the FTR. The consumer pays the LMP at node b and then is paid the difference between the LMP at node b and the CFD strike price p. All that is left is that the supplier earns revenue equal to the CFD strike price times output, while the consumer pays the same amount.

Now let's put some numbers into the example in the table. We will let Q = F = M = 100 MWh and let the strike price of the CFD be $50/MWh. The results are shown in the table below.

To explain what is going on in the table, we will walk through the supplier's revenues and the consumer's payments separately.

Looking at the supplier's revenues, we see the following:

• The supplier earns 100*LMP_a in the real-time energy market
• The supplier earns 100*($50 - LMP_b) through the CFD
• The supplier earns 100*( LMP_b - LMP_a) through the FTR
• When we add everything up, all of the LMP terms cancel (this happens because Q = F = M), and the supplier's total revenues are $5,000.

Now we will look at the payments by the consumer:

• The consumer pays 100*LMP_b in the real-time energy market
• The consumer pays 100*($50 - LMP_b) through the CFD (The consumer actually gets rebates from CFD if LMP_b > $50)
• The consumer has neither payments nor revenues through the FTR
• When we take the CFD payments plus the real-time energy market payments, the LMP terms cancel and the consumer's total payments are $5,000.