Prices are Negative? That’s Un-possible!

As recently reported by Crain’s (, Exelon has a pilot arrangement with PJM ISO to reduce capacity by up to 10% of its nuclear capacity when wholesale prices of electricity were likely to go negative.  Seems fairly logical – who would keep going to work if you had to pay your company for the ability to work there?  But, is that really the whole story here?


In this article, I will be breaking down the market forces that are at work here to try to provide you with a better understanding of what’s going on here.  I’ll be up front; I have a bias here.  I don’t think Exelon should be cycling their nuclear power plants, and I don’t think wind power is to blame for their woes.  So, if you see me start going in that direction, at least you know my bias, and can compensate for it.  So, here goes.


Prices are negative!  That un-possible!

Screen Shot 2013-06-21 at 3.20.02 PM

In the market for wholesale electricity, prices do sometimes go negative.  It seems odd, but understanding a little more about how wholesale prices are determined should allow you to understand how this can happen.

Wholesale prices are determined hourly based on an auction for electricity that happens every hour of every day.  These auctions are managed by the Independent System Operator (ISO) or Regional Transmission Operator (RTO) that manages the high-voltage transmission grid covering an area.  In Chicago, we’re part of the PJM ISO (Pennsylvania – Jersey – Maryland Independent System Operator), while in the southern part of the state they’re part of MISO (Midwest Independent System Operator).

I’m going to have us go through roughly what one of these auction looks like using just three power plants.  In both PJM ISO and MISO, there are many more power plants than this, but this simplified example will get the gist of how the auction operates across.  So, we have the following power plants in our area:

a)     nuclear plant = 50 MW

b)     coal plant = 35 MW

c)     natural gas = 25 MW

During this hour, PJM ISO will contact all these power plants and tell them that they expect the amount of electricity needed during the next hour to be 100 MW.  Each power plant will in turn tell PJM ISO how much electricity they’re willing to supply during the next hour.  So, PJM ISO will end up with something like the following:

a)     nuclear plant = 50 MW @ 3 c/kW

b)     coal plant = 35 MW @ 7 c/kW

c)     natural gas = 25 MW @ 5.5 c/kW

(I’ve shown the prices in kW so it can compare to what we’re all used to paying for, when it comes to electricity.  Typically, these markets are prices in MW, but the conversion it fairly simple.  1 MW = 1,000 kW, and 1 kWh = 1 kW used over the course of an hour.  Conversely, 1 kW consistently used over 2 hours is 2 kWh, while 2 kW used over the course of an hour is also 2kWh.)

Well, PJM ISO only needs 100 MW, but 110MW has been bid.  How does PJM ISO figure out which bids to accept?  It ranks the prices of the bids starting with the lowest bid first, and includes as many bids as necessary to reach the demand.  So, in our example, the nuclear plant will provide 50 MW, the natural gas plant will provide 25MW, and the coal plant will provide 25 MW.  That’s how the demand is met, but how we determine the price?  The market price is the price of the last bid that is needed in order to reach the market demand.  In our example, the last bid had a price of 7 c/kW, so each supplier, regardless of what they bid, will be paid 7 c/kW for each kW they supply to the market for that hour.  So, the nuclear plant makes a little more than twice as much as it needs, the natural gas plant makes a little more than it needs, and the coal plant just makes what it needs.  (This piece will become important once we start looking at Exelon’s impact on the overall market supply and market price.)

So, now let’s take a look at what happens if the amount of electricity demanded by the market is very, very low.  Picture it being 3am on a Tuesday morning in April – everyone’s is asleep, and the heat and air conditioning are all off because the outside temperature is comfortable enough to sleep in without them.  Let’s face it; there’s very little electricity being used at that time, so demand is very low.  Let’s look at what happens to our auction when PJM ISO requests 10 MW of electricity to supply the entire market.  The decision process becomes a little more involved here, so let me take you a little deeper into what each plant operator is thinking at this time.

a)     Nuclear plant = Nuclear plants have essentially no fuel cost, and are very difficult and expensive to shut down and power up.  Some nuclear plants may take as long as 6 weeks before they’re ready to be powered up if they’re shut down!  (It depends on the nuclear technology involved and on the age of the plant.)  So, the plant manager really doesn’t want to shut down his plant, because he really wants to be able to sell power later on that day.  He knows that if he just bids the 10MW that PJM ISO requested, he’ll probably be made to shutdown part of his plant.  (Another issue with a nuclear reactor is that it doesn’t like to operate at much less output than its max output.  It’s pretty close to an all or nothing deal.)  So, he needs to make sure that he can put out full power, and he need to bid a price that will get him there.  If he pays people to use power, PJM ISO won’t make him shut down his plant.  Nuclear plant = 100 MW @ -0.1 c/kW

b)     Coal plant = Coal plants have a fuel cost.  The burn a certain amount of coal for every MW of electricity they produce.  But, the operator of this plant is smart.  She knows that the nuclear plant is going to low-ball the bid.  She also knows that the market price won’t be enough to cover what it would cost her in coal.  So, she throws out her normal bid, but she’s fully expecting to be shut down during the next hour.  Coal plant = 35 MW @ 7 c/kW

c)     Natural gas = Natural gas is in essentially the same boat as the coal plant.  There are two significant differences between the two types of plants.  First, natural gas plants are able to ramp their power output in less time than a coal plant.  (It doesn’t make a difference in this instance, but it does play a significant role in the overall market dynamics that are beyond the scope of this article.)  Second, the cost of natural gas vs the cost of coal.  (Over the last three years, the added natural gas production has allowed natural gas plants to be consistently less expensive than coal plants.)  Natural gas = 25 MW @ 5.5 c/kW

And, the winner is … nuclear plant @ -0.1 c/kW!  And, that is how the price of electricity is sometimes negative.


How hard does the wind blow?


Exelon is claiming that wind is increasingly the reason that electricity prices go negative overnight.  Let me walk you through the logic as to how this could be the case, and then provide some statistics to demonstrate how it’s actually very unlikely to be the case in Illinois.

So, let’s see how our low demand example from above changes if we include a wind plant.  I’ll skip the discussion of the coal plant and natural gas plant at this time, since they weren’t going to bid in the first case anyway.  PJM requests 10MW of electricity for the hour, and this is how the nuclear plant and the wind farm respond:

a.)   Nuclear plant = The factors are nearly identical to those above, except that the nuclear knows that the wind farm will bid really low too.  So, the plant manager is going to try to underbid the wind farm so it can keep producing during that hour.  Nuclear plant = 50 MW @ -2 c/kW

b.)   Wind farm = The wind farm has no fuel costs (wind is free!), but it does get some money from the federal Production Tax Credit and from selling RECs.  The federal Production Tax Credit provides a tax credit to wind farms as long as they produce a certain amount of electricity.  RECs are the greenest of the electricity produced by the wind farm, and they are sold to people who wish to claim their electricity as renewable.  (See this article for a deeper discussion on RECs –  So, the manager of the wind farm knows that as long as amount they get from the Production Tax Credit and from the RECs is more than the “cost” of they have to pay people to take the electricity, then she can still be making money.  Wind farm = 10 MW at -3 c/kW.


And the winner is… the wind farm and the nuclear plant!  Wait, the wind farm’s price is lower than the nuclear plant.  Why didn’t the nuclear plant win the entire thing?  So PJM ISO is pretty smart too.  It knows that the wind farm is dependent on the wind blowing, and that if the wind stopped blowing in the middle of the hour, then there would be a greater demand for electricity than what would be available.  And powering up the nuclear plant at a moment’s notice isn’t going to work.  So, PJM ISO has the nuclear plant and the wind farm share the load for the market price of -2 c/kW.  Including the value of the Production Tax Credit and the sale of the RECs, the wind farm is still making money.  Don’t feel too bad for the nuclear plant.  It seems like its losing money right now, but remember our “high demand” scenario where the nuclear plant was getting 7 c/kW?  That scenario happens is less than 12 hours away.

Exelon is complaining that the subsidies grant to wind farms give them an unfair advantage and should be removed.  (To be honest, I’m not completed opposed to letting the wind subsidies expire at the end of the year.  Wind is competitive with fossil fuels in many markets.  Which means, the subsidies for wind did exactly what they were supposed to do.  They stimulated innovation and scale so that a new technology could compete against the incumbents on equal footing.  But, let’s remove the subsidies from nuclear, natural gas, and coal as well, so that they can all compete on equal footing.  But, I digress…) But, is that actually what’s happening in Illinois?  Let’s take a look at the latest available generation numbers from the Energy Information Agency, part of the U.S. Department of Energy, for February 2013 (

Coal                                    = 46.5%

Nuclear                                    = 45.2%

Natural gas                  =    5.5%

Wind                                    =    5.2%

(Biomass, Other Gas, Oil, and Other make up the balance, in decreasing order of percentage)

So, how is it that 5.2% of the electricity generation in Illinois that’s coming from wind having such a profound impact on Exelon’s nuclear plants that they can’t afford to keep them operating?  Something smells fishy, and it’s not the wind farms.


Impact of Idle Nuclear plants on the PJM ISO


Let’s ask ourselves for a second what the impact is of idling a nuclear plant or two in the PJM ISO.  Well, the most obvious point impact is that there’s now less supply.  Going back to our Econ 101 class, if demand remains constant, but there’s and decrease in supply, then there will be an increase in price.  That seems pretty straight forward.  But, where’s the biggest impact on price going to occur?

Take our “high-demand” and “low-demand” examples from above.  These are two snapshots for the 24 hourly auctions that are run by PJM ISO.  These are the two extreme cases, with the demand in the remaining 22 hours sitting somewhere in between.  Imagine that each of our “plants” actually represented dozens of plants that shared those similar characteristics.  If we assume that Exelon plans to idle 10% of the nuclear “capacity”, then the amount of electricity coming from the nuclear plants will be 10% less.  If you run the example for all 24 hours in the case of full nuclear capacity, and re-run the example for all 24 hours with a reduced nuclear capacity, you’ll see that prices increase for all hours of the day.  Not only that, but you’ll see that coal, the most expensive and environmentally harmful of all the plants, produces a much larger percentage of electricity.

I’m not saying that Exelon is idling nuclear plants in order to create more pollution.  That just doesn’t make sense.  But, what I am saying is that Exelon owns a large number of nuclear power plants in Illinois, Pennsylvania and New Jersey, which are also serviced by PJM ISO.  (The “P” in PJM is Pennsylvania, and the “J” is Jersey.)  The cost of idling a few nuclear plants is more than covered by the increase in revenues Exelon’s other nuclear plants will experience due to the higher market prices.  So, Exelon as a whole will make more money.  Exelon is by far the largest owner of nuclear generation in PJM ISO, and I’d hate to see it take advantage of its market position in order for it to make greater profits.

And, I am vehemently opposed to it if those higher profits are as a result of me having to pay more for my electricity!


Read my lips Exelon – No nuclear taxes (increases) on my electricity bill!


[As an aside, I find it very funny that until 30 days ago, Exelon was asking the Nuclear Regulating Commission for permission to increase the output at some of its nuclear plants ( Very odd that these two things would be happening at essentially the same time…]

Related posts: