US power market outlook
Four veteran power market forecasters participated in a round- table discussion at a Chadbourne conference in June about the outlook for US wholesale power markets, which regions of the US offer the best opportunities for project developers, and what effect carbon controls are likely to have on the market when they are eventually imposed by the US government.
The panelists are Steve Dean, president of DAI Management Consultants, Art Holland, director of power and fuels price forecasting for Pace Global Energy Services, Michael King, a senior vice president and economist with NERA Economic Consulting, and Hugh Wynne, the senior utility analyst with the respected Wall Street research house, Sanford C. Bernstein & Co. The moderator is Keith Martin.
Upswing Through 2012?
MR. MARTIN: I saw an interesting statistic in a book that Bernstein put out last week called “U.S. Utilities: The Outlook for Power Market Prices and Profits 2006 to 2010.”
By 1999, reserve margins in the United States had reached 15%, which is considered the lowest prudent reserve margin. The building boom that was already in progress added 180,000 megawatts of new capacity over the five years from 2000 to 2004. That is against a total generating capacity in the United States — after this additional capacity was added — of 900,000 megawatts. US generating capacity increased by roughly a fourth in five years. That explains some of the trouble the independent power industry went through in the wake of the Enron bankruptcy. It also shows the boom-and- bust nature of our industry.
With that background, let me start with you, Art Holland. You told me during a panel discussion in January 2005 that the United States was about to move into the boom part of the cycle and that you thought the next boom might peak between 2010 and 2012. That was before oil prices surged. It was when gas prices were high. Is that still your view of what the current boom cycle will look like?
MR. HOLLAND: A couple of things have changed in the last 18 months. You mentioned oil and gas. If memory serves me, natural gas at the time we spoke was about $5 or $6 an mmBtu. Then we had Hurricane Katrina. Gas prices went up to $13 or $14 an mmBtu and have since declined from there.
The lingering aftershock of the storm is that there is more uncertainty today about gas prices. No one expected that gas prices would remain at $5 or $6 an mmBtu, but now there is even greater doubt. We recognize the need to bring in LNG, and we have made assumptions in our forecast about the number of terminals that will ultimately be built.
Capacity markets have also continued to develop in the United States. They should provide some incentives for people to build new power plants sooner than they would have otherwise.
The increase in gas prices means that we are likely to see fewer new gas-fired power plants than we would have other- wise, but because other technologies take longer to develop, that means the boom may not run its course as quickly as thought earlier. We expect over the next 20 years that as much as half of the new capacity will gas fired. Gas-fired capacity takes two or three years to build; other technologies take longer. Therefore, while we are moving into more of an upswing, I’m not so sure that it will peak in the 2010 to 2011 time frame. The cycle may take longer to play out.
MR. MARTIN: Steve Dean, Art Holland just said that he thinks as much as half the new capacity will be gas-fired. Do you agree with that assessment?
MR. DEAN: No. I think many more new plants will be base- load generation that will be coal or nuclear. This trend will be helped by the new incentives in the Energy Policy Act of 2005 for coal and nuclear. In the late 1990s, we built intermediate generation. There was no real base-load generation added to the grid. There is more interest this time around in base-load generation, whether it be coal or nuclear. However, what gets built during this phase may be driven by the changes in the regulation of air pollution.
MR. MARTIN: Mike King, one thing that seems to characterize the wholesale side of the electricity market in the United States is boom and bust, and 2004 seemed to be a pivotal year. It was the first year when demand for electricity increased more rapidly than capacity additions. Do you think we’re on the upswing now and, if so, how long will that upswing last?
MR. KING: I think we will see a significant amount of new generation built. I would call this a return to the era of King Coal. I expect that almost all of the new generation will be coal fired. We may see a few nuclear plants, but I don’t expect more than one or two. There are still doubts about the economic viability of nuclear plants in the long run without the government guarantees that will be offered to the first few plants. If coal is the dominant fuel in this cycle, then the cycle will take longer to run its course because it takes longer to build coal-fired power plants.
The real question is whether industry has learned when to stop building. This is not just an issue for independent generators. I was involved in the industry in the 1980s when utilities managed to over build as well. The question is whether we have learned our lesson from the last cycle.
MR. MARTIN: Art Holland, you believe as much as half of the new plants will be gas fired. Mike King, you are almost all coal. Steve Dean, you are also leaning more heavily toward coal. Art Holland, you look like the odd man out.
MR. HOLLAND: I am assuming that parts of the country that need capacity now will not be able to wait for new coal- fired power plants. There is also a lot of uncertainty about what new environmental controls will be placed on plants that use coal. These two factors make me think gas still has a significant role to play.
MR. MARTIN: Mike King, have you failed to take into account the cost of carbon controls in your forecast?
MR. KING: Carbon is definitely an issue. Carbon controls are likely to be imposed more rapidly in this country than I would have said a few years ago. I think there will be some new gas-fired power plants, but this is the era when we need base-load capacity. What are the technologies for base-load power? They are nuclear and coal. There are too many issues associated with building nuclear plants. That leaves coal, notwithstanding the expected carbon controls.
MR. HOLLAND: Let me stress that what I said is as much as 50% gas, not at least 50% gas. I put it that way because everything is on the table today. We used to think that gas-fired capacity was the answer to all of our problems, but there are so many uncertainties with respect to carbon and with respect to other environmental problems, the price of gas and the commercialization of new technologies that you really have to look at every technology today as a resource planner.
MR. MARTIN: Hugh Wynne, let me bring you into the conversation. I think this view that we are on an upswing assumes something about expanding spark spreads and higher electricity prices. Do you have a view on how long the current upswing will last?
MR. WYNNE: The most interesting thing about the current cycle in the northeast quadrant of the United States is that we are not getting new plant construction in time to address what I think will be a fairly severe supply and demand imbalance by the end of the decade. When you look at plants that are currently under construction — which are the only new plants that will be available in that time frame — the rate of capacity growth in the northeastern US is not keeping up with growth in demand.
The necessary consequence of that is that demand will consistently be supplied by incrementally more expensive power plants than are in the generating fleet today.
In the Atlantic seaboard, the additional supply will come from increasingly-expensive gas or even from oil-fired power plants.
That means peak power prices are likely to rise significantly. In off-peak hours and in the midwest, you will probably see a slightly different phenomenon, which is a shift from coal being onthemargintocombined-cyclegasbeing on the margin. That is a very significant shift in terms of cost. It is a shift from $35 a megawatt hour to, say, $60.
The critical fact for me is that plants are not being built. That means electricity prices will have to rise as we dispatch from more expensive plants. The effort to implement capacity markets in places served by the New England ISO and PJM is a reaction to this problem. They would like to try to create revenues today by holding forward capacity auctions to encourage developers to build plants now so that the plants will be ready in 2010 and 2011. The delay in getting those markets up and running suggests that we will see a spike in prices in 2010, 2011 and 2012.
MR. MARTIN: Hugh Wynne, continuing with you, what determines whether a particular region of the United States is a good market for wholesale generators? Is it spark spread, reserve margin or something else?
MR. WYNNE: I think it is the shape of the supply curve and particularly whether there are kinks in the supply curve that reflect a shift from one generation of technology to another. It is also the fuel that sets the price of electricity in a region and the outlook for that fuel price.
To take the second point first, markets like New England and many states in the Atlantic seaboard rely predominantly on gas-fired generation, and gas-fired generation tends to set the electricity price. The same is true of Texas. Gas prices are expected to fall over the period 2007 to 2010. If that proves true, then this will be a moderating influence on electricity prices in gas dependent markets.
There are large parts of the country like the midwest where coal is the dominant price-setting fuel, and the outlook for coal prices is, broadly speaking, more stable. But there may be upward pressure because of the cost of increasingly-strin gent emissions controls on coal-fired plants, possibly including CO2 emissions controls.
The second thing to look for, in addition to where fuel prices are headed, is the potential for a shift in technology from a relatively cheap technology to a more expensive technology. That shows up as a kink in the supply curve. For example, a shift from coal that has a variable cost of maybe $35 per megawatt hour to gas, which may have a variable cost closer to, say, $60 per megawatt hour, can occur because demand increases more rapidly than new coal-fired power plants can come on line. That sort of shift leads to an increasing marginal cost of supply and higher electricity prices.
MR. MARTIN: The supply curve reflects the price at which generators would supply a certain quantity of electricity. As the price goes up, more and more supply will be offered by the market. But the technology may affect how much is ultimately supplied at any given price, right? MR. WYNNE: Yes. What keeps electricity cheap in the midwest today is that the incremental megawatt hour produced tends to come in most states from a rather large, relatively efficient coal-fired power plant. However, by the end of the decade, our expectation is that those large 1,000- megawatt base-load coal plants with scrubbers will probably be running flat out at full capacity. At that point, the incremental megawatt hour will come from a smaller, 500- megawatt 40-year-old coal-fired plant without a scrubber that may have a generating cost that is $10 higher. Those plants will eventually be running flat out, after which the next incremental megawatt hour will have to come from a gas turbine generator, which — even assuming $8 gas — would an operating cost of about $65 a megawatt hour.
As the market is forced to dispatch incrementally more costly units, you climb the supply curve in gradual increments and are forced to pay progressively higher prices so those units can cover their costs of operation.
MR. MARTIN: Art Holland, reserve margins are slipping to 20% in the midwest and they have already fallen below 15% in the mid-Atlantic states. These would seem the most fertile ground for independent generators. Do you agree?
MR. HOLLAND: Let’s not lose sight of an important devel opment. All of the regions you mention are in a part of the country that is interconnected with the PJM grid. That higher degree of interconnection allows for a lowering of reserve margins while maintaining of the same level of reliability.
Those reserve margins sound lower than what we have observed. They may reflect only what gets reported to the North American Electric Reliability Council, or NERC. The NERC figures generally only include plants that are under firm contracts to utilities. The figures do not include mothballed plants. When guessing at future prices, it is important to take into account mothballed plants that can come back quickly when scarcity pricing starts to become an issue in a region.
When interconnectivity and mothballed plants are taken into account, we don’t think prices will spike as quickly as Bernstein does. Also, we think additional capacity will come on line along the mid-Atlantic seaboard fairly quickly.
MR. MARTIN: Steve Dean, what parts of country offer the best opportunity for generators looking to build new plants?
MR. DEAN: The west coast and the northeastern United States where the population density is greatest, but in those areas, you also have other issues. For example, both California and most of the New England states have moved to adopt state limits on carbon emissions. So you have parts of the country that will need additional electricity. They are making it hard to construct new coal plants. They probably don’t like nuclear. They are pushing for renewables, but that will not be enough. The real need is for base-load plants.
MR. MARTIN: Mike King, do you agree that the west coast and New England are the two parts of the country that offer the best opportunity for new power plant construction?
MR. KING: Yes. Let’s not lose sight of the fact that in places like California, you see many proposals by developers to build coal-fired power plants, but in states like Wyoming and Nevada that are close enough to supply their output to California.
Opportunity for Windfall Profits?
MR. MARTIN: Hugh Wynne, you implied in your earlier comments that the best opportunity for profit is to be a generator who uses something other than the price-setting fuel. For example, in regions of the country where coal sets the price of electricity, you are better off being a nuclear generator or perhaps a wind generator. In what parts of the country is coal the dominant fuel?
MR. WYNNE: The entire region
between the Mississippi River east to Ohio and western Pennsylvania is one where coal is the dominant fuel. As you move into eastern Pennsylvania and upstate New York, coal is still the price-setting fuel during off-peak hours, but during peak hours, the fuel is gas. As you move farther east into New England, it becomes predominantly gas. The other wholesale region where you tend to see a lot of coal-fired generation is the southeastern US, excluding Florida, but it would be unusual to see coal as the price-setting fuel for more than 50% of the hours of the year. Finally, in the upper midwest — Wisconsin, Minnesota, Nebraska, the Dakotas — coal sets the prices of electricity perhaps 75% of the time.
The gas-fired regions are the remaining parts of the country — the southwest, much of the west as well as Florida, the Atlantic seaboard and New England.
MR. MARTIN: Mike King, the opportunity to earn a windfall profit from using a fuel other than the one that sets electricity prices sounds good in the abstract, but it assumes fuel prices will remain in constant relation to one another, apart from wind where there is no charge for the fuel.
MR. KING: I think that’s a key issue. The way that you make the most money is to have some other plant setting the price. If the price for electricity it set by a $60 gas plant, that is pretty important to a coal-fired power plant because the $25 spread between the $35 cost of the coal plant and the $60 price set by the gas plant creates a sizable profit margin. Coal prices tend to be stable over the long haul. The big uncertainty is what will happen to gas prices.
I have been in this business now for 25 years. There have been many times during that period when people have been alarmist about oil prices. Wasn’t it Hubbard, the state geologist in Pennsylvania, who said in the 1880s that we had at most 10 years of oil left? These things are notoriously difficult to predict.
We are in an era when gas prices are persistently in a range of $8 to $10 an mmBtu or higher. That makes coal look awfully attractive. If gas prices were to drop to $3 or $4 or, God forbid, Standard & Poor’s downside case of $2.75 an mmBtu, then coal is going to look a lot more troubled.
We made a mistake the last time around. The supply surge that began in 1997 or 1998 and culminated in adding 180,000 megawatts of new generation was almost all gas fired. We know that we probably needed some base-load power plants
out of that. Now we are in a situation where many of our base- load plants are aging and will have to be replaced.
The 900-pound gorilla in the room that no one has acknowledged yet is what is the regulatory scheme in the United States. It is not just a question of gas prices in relation to coal prices, but also whether we will continue to rely on a competitive wholesale power market or whether utilities will own the next round of plants and put them into their rate bases.
MR. HOLLAND: Our expectation at Pace Global is gas prices will fall through the 2012, 2013 and 2014 time frame. We think gas prices will reach bottom sometime in 2015 and then ease back up. However, we do not want anyone to hang his hat on a particular gas forecast. We now build all of the forecasts that we publish quarterly as stochastic bands, and those stochastic bands for gas are extremely wide.
I agree with Mike King. While the relative prices of gas and coal and other fuels are important, the real opportunities for independent generators turn not only on the relative spreads between fuels, but also on scarcity pricing. That is the promise of forward capacity markets to take some of the boom and bust and provide some of that scarcity pricing to developers in advance of when the market is short electricity.
The high price of gas will take you to $100 a megawatt hour. Scarcity will take you to $1,000 or $2,000 a megawatt hour. The key is to watch how effectively forward capacity markets induce developers to build before low reserve margins start being reflected in higher prices.
MR. MARTIN: Steve Dean, we have been talking about how fuel prices might provide opportunities for independent generators. What are other key variables?
MR. DEAN: One factor is the type of carbon controls that emerge. Less than 10% of the capital cost of a coal-fired power plant built in the 1980s went to air emission controls. The mercury controls that the Environmental Protection Agency adopted earlier this year will add another 2% to cost by 2010. Carbon controls are likely to add another 4% to 5%.
MR. MARTIN: That is 5% additional cost to generate a megawatt hour?
MR. DEAN: Exactly. Someone running a combined-cycle gas- fired power plant faces no additional cost due to mercury controls and about half the cost from carbon controls as for a coal-fired power plant. Therefore, whether you use gas or coal, your costs will increase. If you own a coal plant in a gas region, your margin will be squeezed. That is not something that many people have taken into account adequately in their calculations.
MR. MARTIN: Are there other factors that generators should take into account in their calculations? For example, how quickly will new nuclear power plants add significantly to US generating capacity? Are renewables likely to soak up the entire load growth?
MR. DEAN: There are about 35 new nuclear plants being constructed today in other countries. The US Nuclear Regulatory Commission is hiring 4,000 engineers this year to gear up for expected new license applications. Many of the large nuclear utilities are gearing up to submit applications. We will see new nuclear plants. The US government is offering production tax credits and taking on cost-overrun risk as an incentive to build. Any cost overruns above 125% will be borne by the taxpayers.
Effects of Carbon Controls
MR. MARTIN: Hugh Wynne, you said during a very interesting call with institutional investors that carbon controls, which are expected to be imposed eventually by the United States, will add between $3 and $11 a megawatt hour to produce electricity from gas and between $8 and $28 a megawatt hour to produce electricity from coal. However, you tended toward the bottom end of that spectrum. Why?
MR. WYNNE: It was partly a matter of editorial style. It is hard to publish research reports that foresee a change in the price of power on the order of 100%. I don’t think the world works that way. Such changes tend to be mitigated by demand response. There are also substitutions in supply. This makes me gravitate toward the lower end of price forecasts.
Yet, even at the lower end of the spectrum, the changes are likely to be very significant. My point is that even if we accept a very conservative view of the cost of CO2 emissions, the impact will still be huge.
There is another important point. There are likely to be sources of CO2 emissions reductions — and, therefore, sources of CO2 credits — that are much less costly than those that are available to the power industry alone. In other words, it may be very expensive to build power plants that use coal but do not emit CO2, like IGCC plants, but it may be even more expensive to retrofit plants that currently emit CO2 in an effort to minimize their emissions. Further, it may be very expensive to substitute lower CO2 fuels like gas for coal. Those costs may be much higher than the costs that would be incurred by other sectors of the economy to reduce their CO2 emissions.
One of the beauties of the regulatory schemes that are being discussed in the Senate today is that they tend to encompass virtually all sectors of the economy.
The result is the incentives will be greatest for sectors of the economy that are the most wasteful users of energy to cut back first on their uses of hydrocarbon fuels. The cost of cutting energy use in those sectors will be less than if we tried to cut CO2 emissions by focusing on the power sector alone. Residential and transportation uses of hydrocarbon fuels will be curtailed first. That’s another reason why I think the cost increases in the power sector will fall at the lower end of the spectrum.
MR. MARTIN: Steve Dean, you are involved in valuing power plants in acquisitions. Are you taking into account the possibility of carbon controls and, if so, how?
MR. DEAN: Yes, we are. When we value power plants, we assume that a carbon tax will take effect in about 2010. We assign it a 50% probability. It adds 4% to 5% to the cost of a typical 1980s coal-fired power plant and about half that much to the cost of a combined-cycle gas turbine plant. We are also assuming that any coal-fired power plant will use activated carbon injection to control mercury emissions, which will add another 2% to operating costs.
MR. MARTIN: Mike King, you also do valuations. Are you taking the same approach as Steve Dean?
MR. KING: Yes. We believe carbon controls will be put in place later than 2010. We tend to look at these things in terms of the implications of an $8 price for an allowance, or the right to emit a ton of carbon emissions. How an $8 allowance price would affect the market is unclear. It is not as simple as saying the price to generate a megawatt hour of electricity from coal has just risen by $8. Most likely, there will be price pressure on natural gas because it will be a premium fuel. The point is you have to play out a string of consequences to figure out what effect carbon controls might have on the value for a particular plant.
MR. MARTIN: Hugh Wynne made the point during a call with institutional investors that the Senate Energy Committee is proposing carbon controls not be imposed on the generators who buy the fuel but on the suppliers of fuel. They are fewer in number. It is easier to require that allowances be purchased by them.
Another interesting point he made is that he doesn’t see the federal government just handing out allowances, given its budget problems. He thinks that this will work like radio frequencies where allowances are auctioned off by the government to help close the budget gap. Mike King, do you think that is the most likely outcome?
MR. KING: I think that is highly unlikely. MR. MARTIN: Why? MR. KING: If we look at the trading schemes that have been
put in place, we know of no circumstance where allowances have been auctioned. There are lots of political issues associated with carbon controls. We are in an environment where a 72% rate increase in Maryland led to a political revolt, fed largely by The Washington Post. Price increases cause political backlashes. We just don’t believe you will see politicians touching the third rail by making a carbon scheme look like a tax. We think they will have to grease the wheels for any sort of carbon scheme to be enacted by handing out allowance. A tax scheme just makes losers. It makes no winners. An allocation scheme might get some people behind the proposal.
MR. MARTIN: Steve Dean, what calculations should a savvy banker do when lending to a greenfield plant today that uses fossil fuels?
MR. DEAN: There are two issues for the banker. He has to assign a probability to how likely carbon controls are to be imposed. He has to make judgments about when they are likely to be imposed and at what level. In my mind, there is no doubt that such controls will add to the operating costs of both gas- and coal-fired power plants. The issues are how much and when.
MR. MARTIN: And your guess is a 50% likelihood? MR. DEAN: Yes, 50% in 2010. MR. MARTIN: Hugh Wynne, I think you believe carbon controls are almost a certainty, am I right? MR. WYNNE: My guess is 30% to 50% probability by 2010. A key fact that goes unrecognized with CO2 emission
limits is that there is no way for the power sector to comply with them other than by curtailing output or by substituting higher cost fuels. Adding pollution controls to older plants is not really an option because it is not economic to retrofit with the available technologies.
Therefore, if you want the power sector to produce less CO2 — and the same is probably true of other big industrial
facilities and certainly true of houses and cars — the only way they can do it is by using fewer hydrocarbon fuels. That means producing less electricity in the case of power. It means that imposing a cap on CO2 emissions is like imposing a cap on power output. Utilities will have a choice of reducing output or switching from coal to gas plants, but either way, the price of electricity will increase just as it has increased in Europe when CO2 emissions limits were imposed there.
The real question is who will benefit from that price increase. Will the benefit be a windfall profit for certain utilities, which is the way it has worked in Europe?
The only way it will not be a windfall profit in the United States is if the government sells the allowances. The proceeds of the allowance sale will give the government resources to mitigate the effect of the price increases. They could be used for something crudely popular like a tax reduction or rebates on fuel bills to consumers. They could be used to stimulate technology investments that reduce CO2 emissions. They could be used to mitigate the effect on certain sensitive sectors of the economy that might be disproportionately harmed from the emissions limits, like coal miners and owners of coal-fired power plants.
MR. MARTIN: Art Holland, I can’t help but observe that the last time you and I spoke, you pointed out that one lesson forecasters took away from the last merchant plant boom was to be less certain about their forecasts. You spoke today about stochastic bands. This reminds me of what tax lawyers do. We have a range of opinions — this “should” be how the law works, or it is “more likely than not” how it works, or there is “substantial authority” for your position. Is this the grim future for price forecasting — that you will come off sounding like tax lawyers? [Laughter.]
MR. HOLLAND: It depends on how you plan to use the forecast. A developer using such a forecast needs a most likely case, and we will provide that, and he also needs to take into account a worst case.
Meanwhile, a utility using forecasts for resource planning should use the full range of possible outcomes. We want it to look at the entire range of power prices, gas prices, coal prices, emission allowance prices and a number of different portfolio mixes and to assess how each of those portfolios performs across the entire range of possibilities. Most utilities go with the portfolio that provides the best protection on the downside even though it doesn’t necessarily produce the best return on the upside.