The United States is expected to have to take action to limit carbon emissions that contribute to global warming. A number of states are moving to limit such emissions on their own without waiting for the Bush administration to act. Some individual companies are under pressure from shareholders to disclose the expected future costs of carbon controls in their annual reports. The following are excerpts from a conference call that Hugh Wynne, US utilities analyst for the independent research house Sanford C. Bernstein & Company, LLC, held on April 13 for investors interested in the US power sector. The transcript is reprinted with permission from Bernstein & Company.

Likelihood of US Controls

MR.WYNNE: We believe there is an increasingly large probability — maybe 30%,maybe 50% — that emissions limits will be imposed at a national level in the United States within the next five years.

Our reasons for this conclusion are the following. First, a group of states in the northeast has already committed to limiting emissions. The group includes the New England states, New York, New Jersey, Delaware and Maryland, which together account for about an eighth of the population of the country. Second, California has already taken steps to limit carbon dioxide or “CO2” emissions from new automobiles sold in the state. In 2005,Governor Schwarzenegger issued an executive order setting limits on CO2 emissions statewide. There is legislation pending before the state assembly to make the emissions targets in the executive order legally binding. CO2 emissions limits in California would affect another eighth of the US population. Thus, these two state-level initiatives alone are expected to affect a quarter of the United States.

At the federal level, it is noteworthy that both the leading Republican and Democratic candidates for the presidential nominations in 2008, McCain and Giuliani on the Republican side and Clinton on the Democratic side, all favor limits on greenhouse gas emissions. The Senate has also begun taking action. Last year, the Senate passed a nonbinding sense-ofthe- Senate resolution supporting a national program of mandatory,market-based limits on emissions of greenhouse gases. Over the following 12 months, the Senate Energy Committee did extensive research on how to implement a program of emissions controls. It has published the results of some of that research in a “white paper” that can be downloaded from the Senate Energy Committee’s website. That white paper is called “Design Elements of a Mandatory Market-Based Greenhouse Gas Regulatory System,” and it goes into remarkable detail about just how such a system could be structured.

If a national program of emissions controls is put in place, our view is that it is likely to be an economy-wide one and will not single out individual sectors. While the power sector accounts for about 40% of greenhouse gas emissions, other sectors, including transportation and industry, are also important contributors, and it will be more efficient economically if those sectors are also covered by any emissions limits.

We believe that policy makers are likely to favor emissions caps rather than looking to reduce the emissions intensity of economic activity as proposed by the Bush administration, because the objective ultimately is to limit atmospheric concentration of greenhouse gases and that can only be achieved by capping and then reducing the CO2 output of the country.

It is highly likely that a large portion of the emissions allowances will be sold, possibly through an auction — much like access to the radio spectrum is auctioned — rather than be granted to emitters as has been done in the European Union. The reason we think that a sale of allowances is more probable than a grant is that governments at both the state and federal levels are keen to maximize their fiscal revenues. Other objectives include avoiding windfall profits by the emitters of greenhouse gases and raising resources with which to fund programs to protect consumers from energy price increases as well as to foment technologies that limit the output of greenhouse gases.

Finally, I think it is likely that to protect affected industries, we may find that any emissions control program put in place includes a safety valve that allows the government to issue additional allowances if the price of allowances exceeds a certain level. The idea is to cap the cost to producers and consumers of CO2 emissions limits.

Impact by Fuel

Figure 1 (page 16) shows the average US coal-fired power plant emits about a ton of CO2 per megawatt hour produced. A gas fired power plant emits a little over half a ton.

One can take the estimated cost of emissions allowances and multiply them by these factors to estimate the increase in the cost of producing electricity at coal- and gas-fired power plants. The range of cost estimates runs from about $8 to $28 per ton of CO2 emissions, $28 being the dollar cost per short ton of CO2 at which emissions allowances currently trade in the European Union. The lower estimate of $8 corresponds to studies and analytical work done in the United States about the cost of complying with a program such as one proposed by two influential US senators, John McCain (R.-Arizona) and Joseph Lieberman (D.- Connecticut). The reason for the difference is the less onerous emissions limits that are being proposed in Congress compared to those adopted in Europe.

Thus, to complete the picture, given an allowance price of $8 per ton of CO2, the cost of generating electricity at a coal fired power plant should rise about $8 per megawatt hour. The cost of generation at a gas-fired power plant should rise by approximately half that or $5 per megawatt hour.

CO2 emissions limits have important implications for the profitability of different types of power generators. In markets where gas-fired generators are the marginal or price-setting units, their cost of compliance with the CO2 emissions program will be reflected in the price of power. That would be true, for example, in markets like New

England, New York, the mid-Atlantic states and Texas. In markets where coal-fired generators are the marginal or price-setting units, such as the Midwest, their cost of compliance will again be reflected in the cost of power.

Because of the differential in the CO2 emissions of gas plants versus coal plants, power prices will likely rise more in coal dependent markets than they will in gas dependent markets. The implication is that coal-fired generators operating in markets that rely primarily on gas-fired generation, such as Texas or New England, may find their margins squeezed. If their cost of generation were to go up by $8, while the marginal producer sees his cost of generation go up by only $4, then the likelihood is that prices will rise by less than the operating cost of the coal plant, and its margin will be therefore be reduced.

Conversely, the obvious beneficiaries of any program such as this would be the nuclear power plants with no CO2 emissions that are operating in coal-fired markets, where the increase in the cost of generation will be the greatest, let’s say $8 per megawatt hour. That increase in the cost should drive a similar increase in price that would basically add directly to profit for nuclear generators. In gas-fired markets such as Texas or New England, that cost and price increase would be less, let’s say $4 or $5, but the impact would be the same, adding directly to profit for nuclear generators without any increase in their operating costs whatsoever.

The primary beneficiary we believe would be Exelon, which operates a large nuclear fleet in Illinois, where its

competitors are coal-fired power plants. Those plants would face the full cost implications of an $8 per ton allowance price, driving up the price of power in that market and driving up the margins of Exelon by a commensurate amount. We estimate the increase in Exelon’s EBITDA (earnings before interest, taxes, depreciation and amortization) from an $8-per-ton allowance price at 40%. There is no cost associated with that to Exelon, so that EBITDA increase should flow through to pre-tax profits.

Smaller but still very significant earnings gains could be expected at companies like Constellation, First Energy, Dominion, PP&L, Entergy and PSEG.

Their gains are less in part because these companies have smaller nuclear fleets, but also because those nuclear fleets are located in regions where gas-fired generators set the price of power and, therefore, the expectation would be that power prices would rise by less.

At risk on the other hand, are companies like TXU, Xcel, Reliant, Northeast Utilities — although its generating fleet is up for sale — Dynegy and NRG. In cases such as TXU and NRG, what we are talking about are the potential pressures on profitability created by operating coal-fired power plants at higher cost in markets where gas is setting the price of power.

Effect on Valuations

Let me move on to the second section of the presentation. Why worry about CO2 emissions limits now? What are the developments in state and federal policy that lead us to believe that this is becoming an increasingly likely initiative on a national level?

The northeast has already agreed to caps. The first state to enact CO2 emissions limits was Massachusetts in June

2002. Those limits took effect on January 1 of this year with respect to coal-fired power plants. In December last year, several states in the northeast — New York, New Jersey, Delaware, Connecticut, New Hampshire, Vermont and Maine — committed to limit their emissions of greenhouse gases under a regional initiative called the “Regional Greenhouse Gas Initiative” or “RGGI.”That initiative would cap greenhouse gas emissions at current levels from 2009 to 2014 and then cut them by 10% by 2019. In April of this year, Maryland passed legislation requiring that state to join RGGI in 2007. These initiatives have committed states to cap CO2 emissions in a region that encompasses an eighth of the US population.

In California by state law, greenhouse gas emissions from automobiles sold in the state are to be regulated commencing in 2009. The law mandates emissions cuts of 22% by 2012. The state public utilities commission, moreover, already requires utilities evaluating different types of power plants to add a carbon adder of $8 per megawatt hour when evaluating the economics of coal-fired power plants. The governor of the state, Arnold Schwarzenegger, in June of last year signed an executive order capping greenhouse gas emissions at 2000 levels. The executive order does not have the force of law. In April, a bill was introduced in the state Assembly that would make the governor’s emissions targets mandatory ones statewide. If that law is enacted and California adopts mandatory CO2 emissions limits, another eighth of the country’s population would be affected.

It is interesting to see that the electric generating capacity of the states that are moving to limit CO2 emissions is a much smaller fraction of the total than the population of these states is of the nation’s population. The states that have adopted CO2 emissions limits encompass about a quarter of the nation’s population, but they account for only about 14% of US generating capacity. Much of that capacity is in fact gas-fired capacity. These states account for about a fifth of the country’s gas-fired capacity, but only about 5% of the nation’s coal capacity. Perhaps these states are more comfortable imposing CO2 emissions limits because they are large importers of power and because the in-state generation that they are regulating is not a heavily-emitting fleet. It is a gas-fired fleet as opposed to a coal-fired fleet.

Federal action is increasingly likely. In April 2005, the federal Energy Information Agency published a study of a proposed CO2 emissions limit program, proposed by the National Commission on Energy Policy, and concluded that the proposed limits would not materially affect economic growth in the United States. That had an important impact on the debate around CO2 emissions limits and, in June 2005, the US Senate passed a sense-of-the-Senate resolution stating that global warming has a great deal of scientific evidence supporting it, that evidence indicates that global warming is being aggravated by man-made emissions of greenhouse gases and that something should be done about it.

Rather than simply passing that resolution as a bit of political drama, the Senate Energy Committee has followed through with extensive research and hearings on how to design and implement a program of CO2 regulation. The committee published a white paper on the subject. Following up on the various hearings that the Senate Energy Committee has had, one of which took place in early April, Senator McCain now plans to re-introduce his McCain-Lieberman “Climate Stewardship Act” that would cap greenhouse gas emissions at 2000 levels. The leading Republican and Democratic candidates for the 2008 nominations all favor restrictions on greenhouse gas emissions.

Thus, taking into account these state-level initiatives and taking into account the level of federal interest in both the executive and the legislative branches of government, we think that the probability of some type of national limit on CO2 emissions has increased appreciably and now needs to be factored into the valuations of power companies in this country.

The companies that would be most affected by CO2 emissions limits are taking the possibility of a nationwide cap quite seriously. The largest US power utilities and their lobbying group, the Edison Electric Institute, are already engaged in a dialog with legislators at the federal level about greenhouse gas emissions.

Moreover, a number of the largest utilities have expressed their support; Exelon and Duke are both advocating economy-wide limits on greenhouse gas emissions, supplemented by a system of tradable emissions allowances. Those are the two largest utilities in the United States. Other large utilities, like Edison International and Entergy, have taken similar positions.

Some utilities remain opposed to a mandatory program of emissions limits, such as AEP and Southern, but those utilities are also being very active in designing strategies to cope with a national program of emissions limits were it to be implemented. Most notably, AEP has recently obtained approval from the Ohio Public Utility Commission to recover in its rates the pre-construction cost of an integrated coal gasification, combined-cycle power plant that the company plans to build in Ohio. The choice of that technology specifically addresses the expected need to contain the emissions of CO2 from future coal-fired power plants.

Competing Proposals

Next, I would like to analyze the various policy alternatives that have been put forward and try to derive from that analysis an idea of what the future may hold for us by way of the structure of a greenhouse gas emissions program.

The first initiative to consider is the Regional Greenhouse Gas Initiative opted by the states in the northeast. It has several important elements. First, it is a cap-and-trade program similar to those in effect at the federal level for sulfur dioxide and nitrogen oxide emissions. Second, it only covers the power sector. It is not economy wide. Third, it allows member states to sell rather than grant allowances. Fourth, it has a safety valve mechanism of a sort designed to soften spikes in allowance prices.

Greenhouse gas emissions under this program would be capped at current levels from 2009 to 2014 and cut by 10% through 2019. Each state may allocate its greenhouse gas emissions budget as it sees fit, but the initiative requires that 25% of the allowances be set aside by those states for sale and the proceeds used for the benefit of consumers or to foment low-emissions power sources. There has been a lot of grass roots pressure put on legislators in these states to ensure that the amount of allowances that are sold is maximized in order to maximize the fiscal revenues available for compensation to consumers, to fund new technology initiatives and to prevent a windfall profit from falling into the laps of the very companies that are emitting the greenhouse gases. The final point is that the program is designed so that if the allowance price exceeds $10 a ton, generators may use offsets — credits from programs to capture CO2 through reforestation or to capture methane through landfill gas capture — to cover up to 20% of their emissions.

Moving on to the McCain-Lieberman “Climate Stewardship Act,” the approach is materially different. This is an economy-wide program: the only sectors of the economy that are not covered by McCain-Lieberman are the residential and agricultural sectors. It is similar to RGGI, on the other hand, in that it is a cap-and-trade program. It allows the government to sell rather than grant allowances. Under McCain-Lieberman, greenhouse gas emissions would be capped at 2000 levels for the years 2010 to 2015. The Secretary of Commerce is empowered to determine what share of allowances should be granted to emitters versus what share should be withheld by the government for sale. The proportion may vary year to year. That, in turn, serves the purpose of allowing the government to grant more emissions when the price is high. In addition, the legislation would allow emitters to satisfy up to 15% of their allowance requirements by purchasing allowances from other countries or by obtaining credits for reforestation and other carbon capture projects.

There is an important private sector initiative put forward by the National Commission on Energy Policy, which is a group that encompasses electric utilities, labor unions and academics. Its proposal has stimulated a lot of debate in Washington. The program recommended by the National Commission on Energy Policy is also economy wide. It also involves tradable allowances. Unlike RGGI and McCain-Lieberman, it does not seek to cap greenhouse gas emissions but rather to slow their rate of growth. Unlike RGGI and McCain-Lieberman, it skews the allocation of emissions allowances towards grants to emitters rather than encouraging the government to hold allowances back for sale. Finally, it specifically calls for a safety valve to avoid spikes in the price of allowances.

Specifically from 2010 to 2019, emissions targets would be set that would reflect a reduction in the rate of greenhouse gas emissions per dollar of gross domestic product or “GDP.” The idea is to reduce the carbon intensity of the economy, but not necessarily to cap the absolute level of greenhouse gas emissions. The effectiveness of that approach in terms of reducing the level of CO2 concentrations in the atmosphere is questionable. The advantage, of course, is that it would set less stringent limits that favor the utilities that are behind the proposal.

Another important element is that 95% of the emissions allowances would be issued at no cost to emitters. Over time, that would be reduced to 90% of the allowances being issued at no cost to emitters. Again reflecting the provenance of the proposal, this would maximize the profits of the utility sector.

To prevent price spikes, the government would sell unlimited amounts of allowances at a price of $8 per ton. Thus, it would be possible to quantify in advance the maximum price that emitters would incur to acquire emissions allowances and also to quantify the maximum price that consumers would pay as a result of the program.

Finally, the Senate Energy Committee white paper has set out some very thoughtful observations about the structure of an emissions limit program. One of the things that this white paper recommends is an economy-wide approach, but it also suggests that emissions limits be placed on the sellers rather than the consumers of hydrocarbon fuels. The way that would work is that anyone selling a hydrocarbon fuel, whether it is coal or gas or refined petroleum products, would be required to own allowances for CO2 emissions equivalent to the CO2 content of their fuels being sold. What would happen, therefore, is that regulation would take place at the upstream points — the coal mines, the gas gathering companies and the refineries. This would be more efficient than trying to regulate emissions at every final emitter’s tailpipe or chimney, particularly when a lot of those final emitters are automobiles or residences and the number of such emitters runs into the millions and the administrative difficulty of monitoring and limiting their emissions is exponentially higher. By capturing all sources of emissions, moreover, the upstream approach may stimulate a wider range of emission reduction responses and thus achieve emission reduction targets at a lower cost.

A government auction of allowances is favored over allocations. The reasons for that are to reduce administrative costs, to limit the potential for windfall profits to emitters and, finally and most importantly, to raise government revenues. Government revenues I estimate could be $50 billion to $175 billion annually from the sale of these greenhouse gas emission allowances. A portion of the proceeds from the sale of allowances would be set aside to provide technology incentives, to protect consumers and possibly also to compensate energy intensive sectors that would be materially disadvantaged by CO2 emissions limits. That would include potential compensation to electricity generators.

Taking into account these various initiatives, what can we perhaps foresee in a future program to eliminate emissions of greenhouse gases in the United States? Importantly, McCain-Lieberman, the National Commission on Energy Policy and the Senate Energy Committee all favor an economy-wide program over a sectoral program. That seems far more likely to me because it is the cheapest and most effective way to achieve emissions targets in the future.

I also think it is more likely that any such program would have emissions caps as proposed by RGGI and McCain- Lieberman as opposed to limits on emissions intensity, as proposed by the National Commission on Energy Policy, simply because limits on emissions intensity do not address the real problem behind global warming, and that is the level of CO2 concentrations in the atmosphere.

Both RGGI and the white paper produced by the Senate Energy Committee favor the sale as opposed to the granting of emissions allowances. That, I believe, has high probability because of the fiscal benefits to the government. McCain-Lieberman would delegate the decision about how much to sell and how much to grant to the Secretary of Commerce. Only the National Commission on Energy Policy, which includes utilities among its sponsors, favors the outright grant of allowances to emitters, allowing the emitters to enjoy the benefit of any windfall profit from the increase in the price of fossil fuels or electricity.

RGGI and the National Commission on Energy Policy provide for mechanisms to limit spikes in the price of allowances. I think it’s likely that something like that will be included in the final program simply because it will create a level of certainty regarding the cost of the program that both consumer and producer groups will find attractive.

Finally, many of these proposals — RGGI, McCain- Lieberman and the Senate white paper — call for a set aside of the proceeds from the sale of allowances to protect consumers, to promote technology development and to compensate energy-intensive industries.

Implications for Generators

Let me move to the final subject, which is the implications for US generators of limits on CO2 emissions. The implications are very much a function, of course, of the price of the allowances. In 2006, the price of CO2 allowances in the European Union has averaged about $28 per short ton. Estimates and studies done of the cost of compliance in the United States with programs such as that proposed by McCain-Lieberman tend to have much lower prices for CO2 emissions — about $7 to $9 per ton. We have used $8 therefore to estimate the cost to generators in the United States.

The impact of CO2 controls in the United States will fall much more heavily on coal-fired generators whose emissions of CO2 per ton - or per megawatt hour produced — are approximately twice those of gas-fired generators. The advantage that coal enjoys over gas will be eroded. How much it erodes will depend on the price for allowances. The gap will erode but not be eliminated. Figure 2 shows the range of possible erosion in the gap.

We looked at how the generation cost of individual utilities would be affected by CO2 emissions limits. Our conclusion is that the markets that are going to be most heavily affected by limits on CO2 emissions will be the Midwestern markets of ECAR and MAIN and to a lesser extent the mid- Atlantic market or MAAC, because these are markets where coal-fired generators set the price of power and, therefore, the

cost increase per megawatt hour produced will be greatest. Markets that will be somewhat protected because they are primarily gas dependent markets will be markets like Texas and New England. These are markets where gas-fired generators are the marginal producers and, therefore, the cost increase will be less.

In summary, the implications of a program of CO2 emissions limits will be, first, an increase in the price of power to reflect a cost increase to the marginal producer in the region of paying for the allowances. Second, there will be a tendency for coal-fired generators in certain markets to see their margins erode. That will be a problem primarily for coal fired generators that are operating in markets where gas fired generation is the price setting technology. Third, the greatest gains will be enjoyed by nuclear generators in markets that are heavily coal reliant, such as the Midwest and, to a slightly lesser extent, the mid-Atlantic. Gains will also be enjoyed, of course, by nuclear generators in markets that are gas reliant, but those gains will be less per megawatt hour because of the lower CO2 emissions of gas fired generators.

We calculated the erosion in gross margin that will be suffered by coal-fired generators that are competing in markets where gas sets the price of power. We then calculated the expected gross margin gains likely to be enjoyed by nuclear generators in markets where gas or coal is setting the price of power. Figure 3 shows the combination of the two. It shows the expected net effect on gross margins of unregulated generators in this country. The biggest beneficiaries will be companies like Exelon, operating nuclear fleets in coal-fired markets, and to a lesser extent companies like Constellation, First Energy, Dominion, PP&L, Entergy and PSEG, operating nuclear fleets in gas-fired markets. The biggest profit erosion will likely occur at coal fired generators operating in markets where gas sets the price of power, as would be the case, for example, with NRG in Texas.

Questions

AUDIENCE MEMBER: Duke has said it prefers a carbon tax because a tax would more fairly spread the cost of controls across all industries. What do you think of a tax versus a cap-and- trade system?

MR.WYNNE: The benefit of a cap-and-trade system is that to the extent a company can comply with the emissions reductions at a low cost, its compliance will benefit other emitters whose cost to reduce emissions would be much higher. The point is to find the cheapest way for the economy  as a whole to comply.

The concept is that if I am operating an extremely inefficient coal-fired power plant with a very high heat rate, and thus producing very high emissions of CO2 per megawatt hour, by shutting that plant down, I can reduce CO2 emissions at a cost to the economy that would be a great deal less than would be the cost to a highly-efficient facility with much lower allowances.

The classic example of how this has worked is what has occurred under the Kyoto treaty where the collapse of highly inefficient manufacturing industries and power plants in the former eastern bloc countries has meant that CO2 emissions in Europe have declined dramatically.

Those countries have been able to sell allowances for CO2 emissions that they are no longer using to western Europe where the cost of achieving similar emissions reductions would be much greater because western European countries are much more energy efficient— they are much cleaner in terms of the operation of their manufacturing industry and their generating stations.