Carbon Wargames


U.S. utilities gain strategic insights by playing out a carbon-constraint scenario.

Fortnightly Magazine - December 2007

The U.S. power industry stands frozen in the headlights of the onrushing worldwide CO2 crisis. While power demand continues growing, and reserve margins continue shrinking, uncertainties in long-term natural-gas prices, carbon regulation, and clean-technology alternatives create a prohibitive level of investment risk for most power companies.

Moderate changes in the costs of gas or carbon credits can have a dramatic impact on the economic viability of coal-fired power plants (see Figure 1). If coal plants shift positions with gas plants on the dispatch curve, the economics of all generation sources will be affected profoundly.

In the present state of policy uncertainty, companies are prudent to try muddling through. A sharp drop in planned generation investments between 2005 and today suggests many companies in America’s power industry have adopted this approach (see Figure 2).

But the industry can’t postpone investments to meet growing power demand and replace aging capacity indefinitely. Although the current RPS-driven investment in renewable-power capacity is useful and important, it is at best a very partial response. Society’s long-term power needs dwarf the current scope of “no-regrets” moves.

It’s not too soon to begin assessing the long-term strategic options that may arise if and when the expected public-policy carbon intervention occurs. That intervention likely will redefine the generation playing field for the next several decades. What a particular company’s options might be in such an event will depend not only on the particular policy that is imposed, but also on the responses that other power companies might make. The collective responses of these companies in the short term will affect power prices and emission-allowance prices, and in the long term the further evolution of public policy.

For companies that must make multi-billion-dollar investment decisions in generation over the next several years, it’s important to explore how these dynamics might play out. For this reason Booz, Allen & Hamilton recently hosted an emissions “wargame,” drawing on a technique the firm has long used with the Pentagon, to explore how competitive (or even partially cooperative) actors might respond to each others’ moves when stressed by large, multidimensional, and uncertain changes.

A wargame provides an efficient and intuitive way to begin the process of thinking through the strategic dynamics those actors will encounter. It challenges players to think systemically, and yields insights into the issues that can operate, often in unexpected ways, to validate or to undermine a given strategy.

The Set-up

The carbon wargame began with a view of reality that was radically simplified for “playability,” yet faithful to the principal drivers of future strategies (see sidebar, “Designing the Game”). For example, players’ actions occurred in two mega regions—Coal Land, where coal-fired generation sets the normal marginal cost, and Gas Land, where gas generation sets the normal marginal cost. Of course, such a simplified model is not intended to predict future outcomes in the real world, but rather to identify likely future issues and to inform the strategic perspectives participants will bring to those issues when they arise.

A valuable game, or series of games, could be built around the effects of different types of climate regulation. Because the purpose of this particular game, however, was to illuminate industry strategy issues rather than public-policy issues, the game used a single set of regulatory assumptions, reflecting a rough mid-point among the more stringent carbon reduction proposals currently before Congress (see sidebar, “Regulatory Assumptions”).

The players were five fictional power generators, mostly integrated utilities with a blend of regulated and unregulated generation portfolios. These players interacted with a regulatory team and (in the background) a control group. The aggregate portfolio of all players mirrored the U.S. national generation profile. Players operated in a closed system in the sense that their moves alone set power prices and determined industry emission levels. The object of play for each team was to maximize shareholder value, consistent with federal policy strictures and as permitted by state and federal regulators.

Play began in 2009, with the hypothesized enactment of a national scheme of carbon regulation. The game proceeded in three seven-year moves—the first spanning the period 2009 through 2015, the second 2016 through 2022, and the third 2023 through 2029. As players added or retired generation, market prices shifted. Meanwhile, emissions credits were bought, banked, or sold. Player balance sheets and income statements changed accordingly. At the end of each period of play the game-control group calculated net change in carbon emissions compared with 2009, and the annual total shareholder return (TSR) from 2009.

Move 1: Green Bandwagon

At the risk of over-simplifying a complex set of moves and strategies, the three stages of play seemed characterized by three different strategic impulses. In the first move (2009 to 2015), the main strategic impulse was a “rush to compliance.” The principal strategic actions were:

• Strong investment in demand-side management (DSM);
• Strong investment in renewable generation;
• Selective initiation of nuclear construction;
• Substantial retirement of coal facilities; and
• Major reliance on new gas-fired generation to meet near-term load needs.

While none of these moves is startling, the combined effect—an overall reduction of industry carbon emissions of nearly 20 percent (well in excess of mandated levels for that period), and an average annual real TSR exceeding 4 percent—was notable. As a consequence of this overall success in compliance, the industry finished the first move in a long position with respect to emissions credits.

The downside of this picture was that wholesale prices roughly doubled in real terms over the seven-year period, both in Coal Land and Gas Land. Notwithstanding substantial investment in renewables, capacity additions lagged demand growth. While retiring a number of coal facilities, companies remained cautious about large-scale investment in compensatory new capacity. As reserve margins tightened, unregulated assets yielded exceptionally strong returns. The stage was set for a correction.

Move 2: Capacity Correction

In the second move (2016 to 2022), that correction arrived with a vengeance. The dominant strategic impulse was plant construction to replace the capacity lost as a consequence of stage-one coal-plant retirements. The principal strategic actions were:

• Continued nuclear investment;
• Major investment in IGCC plants;
• Significant ramp-up of renewables investment; and
• Significant ramp-up of DSM efforts.

This strong wave of investment in low-emissions assets, along with continued retirement of non-reformed coal plants, pushed overall emissions levels well below the target for that period, and also lowered financial returns.

For purposes of this game, technology risk was assumed to be minimal, and companies therefore invested in IGCC and renewables without significant fear of cost overruns, construction delays, or technology inadequacy. Even without this risk, average TSR fell substantially.

The collective surge to investment contributed to that result, as the market moved to a condition of overcapacity. Price increases moderated, so that wholesale prices grew less than 10 percent in real terms during this second stage of play. The over-shooting of emissions mandates left the industry holding significant quantities of banked carbon allowances.

The impulse behind this stage-two overbuild is familiar to anyone who has been part of the real-world power industry over the past decade. High price signals at the end of Move 1 yielded high spreads between marginal cost and revenue. Those short-term spreads invited utilities to make long-term investments in Move 2 that were far less profitable than hoped when the aggregate capacity caught up with demand—a demand that the same companies were simultaneously spending lavishly through DSM to reduce.

Moreover, the market rules in the game provided capacity payments that varied inversely with reserve-margin levels. At the beginning of move one, those capacity payments offered insufficient incentive to invest. At the beginning of move two, they augmented a message already conveyed loudly by market signals, and simply reinforced the already over-eager urge to invest.

Move 3: Diminishing Returns

In the third move (2023 to 2029), the industry returned to basics. The principal strategic actions were:

• Moderate shift back to coal investment;
• Attempted monetization of banked allowances;
• Diminished investment in renewables and nuclear power; and
• Discontinued DSM efforts.

Emissions during this period continued declining as earlier investment in renewable and nuclear power yielded significant production. However, financial returns for the industry also continued declining. Capacity proved more than adequate to the demand, thanks in part to large stage-two capacity investment and in part to substantial demand reduction.

Each company followed a different path, based on its starting portfolio and its particular strategic decisions (see sidebar “Two Hypothetical Utilities: Different Approaches, Different Results”). Overall, however, the players charted a trajectory of steadily decreasing emissions and steadily decreasing financial returns. Beginning with stage two, an unfavorable gap opened between shareholder return and shareholder cost of capital—a gap that steadily widened.

Paradoxically, two artifacts of the game that appeared advantageous to companies at the time probably contributed to these declines. First, today’s new technologies were assumed to work. A few early mishaps and disappointments might have tempered the companies’ enthusiasm for investment, although obviously complicating their efforts at compliance. Second, while capital of course had an economic cost, capital markets remained sanguine throughout the period of play and indulged company investment plans. The brakes that such a market might have imposed on overly ambitious investment plans never were applied. In the real world, technology uncertainties and rating agencies might reasonably be expected to restrain investment to a considerable degree.

The industry’s generation portfolio changed significantly between 2009 and 2030 (see Figure 3). During this 21-year span:

• Net capacity increased by 45 percent;
• Renewable generation increased from less than 1 percent to 18 percent of the total portfolio;
• Coal remained stable as new investment balanced retirements. However, coal-based emissions declined significantly due to CCS retrofitting; and
• Significant investment occurred in IGCC and nuclear.

Prices roughly doubled in both Coal Land and Gas Land, with most of the increase occurring during the first rush-to-compliance stage. The sharp, temporary increase observed in Move 2 resulted from an exogenous spike in the price of natural gas. While this exogenous development had only a temporary impact under the circumstances designed into this particular game, it served to demonstrate the vulnerability to world price movements of a wholesale-power market heavily dependent on natural gas generation.

Additionally, simultaneous actions by several players collectively reshaped market circumstances with each move. This served to mitigate the success of any single player’s strategy. The tendency to follow similar build cycles led to excesses in capacity. Similarly, a uniform move toward compliance through generation investment rather than through purchase of emissions credits left players in a situation where each had a surplus of allowances to sell, with no industry buyers. (A cross-industry market in allowances was assumed, which permitted players to unload their allowances at a discount.)

Real-World Dynamics

No single game offers a sound basis for prediction. The time constraints of game play, the inevitable sketchiness of competitive information, the simplifying assumptions employed in the particular game, and the number of variables that each team nonetheless must manage, yield results unique to any particular game. Future iterations of this or similar games might yield significantly different outcomes.

The current game does suggest, however, that significant reductions in industry-wide carbon emissions are feasible, at least under favorable technology assumptions. The financial consequences for the industry are disturbing, and multiple-player interaction will be an important strategic factor in companies’ calculations.

Several additional dynamics emerged in the game.

• Retail rate shock: Retail price increases varied by region and by company. At the low end, real increases averaged only 0.2 percent a year compounded—a relatively modest increase, although a reversal of the real-price decline U.S. consumers have experienced, and have become used to, over the past 20 years. At the high end, real increases exceeded 5 percent a year compounded. This level of increase, experienced particularly in those regions where coal plants initially set the marginal cost of power, would pose significant regional economic concerns, impacting particularly the energy-intensive industrial base currently located in coal-dominant regions.

In the real world, an effect of this magnitude—affecting different sections of the country with vastly different levels of severity, in the service of a generic social benefit of global scope—would pose enormous challenges for public policy, extending well beyond the purview of traditional regulatory institutions. In contrast to the generally accommodating posture adopted by regulators in this game, the regulatory attitude created in real life by this level of rate increase almost certainly would be skeptical and severe. The unfavorable pressure on shareholder returns already indicated by the game would be exacerbated.

• A steeper supply curve: Apart from the obvious overall price increase, costs increased sharply at relatively low levels of demand. “Mid-merit” generation shrank, and the supply curve became almost binary—either base or peak. Under such circumstances in the real world, significant price volatility would emerge at even moderate demand levels—significantly increasing the opportunities for traders and the risks to spot buyers. Moreover, the value of load-shifting initiatives would diminish, as most reductions in the highest peak demand would have only modest impact on marginal cost.

• The role of demand: Initially, DSM programs represented an important compliance strategy, used extensively by each company serving retail loads. Over time, however, demand levels were affected less by these interventionist initiatives than by the straightforward economics of high prices. Under the game’s long-term price-elasticity assumptions (on average, a 10 percent decline in power consumption for a doubling of power price), by 2030 the impact of power price on demand was roughly four times the impact of DSM initiatives. While utility DSM investments typically are compensated through regulatory processes, price-driven falloff in demand typically is not. How regulators, legislators, and the industry will accommodate the potentially serious impact on utility revenue—particularly under circumstances when ratepayer tolerance for compensatory rate increases will already be strained—will be a dominant issue in any carbon-constrained future.

• Alternative energy: The market forces operative in this game brought the portion of renewables in the industry’s overall generation portfolio to 18 percent, without recourse to explicit renewable portfolio standards (RPS). With growing deployment, scale-based cost efficiencies made renewable technologies increasingly attractive. The potential of such technologies for distributed rather than concentrated application, combined with the rate pressures noted above, could fundamentally reshape today’s utility model—plunging it into an active intermediation role among myriad suppliers as well as myriad users of power.

Endgame: Strategic Questions

Typically, games of this kind aren’t one-shot efforts. Instead, by varying the fundamental assumptions, introducing different “game events,” observing the strategic patterns that emerge and adjusting individual strategies accordingly, players gain a progressively richer picture of the strategic landscape ahead. That picture informs the modeling, the options analysis, and the scenario hypotheses that frame robust strategic analysis and foresight.

The results of this single game nonetheless tend to validate the hesitations of today’s utilities. Uncertainty regarding carbon regulation makes it difficult to place big generation bets today. Allowing sound tactical incrementalism to narrow a company’s strategic vision, however, would be a mistake. While the game justifies current caution, it also conjures a future of fundamental changes. Industry leaders who hope to capture the opportunities accompanying those changes need to stretch their organizations’ strategic horizons.

As a start, companies can ask five questions:

1) How could different potential regulatory schemes affect our particular existing portfolio? To what degree will that impact depend on how other portfolios are affected, and on how those other portfolios respond?

2) What are the actual technology choices we will face for our portfolio? What estimates can we make now about evolution in the cost and efficacy of those options?

3) How can we build our portfolio planning around the kinds of uncertainty we face? How should we identify the option value of particular investments, and how do we initiate and manage those options to take maximum advantage of timing opportunities?

4) How will the forces set in play by carbon regulation change our relationship with customers? Higher prices, ramped-up DSM, technology improvements, and carbon awareness inevitably will complicate that relationship—likely pushing the utility away from its current role as a pure deliverer of power toward a role as manager and facilitator of myriad customer power-related choices. Will we have the tools and skills needed to be as good at that job as we have been at delivering the commodity?

5) What regulatory approach will best accommodate the tangled public and private purposes at issue? Investor-owned utilities always have served a public interest, but that public-private nexus has been managed within a reasonably straightforward company-specific economic framework. Carbon control imposes a far more diffuse public purpose that will strain today’s regulatory institutions severely. Without early discussion of how the strains might play out, and what institutional changes will help manage those strains effectively, the industry is heading for a highly contentious future.

It’s a good bet federal carbon regulation lies in the industry’s future. Already the prospect of that regulation casts a shadow over all long-term generation and infrastructure planning. The carbon wargame illustrates the magnitude of discontinuities introduced by such regulation, and the location of particular pitfalls. Also, it offers hope that some winning strategies exist, new opportunities will be created, and deep emissions reductions are achievable. Above all, it suggests how thoroughly the industry’s strategic challenges will enmesh with the fundamental social choices that lie ahead.