Is the electric sector capable of rapid, large scale reform? Many policies implicitly assume the answer to that question is No, especially when it comes to greenhouse gas (GHG) emission control.

The result is a policy conversation that hinges on the assumption that it is hard to change. How much must we spend to accelerate new technology? How many decades should we allow for a phase-in of new regulations?

As it turns out, the industry can change — and indeed, has changed — at a much faster pace than you might think. Contrary to conventional wisdom, it turns out to be quick and fairly painless to replace meaningful fractions of our power fleet in very short time frames.

Why should that be surprising?

The electric sector is arguably among the most regulated part of the U.S. economy. From municipal light boards to state utility commissions to the Federal Energy Regulatory Commission (FERC), there are layers upon layers of regulatory bodies designed primarily to ensure electricity reliability and cost recovery for what have historically been monopoly franchises. What those bodies were most certainly not created to provide is a rapid rate of change.

By and large, those bodies have delivered on their promise. They’ve kept the lights on, kept electric utility profits low enough to protect consumers but high enough to attract capital, and maintained a fairly sleepy industry with very little default risk, virtually none of the “creative destruction” that idles assets in competitive industries, and virtually no significant technological innovation. (The power plant serving your town today not only looks like the power plant that served your town 50 years ago, but most likely is the same power plant.)

While these regulations have maintained predictability within the regulated industry, they have not prevented innovation and change external to the industry. Like flood levees, these regulations have kept the external weather at bay — but they haven’t changed the weather. From new generation technologies to smart grids to emerging concerns about the environment, volatility outside of the regulated enterprise has been persistent, invisible to customers of regulated utilities only to the degree that the regulatory levees hold.

Every once in a while, the levees are breached, exposing regulated markets to the volatility those of us who live in normal markets have come to take for granted. Perhaps unsurprisingly, those historic events have brought about the most dynamic periods in the industry. GHG regulation is, without question, a massive external change to the regulated enterprise. As such, rather than presuming a static, lumbering industry response, we ought to be looking at what happened the last time external changes breached the regulatory levee.

Specifically, let’s look at two recent events: the advent of wholesale market competition in the late 1990s and the creation of capacity markets in New England in the early 2000s.

1992 EPACT and FERC 888

In 1991, the U.S. had 581 GW of combined coal, natural gas, and nuclear capacity (307 GW coal, 174 GW natural gas, 100 GW nuclear). New additions were essentially zero, as the combination of Three Mile Island, the Clean Air Act, and a high fleet reserve margin gave little incentive for new construction. Meanwhile, a broader political push for deregulation was afoot. Into this environment came the 1992 federal Energy Policy Act (EPACT), which — among other things — provided full market access for any electric generator. (Previously, such access had been limited to regulated utilities and the narrow suite of technologies allowed under the 1978 Public Utility Regulatory Policy Act, or PURPA.)

After EPACT became law, there was essentially no discernible impact on new generator deployment; by 1995, we still had 100 GW of nuclear, had 311 GW of coal, and were up to 196 GW of natural gas.[1] It became apparent that while generators were now allowed to sell into deregulated power markets, access to the transmission grid — which was still largely controlled by regulated monopoly utilities — was being constrained for non-utility generators. FERC responded with Order 888, mandating non-discriminatory access to the transmission system for all power plants in 1996. That ruling was contested in the courts, but became final in 1998.

Within just 10 years after the final implementation of Order 888, nearly 200 GW of new generation capacity was added to the U.S. power grid, or 20% of the entire fleet. Nearly all was natural gas. This is a remarkable statistic: having taken nearly a century to build the first 800 GW[2] of total U.S. generation, it took us just one decade to build an additional 200 GW. Moreover, our generation fleet, which had to that point been dominated by coal, was now dominated by gas.

US installed generation capacity by fuel type

This is a massive rate of change in any industry, but especially in one that is supposedly resistant to quick change. Today, we take it for granted that much of our power grid is gas-marginal, but it was not self-evident that this would happen in 1995 (or, for that matter, in 1991). Arguably, we didn’t even have the lens to contemplate this type of change prior to deregulation.

Note, after all, that the big, capital intensive plants that had historically been built by regulated utilities (coal and nuclear) weren’t built prior to EPACT/888 and weren’t built after. In that narrow sense, our belief that the industry was incapable of quick change was correct; what we failed to recognize was the scope of innovation that would occur once new players entered the industry. Those 200 GW of new gas plants were built largely by unregulated companies with fundamentally different appetites for risk than the companies that had heretofore dominated the space. And while many of those new entrants subsequently ran into financial constraints, it bears noting that in many parts of the country, the lights are on today precisely because of this unpredicted, largely unregulated construction of new natural gas facilities.

Building out 20% of the generation fleet in 10 years was a remarkable and unprecedented rate of change. But just as the deployment of new natural gas assets was starting to level off, ISO-New England would make that rate of change look downright glacial.

ISO-NE Forward Capacity Markets

In the early 2000′s, ISO-New England began to consider markets for capacity services (e.g., MW, as distinct from MWh), the better to encourage long term investments in the New England grid. The Forward Capacity Market (FCM) that was ultimately developed had several noteworthy features:

  1. It had a low cost-of-entry, to facilitate participation from smaller resources.
  2. It explicitly recognized the value of “negawatts,” allowing load-sited resources and conservation to participate on the same terms as remote power plants.[3]

ISO-NE has now completed two years under their FCM, and two corresponding forward capacity auctions (FCAs). As of their most recent auction, they had brought forth a total of 2,936 MW of demand-side resources. To put that total in perspective, the peak demand on the New England grid ranges from 19,000-24,000 MW in a typical year, with the all time peak demand recorded on August 2, 2006 of 28,130 MW.[4]

In other words, in just 2 years, the FCM program has brought forth more than 10% of the all time peak capacity demand on the New England grid, without building a single central power plant. Put another way, that’s equivalent to bringing on line more than two Seabrook Nuclear plants (a 1200 MW facility in New Hampshire) in just 24 months.[5]

Note the similarity with the natural gas fleet deployment in the wake of EPACT/888. In both cases, minor market reforms allowed non-traditional entities to participate in power markets, and in both cases, the rate at which those entities engaged vastly exceeded any historical precedent.

What it means for GHG policy

Successful greenhouse gas policy will require, first and foremost, a massive reallocation of capital in the electric sector. Electricity generation accounts for over 40% of U.S. CO2 emissions, thanks to an antiquated, inefficient, fossil-fuel dominated fleet. The discussion of possible CO2 policies tends to be framed around a handful of technologies (coal, nuclear, renewables, carbon sequestration, etc.), most of which have historically been dominated by regulated monopolies. Noting the slow pace of change in that sector, this conversation inevitably turns to near-term winners and losers, with the presumption that there will be no short-term change in the fleet — just a differential dispatch order as we migrate to lower-carbon sources.

But as the two examples above show, this assumption doesn’t wash. Like New England’s FCM, GHG pricing is nothing more than the monetization of an externality that was previously subsidized by the system. Like EPACT/888, it contemplates revenue streams and market participation by a host of companies and individuals who are not currently a part of the traditional power industry. Both factors suggest that the pace of fleet overhaul will be vastly quicker and cheaper than we anticipate. Will we replace 20% of the fleet in 10 years, like we did after 888? Will we move 2.5 times as fast, as we did in New England after FCM? Might it be possible to move faster still?

The one thing that is certain is that it will be decidedly faster and cheaper than we think.


In addition to speed, there are two broad lessons that can be taken from the examples above.

First, note that in neither case did the reform require a drain on government coffers. Governments did not have to throw money at natural gas generators, nor at demand-side resources. They simply needed to modify regulations to allow market participation by non-traditional actors.

We ought to bear this in mind as we move towards a national GHG policy. Regulators and commentators, schooled in the merits of cost-benefit analysis, have a chronic temptation to assume that any GHG reduction will cost money, and fiscal prudence demands that those costs be minimized per unit of CO2 reduction. That’s a healthy approach, but one that paradoxically tends to overlook the lowest cost forms of CO2 reduction — namely, those which cost nothing more than the political capital necessary to remove existing regulatory barriers. In a market as heavily regulated as the electric sector, one can safely presume that massive volumes of private capital stand ready to invest as soon as those barriers are removed, even before providing any explicit fiscal incentive.

Second, note that neither of the reforms that led to these investments were preconditioned on the removal of the entire regulatory edifice. A common skepticism with respect to the potential for barrier removal derives from the sheer scale of regulatory barriers. We have 100 years of regulated power monopolies in this country, with regulations at the state and federal level (not to mention jurisprudence in courts and utility commissions) designed to sustain that model. The sheer magnitude of those barriers compels one to question the hubris of anyone who thinks that reform is easy.

However, if we stand back to look at the data above, we discover the obvious: you don’t need to tear down an entire dam to restore the flow of a river. You need only remove enough bricks to let the water pressure behind do the rest of the work for you. Modest regulatory reform, targeted only at the critical barriers, is sufficient to unleash massive energy sector reform.

Both lessons are cause for great optimism. Fundamentally changing the GHG signature of our economy will undoubtedly be easier, cheaper and faster than we think … once we start.


[1] Data here and throughout on generator fleet capacity taken from U.S. DOE/EIA.

[2] I’ve omitted hydro and oil capacity from this discussion, which account for the remaining ~200 GW up to the 1998 800 GW base (and were largely unchanged during the period in question).

[3] In fact, load-sited resources participate on more favorable terms than central plants, as the FCM explicitly recognizes the savings in line losses and reserve margins innate to locally-sited capacity investments.

[4] Source: ISO-NE website and personal correspondence.

[5] For comparison, 14 years elapsed between the issuance of Seabrook’s permit in 1976 and full power production in 1990, and was directly responsible for the bankruptcy of Public Service of New Hampshire.