Economists take sides in the battle for DR’s soul.
Bruce W. Radford is publisher of Public Utilities Fortnightly.
Back when the U.S. economy and power consumption still were bubbling, PJM reported in August 2006 that customer curtailments during a week-long August heat wave had generated more than $650 million in market-wide energy savings—all at a mere $5 million cost, as measured in direct payments made to the demand response (DR) providers, set according to wholesale power prices prevailing at the time.
Where else but the lottery can you get an instant payoff of 130-1?
Yet, it wasn’t long after that gain that FERC allowed PJM’s economic load response program (ELRP) to expire, explaining that paying the full spot market price to customers for voluntary curtailments undertaken for their own economic benefit amounted to an unwarranted subsidy, thereby denying a complaint by the PJM Industrial Customer Coalition (ICC) that had sought to extend the life of the program. (See, Order Denying Complaint, Docket EL08-12, Dec. 31, 2007, 121 FERC ¶61,315.)
Ever since, a debate has raged in policy circles about how much compensation grid operators ought to pay to customers who volunteer to curtail demand strictly on the basis of price under so-called “economic” DR programs—as distinguished from DR programs designed to preserve reliability, or to ease temporary supply shortages.
This debate has become fiercely rhetorical, pitting populists versus economists, arguing over competing definitions of “consumer welfare,” “market efficiency,” and “value of trade,” and whether electric customers are unfairly “double-dipping” when they receive a direct payment for curtailing demand that’s keyed to the market price of power, because at the same time they avoid a charge on their retail power bill for the electricity they don’t consume.
The debate was given prominence by none other than FERC Chairman Jon Wellinghoff, who dissented from the 2007 ICC case that let PJM’s economic DR program die. Aided by Commissioner Suedeen Kelly, who joined the dissent, Wellinghoff chided his colleagues on the majority for “mischaracterizing” the market-based payments to DR providers as subsidies, and for “squandering” the chance to develop a factual record to document the exact effects of DR curtailments on regional power prices and the scope of consumer benefits they might provide.
Ñow Wellinghoff might get a second bite at the apple—assuming, of course, that he hasn’t changed his view. That’s because PJM now has proposed to restart its supposedly discredited economic DR program, complete with the same sort of “incentive” (or “subsidy,” if you will) that Wellinghoff’s colleagues struck down in 2007.
To be precise, PJM now has proposed, during the year’s highest-priced 9 percent of hours, to pay the full spot market price without any setoff (i.e., the full locational market price, or LMP) as a reward for voluntary curtailments by so-called fixed-rate customers—that is, the typical retail ratepayer, whose bill doesn’t vary directly with the ups and downs of LMPs. For other retail customers—those high-demand industrial and commercial (C&I) customers whose retail rates track dynamic price changes in the day-ahead or real-time LMP—PJM would pay a flat DR incentive of $75 per megawatt hour, a figure said to approximate the typical retail cost of energy paid by fixed-rate customers.
Sensing the issues at stake, PJM’s new proposal has drawn the policy glitterati out of their ivory towers. Weighing in on PJM’s proposal are none other than the celebrated utility economists Alfred Kahn of Cornell University and William Hogan of Harvard, who each have filed opposing affidavits, pro and con, on the subsidy question. (See, FERC Docket EL09-68, filed Aug. 26, 2009.)
This remarkable disagreement stems, perhaps, from a lack of a common vision as to what demand response really is, or ought to be.
Is DR a virtual physical resource—an electric capacity asset dedicated to the public good—or rather a tool available to market operators to moderate power prices, temper the market-power of generation suppliers, and ameliorate some of the well-recognized flaws in today’s wholesale power markets?
Under the first view, DR pricing should be geared primarily to give access to retail ratepayers to the huge consumer surplus that can be mined during high-priced hours at a relatively low cost (see Figure 1.)
Others warn, however, that too-generous payments for DR actually will destroy consumer welfare by eroding market efficiency.
Joseph Bowring, president of Monitoring Analytics and PJM’s independent market monitor (IMM), argues together with his general counsel, Jeffrey Mayes, that the highest purpose of an economic DR program lies in offering a price signal to retail electric customers—a price signal that now is missing because today’s retail ratepayers pay averaged rates that bear only a very tenuous relationship to the dynamic wholesale LMP market price.
In this view, administrative DR programs run by utilities and system operators serve only as a temporary transition vehicle on the road to an ultimate destination known as “price-responsive demand,” whereby retail customers would be exposed to dynamic energy prices that reflect real-time power prices in regional wholesale spot markets. Electricity would become a simple everyday consumer product; you buy what you can afford, and what you can’t afford, you do without.
PJM CEO Terry Boston explained the idea in a statement he issued to the PJM board of managers on June 26: “PJM’s long-term vision is that ‘Price-Responsive Demand,’ which allows more customers to respond directly to market prices and to voluntarily reduce their consumption when wholesale prices rise, is the ultimate solution to demand participation.”
In fact, PJM has asserted that a key reason for resurrecting incentive payments for DR providers is to “facilitate the move toward an ultimate goal of more extensive price-responsive demand at the retail level.”
And where does FERC Chairman Wellinghoff stand?
At FERC’s technical conference on DR in RTOs and organized power markets, convened in May 2008 before he became chairman, Wellinghoff hinted that he still saw DR as a public good, and worthy of incentive payments of the sort that economists define as “subsidies.” The moment came when he dueled with expert witness Robert Borlick, a Maryland-based consultant advising the Midwest ISO in developing its own economic DR program, on whether incentive payments for DR might be warranted to help integrate the massive influx of wind power in the MISO region.
Commissioner Wellinghoff: Wouldn’t it be appropriate for us to look at perhaps payments over the level that you are discussing for demand response, to ensure reliability in the MISO grid?
Mr. Borlick: I would say no …
Commissioner Wellinghoff: To the extent that wind power generation could increase volatility, as it has in Texas … couldn’t additional quick demand response reduce that volatility, and in fact make that MISO system more stable and more able to take more wind? … Wouldn’t that be appropriate?
Mr. Borlick: Yes, but you don’t pay them a subsidy …
Commissioner Wellinghoff: Wouldn’t it, in fact, be appropriate, if we needed to get significant wind into the system quickly, to take care of the carbon problem, to perhaps pay demand response more?
Realm of Collusion
In a white paper prepared for the Electric Power Supply Association and filed at FERC with EPSA’s comments in answer to the PJM proposal, Bill Hogan provides numerous charts and graphs to show why economists believe it’s inefficient to pay full LMP for DR, but instead should offset any LMP payment by a value equal to the variable energy charge as reflected in retail rates that the customer avoids by reducing consumption. They identify this offset as “G” and give a simple equation to express the proper price of DR service: LMP – G.
As Hogan explains, this price is the most market-efficient because it produces the maximum level of social welfare, defined as the sum of consumer surplus plus producer surplus. Such surplus represents economic rent—the degree to which the market-clearing price is both below what some consumers would be willing to pay, and higher than what some producers would be willing to receive. (William H. Hogan, “Providing Incentives for Efficient Demand Response,” Oct. 29, 2009.)
Opponents are quick to point out, however, that DR is functionally equivalent to self-generation by retail customers, who should receive a full LMP payment, without any offset. So why not the same payment for DR?
The economists agree, but counter that a retail customer providing DR service also avoids the retail energy charge—G—the same as a direct payment equal to G. So to design a DR program that compensates providers in an amount equal to LMP, the payment should include an offset, as follows: (LMP – G) + G = LMP.
As for large-scale C&I customers, many of whom have interval meters and take service under variable retail rates that capture dynamic movements in the wholesale LMP price, the economists favor no DR payment at all, since such customers already earn an implied payment of full LMP when they curtail consumption, simply by avoiding payment of an LMP-based retail rate. In this case, Hogan would treat the entire amount of PJM’s proposed DR payment of $75/MWh as an unjustified incentive or subsidy, incompatible with an efficient market.
A group of DR advocates, including Comverge, EnerNOC, EnergyConnect, Viridity Energy and Wal-Mart, counter that a DR payment of LMP – G would require PJM (or any other grid system operator purchasing DR services) to interpret and calculate state retail electric charges, so that FERC must intrude upon state PUC jurisdiction when it enforces wholesale DR tariffs. (See, Protest of Demand Response Supporters, FERC Docket EL09-68, filed Sept. 16, 2009.) The PJM IMM defeats that argument, however, by showing that neither PJM nor FERC need interpret or calculate the state-approved retail rate, as the ultimate calculation of compensation earned by DR providers is simply a wash:
“This arithmetic [LMP – G + G] involves no attempt to evaluate the level or the appropriateness of G. Consequently, the argument raised by Demand Response Supporters that PJM’s proposal would encroach on the jurisdiction of the [state retail] jurisdiction … is a red herring.” (Answer of IMM, FERC Docket EL09-68, Oct. 16, 2009.)
Nevertheless, such mathematical niceties ring hollow for many DR advocates, such as the PJM-area steel producers, Nucor and Steel Dynamics. In their view, DR’s “true leverage … resides in its potential to eliminate or reduce the need for the most inefficient generators, and the associated outsized reduction in market-clearing prices.” (See, Answer of Steel Producers, FERC Docket EL09-68, Nov. 2, 2009.)
Nucor and Steel Dynamics refer to an oft-cited 2007 report from the Brattle Group, produced on behalf of the Mid-Atlantic Distributed Resources Initiative (MADRI), that curtailing just 3 percent of the load in the BG&E, Delmarva, PECO, PEPCO, and PSE&G zones in PJM during the top 20 five-hour price blocks in 2005 would have cut wholesale energy prices by between $8 and $25 per MWh—price cuts that could save between $57 million and $182 million a year for non-curtailing retail customers within PJM.
The potential market-wide savings stem from the fact that most retail customers pay a fixed retail rate that doesn’t track day-to-day or hour-to-hour movements in wholesale spot prices, so that, as the Brattle report states, “the market demand curve is distorted into a nearly vertical, inelastic curve”(see Figure 1.) The Brattle report adds that savings from DR also can extend beyond the LMP-based energy price, to include capacity markets: “Demand response could reduce capacity prices by reducing peak loads and therefore reducing the demand for capacity, as determined by PJM’s resource adequacy requirements.” (See, Brattle Group, “Qualifying Demand Response Benefits in PJM,” Jan. 29, 2007.)
In this light, say the steel producers, “It should be clear that the $75 fixed incentive payment proposed by PJM is likely to be more than offset by the value of the demand response itself, particularly given that the payment of an incentive is restricted to a small number of the highest-priced hours each calendar year.”
Some of Brattle’s claims are questioned by consultant Borlick, who for the last four years has been advising Midwest ISO on how best to develop DR programs; MISO recently proposed its own economic DR program at FERC (see, Docket ER09-1049, filed Oct. 2, 2009). Borlick says Brattle overstated DR benefits by adding together the energy savings achievable in 2005 and the capacity savings that wouldn’t occur until past 2010, and states in any event that the gains enjoyed by non-curtailing customers due to market-wide energy price reductions always will prove short-lived, since they can exist only when generating capacity is in surplus—as in PJM in 2005—and will be given back once demand growth absorbs the surplus and new generation is needed, thus driving capacity prices back up. However, Borlick’s most telling indictment of DR incentive payments concerns the possibility of market price manipulation.
Proponents often tout DR as the best method to counterbalance supplier market power, which arises when power producers can benefit by withholding a portion of production from the market, so as to boost prices captured by remaining capacity. But as Borlick points out, PJM’s DR incentive proposal might well be tried for the same charge: “PJM is proposing to make ‘incentive payments’ to DR providers in order to drive down spot market prices, so that the loads remaining in the market derive greater value.” (See, Protest of Robert L. Borlick, FERC Docket EL09-68, Sept. 16, 2009.)
In effect, as Borlick concludes, “PJM is proposing to use its unique position … to manipulate market prices in a way that benefits one class of market participants (load-serving entities and their retail customers) at the expense of another class of market participants (suppliers).
“When this is done with the explicit intent of optimizing the net gains from changes in market prices, we are in the realm of collusive behavior.”
Eric Woychik, v.p of regulatory affairs for Comverge, offers perhaps the best argument for paying an incentive or subsidy to DR providers—that the sheer complexity of state-approved retail rate designs makes it much too difficult for curtailing customers to calculate the dollar amount of the avoided retail energy charge, in order to net out the true value of LMP – G, and so puts load at a disadvantage as against generation suppliers, who can log on to the Internet and get an instantaneous, minute-to-minute readout of LMP prices, to guide their decisions on how to bid their supply into the market.
Indeed, state utility commissions aren’t uniform in their choices of rate designs. Sometimes fixed costs are recovered through the energy charge, or variable costs are recovered through the demand charge. Decoupling efforts by some state PUCs further complicate the task of calculating the avoided energy charge (G) and thus put in doubt the exact dollar payoff for the would-be DR provider.
“The deck is stacked against customer and DR providers,” notes Woychik, testifying on behalf of the Demand Response Supporters. “Many customers and DR providers that would otherwise participate in the market cannot easily translate the existing and proposed PJM economic compensation policies [such as LMP – G] into day-to-day operational rules.
“These net-pricing results are lagged and resolved with certainty only after the PJM settlement period is over.”
Imagine the difficulty faced by a large-scale industrial facility, as described in testimony given by Ron Belbot, of the Severstal Sparrows Point steel plant in Baltimore, explaining how the plant managers decide on whether to submit a DR bid: “When loads are forecast to be high, with correspondingly high LMPs, the operation of each of the facilities is reviewed … Information like where a unit is in a particular production run, and if that run can be interrupted without affecting quality or damaging equipment … If a product is being produced for shipment at a later date, it is much easier to delay and reschedule.
“Factors such as inefficiencies and quality impacts of stopping and starting an operation, the ability to perform needed maintenance during the curtailment … are all considered.
“Once this is established, the actual real-time price is monitored and when it is high enough … the curtailment notification is sent to PJM and the appropriate units are shut down.” (See, Affidavit of Ron Belbot, Protest of Demand Response Supporters, Sept. 16, 2009.)
Adding his voice in support of DR advocates is legendary economist Alfred Kahn, who also furnished an affidavit on behalf of the Demand Response Supporters, and who argued that while PJM’s rationale for its proposed DR incentives is perhaps “mistaken” from the point of view of theory, that “serendipitously, its proposed incentive payments are economically correct.”
In his testimony, Kahn gave what appears to be two distinct reasons for favoring an incentive payment of full LMP for retail demand response, without a G-style offset for avoided energy charges. Or, in his words, “why the mere prescription of equating the price of power to marginal supply cost is not a sufficient recipe for maximizing economic efficiency.”
First, Kahn appeared to suggest that DR incentives might well be politically necessary—that is, needed to overcome “the typical unwillingness of state commissions to subject customers to genuine marginal cost pricing.”
Second, Kahn asserted that “as output gets closer and closer to physical limits of generating capacity,” thus increasing the likelihood of shortages or load losses, that “the true marginal costs of energy far exceed LMP,” thus suggesting that DR payments of LMP – G might well prove insufficient to elicit the desired market response from customers.
When asked whether the incentive payments that some see as subsidies will distort economic behavior and produce market inefficiencies, Kahn answered that for his former constituents in New York State, where he once served as chairman of the public service commission, “a sufficient response to your question would probably be subsidy, schmubsidy.”