Rising energy demand could spur investment in waves, but a fixed capacity charge might flatten the curve.
Construction cycles occur in many industries. Examples include automobile manufacturing, metallic commodities, agricultural commodities, and real estatesome of which may differ in fundamental ways from the electric industry. Yet that does not make the power business immune from boom and bust.
The differences that separate these other industries from the power business may arise from their long supply chains or their ability to store products in inventory prior to delivery. The real estate industry, however, appears similar to electricity in many respects. Investors consult market data in deciding whether to construct new buildings, just as a power producer might do. And in real estate, delays in completing construction have given rise to cycles that date all the way back to the early 1800s (See bibliography, Hoyt 1933, DiPasquale 1996, Sterman 2000).
Construction cycles have not been as prominent in the long history of the electric industryfluctuations in reserve margins have not been as dramatic as changes in inventory in other industriesbut that fact may change. When restructuring is completed, electric utilities no longer will be obliged to build the new power plants needed to serve demand. They will consult the market before building the plants that will be needed in the future. Some believe the market will respond in a cyclical manner
This article describes computer simulations to help us understand the key feedback mechanisms that control power plant construction in a restructured electricity market. The model was originally developed for a western utility to help its planners think through strategies to achieve its goals for electric rates and cash flow. The model is used in this article to help us appreciate the potential for construction cycles in the western United States. Summary results are given here; detailed results can be found in an earlier article that I published in Energy Policy, which is listed at the front of the bibliography.
In short, the research in this article suggests that in a restructured market, power plant construction may come in waves, causing alternating periods of over- and under-supply of electricity that may prove severe enough to require operators of spot energy markets to impose artificial price caps. The end result would be major swings in market prices as the industry moves through the phases of a construction cycle.
This research indicates that construction cycles are a potentially serious problem if the western electricity markets follow the example set in California. Under some circumstances, the cycles would take the extreme form of a "limit cycle." In this unfortunate situation, the industry would face repeated periods of under supply, and regulators would be forced to intervene with administrative limits on the price of electricity.
These problems arise from the inherently unstable interactions between the market and investors. But construction cycles are not inevitable. They could be dampened substantially if western markets allow for additional incentives.
This article considers a constant capacity price as a vehicle to deliver the added incentive. Simulation studies have shown that a capacity payment in the range of 5-8 mills per kilowatt-hour could stabilize the market in the long run. The "price of stability" turns out to be only slightly higher wholesale rates in the short run.
Moreover, this strategy of adding capacity price incentives would seem to have minimal effects on consumers in retail markets. Based on the ratemaking rules in California, retail customers would be shielded from the short-run increase in wholesale rates.
Reliance on Energy Prices: Cyclical Behavior in the Long Run
The figures that follow each assume a particular rate of demand growth and then show the effects on (1) the PX energy price, (2) investment in power plant permitting and construction, and (3) new plants coming on line, which are assumed to be gas-fired, combined-cycle combustion turbines (CCCTs).
They also show how the energy price compares with the total, levelized cost of new CCCTs.
Short-Term Reaction. Figure 1 begins with just two variables, the cost of CCCTs and the PX energy price, and considers the reaction of the construction cycle only over the short term, for the eight years from 1998 to 2006. It assumes the simulation with base case assumptions as explained in the Technical Appendix. In Figure 1, both the PX energy price and the levelized cost of a new CCCT are measured in dollars per megawatt-hour, in constant 1997 dollars. The PX price begins the simulation at $20 per megawatt-hour, around the same result obtained in a detailed analysis by the Northwest Power Planning Council (NPPC 1998). The growth in demand forces the PX to increase the price to bring more of the existing gas units into operation. The price of natural gas escalates at 1.0 percent per year, which also contributes to the upward trend in both the PX price and the levelized cost of a new CCCT. By the end of eight years, the PX price has increased to exactly the levelized cost of a new CCCT.
Three Predictions on Market Price Effects
Northwest Power Planning Council
The staff of the NPPC warned in 1998 that most commodities go through cycles of over- and under-supply, and that electricity is not likely to be different. They repeated the warning in their 1999 report on adequacy of power supply in the Northwest.
This consulting firm took a similar position in its 1999 report on electricity markets in "exuberant regions" like Texas and New England. Analysts observed that New England was the leading market for gas-fired, combined-cycle projects, where the capacity planned on the drawing board appeared sufficient to double the installed capacity base.
Brattle also warned that electricity prices in the future would no longer reflect average total costs, but would likely be highly volatile, with unpredictable price spikes in the short run and boom-bust cycles over the long run.
California Energy Commission
A third example comes from market price scenarios issued by the California Energy Commission in February 2000.
The commission staff warned that additions of new generating capacity would not occur in a smooth, even manner, but would more likely follow a cyclical pattern, producing periods of excess and lean capacity.
The staff also argued that this cyclical pattern would occur primarily because profitability of new generating units would remain heavily dependent on the prices during the summer peak demand season. They warned that summer peak prices would not necessarily reach a level to support new entry until reserve margins drop below the levels normally associated with reliable service. A.F.
The results in Figure 1 are typical of the short-term results from more detailed models.1 These models predict that the PX price will remain exactly at the levelized cost of a new CCCT, even as the horizon extends beyond the short term, but that result is questionable. This prediction is plausible only if we believe that new units will come on line in exactly the amount to counter the growth in demand. But in a competitive market, investors are not striving to bring exactly the right amount of capacity on line. Their goal is to invest in a profitable manner. We need to extend the simulation to learn if their actions lead to a stable price trajectory.
Long-Term Reaction. Figure 2 extends the simulation over 20 years to reveal that in the longer term, investor activity would likely appear in waves, causing cyclical variations in the PX price.
The first wave of investor activity appears around the years 2004_2009. Capacity in the permitting process reaches a peak of around 10,000 MW in 2008. Capacity in construction reaches a peak one year later. Installed CCCT capacity begins to grow around 2007. By this time, however, the PX price has climbed to $32 per megawatt-hour, well above the cost of a CCCT.
By the year 2008, installed capacity is growing rapidly. These new units allow the PX price to fall from the peak. By the year 2009, the PX price has returned to the cost of a new CCCT and is on a downward path. The PX price reaches a low of $22 per megawatt-hour before returning to a gradual upward trend. This renewed upward trend is similar to the upward trend at the start of the simulation; it is caused by the gradual growth in demand.
When the first wave of construction ends, about 14,000 MW of CCCTs are in operation, but there are no new CCCTs in the permitting process. Investors are reluctant to get involved in new projects until the PX price approaches profitable values. This reluctance does not reverse again until around the year 2013, when a second wave of activity appears. This wave appears near the end of the simulation, when the PX price peaks at $37 per megawatt-hour. This second wave of CCCTs comes on line around the year 2017, and drives the PX price down to $23 per megawatt-hour by the end of the simulation.
This base case simulation reveals an inherently unstable systemnot only do we see oscillations in the PX price, but the oscillations grow larger over time. The oscillations appear even though we have created an artificial environment with no surprises. There are no unexpected changes in the rate of growth of demand, and hydro generation is the same, year after year. With these factors held constant throughout the simulation, we know the oscillations come from the interactions inside the system.
High-Growth Scenario. The pattern in Figure 2 represents only one of several patterns of cyclical behavior. If the demand growth rate is lowered to 0.5 percent per year, for example, the simulation shows a gradual growth in the PX price over time. It is not until the very end of the simulation that the PX price would interest investors. The model shows no signs of a construction cycle under such low growth conditions. But a more volatile simulation is shown in Figure 3, where demand is assumed to grow at 2.5 percent per year.
In Figure 3 all other assumptions are the same as before, but the simulation indicates that the PX price will follow an entirely different trajectory. It grows rapidly during the early years of the simulation, reaching the cost of a CCCT in just five years. It then "shoots past" the cost of a CCCT hitting an administrative limit by the year 2004. The circuit breaker is imposed because the region is short of energy. This price cap remains in effect for over a year before sufficient capacity comes on line to eliminate the shortage. The new capacity then causes the price to fall rapidly, reaching a low of around $22 per megawatt-hour around the year 2006.
Figure 3 shows the same pattern recurring after 2006. Continued growth in electricity demand causes the PX price to return to an upward trend, and the trend accelerates around 2010. Administrative limits are imposed again around 2012, and the price is capped until another wave of construction eliminates the shortage of energy.
The oscillation in Figure 3 is known as a "limit cycle." Limit cycles are common in natural systems. They appear when a system prone to growing oscillations encounters a natural limit. Once the system reaches its natural limit, the cycle can repeat itself in an amazingly stable manner (as in the beating of the human heart). But the limit cycle in Figure 3 is not particularly "natural" since the price caps are imposed administratively on a market that has failed to stimulate the needed investment in a timely manner. It is highly doubtful that the industry could survive this pattern of behavior. It is important that we understand the factors contributing to the unstable behavior, and that we design policies to stabilize the system.
Adding a Capacity Incentive: A More Stable Market
The simulations shown so far envision an electric industry based on the premise that investors must rely only on the energy price in a competitive spot market in order to recover the full cost of a new CCCT. They reveal a market plagued by boom and bust. These cycles could prove particularly troublesome if a new CCCT should lose some of its competitive advantage, for example, because of an increase in construction costs. (See Ford, 1999.)
These results suggest that investors need an additional incentive to construct power plants, beyond the bare PX energy price. Let's consider a capacity payment as one way to deliver the extra incentive.
Capacity Payments. Capacity payments provide incentives for generators to be available when the system needs generating capacity; they also provide extra revenue to the generator to cover the capital and other fixed costs that are not covered by the energy price. Where capacity markets do exist, they tend to be less formal and provide less price reporting than electric energy markets.2 For example, Rose (1995) predicted that wholesale markets would respond to a tighter supply/demand balance with prices that "develop a hidden premium over marginal fuel costs." He calls the hidden premium a "pure capacity price," and he inferred the size of the pure capacity price "hidden" in energy price data from 1997. Rose asks us to "think of the capacity price as being spread equally over all hours of the year, rather than occurring in certain peak hours, as it does in the real world." He then calculates the pure capacity price at just over 5 mills per kilowatt-hour based on the annual capital cost of a new combustion turbine.
A capacity payment measured in mills per kilowatt-hour is used in the construction cycle because "spreading the payment" over the year corresponds to the average annual operation of the model. Generators now receive the sum of an energy price and a capacity price. The energy price is based on the same rules as before. It's the price needed to bring forth the generation from existing units to meet the demand for energy imposed on the PX. Now, what about the capacity price? Perhaps we should allow the capacity price to vary over time due to variations in the need for generating capacity. But the previous simulations show a highly volatile system if we rely on energy prices alone. The last thing we need is additional volatility from a capacity payment that rises and falls with swings in the construction cycle.3 Investors need additional incentives, not additional volatility. It makes better sense to provide the additional incentive as a capacity payment that remains constant over time and use the simulation model to search for the size of the capacity payment to achieve improved stability. The search (Ford 1999) has shown that capacity payments in the range of 4-8 mills per kilowatt-hour are sufficient to stabilize the system.
Figure 4 shows an example with a 5 mill per kilowatt-hour payment. The simulation shows the stable pattern of construction that we would like to see when demand is growing at a steady rate year after year. Investors begin applying for construction permits around the year 2001, and the capacity begins to grow around 2003. The pattern of growth is smooth and steady because investors fill the "pipeline" by the same amount year after year. The steady growth in new generating capacity allows the energy price to follow a smooth upward trend. The upward trend is caused by a 1 percent per year real escalation in the price of natural gas. The total PX price is 5 mills per kilowatt-hour above the energy price, so it follows the same upward trend.
Wholesale Impacts. The 5 mill per kilowatt-hour capacity payment is clearly evident in Figure 4 because the total price is always 5 mills per kilowatt-hour higher than the energy price. This may leave you with the impression that the "price of stability" is a constant 5 mill per kilowatt-hour penalty, year after year. But that would be a false impression. The real "price of stability" is to be found by comparing a simulation relying exclusively on energy prices with simulations that allow an additional payment for capacity. Such comparisons show that capacity payments lead to long-term price stability, and the eventual price converges to the levelized cost of a new CCCT. In other words, a constant capacity payment can convert an inherently unstable system into a stable one, and the "price of stability is limited to somewhat higher prices over an eight-year period.
One way to interpret this short-term price penalty is to estimate the increase in wholesale earnings by large utilities in California. The annual payments would be highest in the years immediately after capacity payments are adopted. They would then phase out automatically over time. The total impact could range from $1.6 to $4.0 billion depending on the simulated conditions.
Retail Impacts. A $4 billion penalty could be interpreted as the extra wholesale cost to serve the energy needs of the retail demand of the California utilities. If wholesale costs increase by $4 billion, one would normally expect that retail costs would increase in a similar manner. But retail ratemaking in California is governed by the state's key electric restructuring law, AB 1890. That law requires consumers to cover the stranded costs of the three utilities. The stranded costs have been estimated at $26 billion to $30 billion. The utilities are allowed to recover these costs by adding a competitive transition charge (CTC) to the retail bill.
Figure 5 places the CTC in perspective by showing charges on a typical bill in 1998. The largest component is distribution charges (36 mills per kilowatt-hour.) The CTC is the second-largest contributor at 34 mills per kilowatt-hour. The PX energy charge, the focus of this article, is the third-largest charge and is substantially smaller than the CTC. The PX energy charge and the CTC are the charges that would change due to the capacity payments. The PX energy charge would be replaced with the total PX charge (for both energy and capacity), and the wholesale impact might be around 5 mills per kilowatt-hour immediately after capacity payments are adopted. The CTC charges, on the other hand, would decline by a corresponding amount because the utilities would earn more from selling to the PX.
The compensating change in the CTC charge arises from the rules of AB 1890, which were written to protect the utilities from the adverse cash flow of a competitive market. If the market clears at lower prices, for example, the utilities may impose a larger CTC to meet their fixed cost obligations. If the market clears at higher prices, however, a smaller CTC would be required. In this article, I envision that market prices would clear somewhat higher in the first few years after adoption of a capacity payment. When the utilities earn increased revenues selling to the PX, regulators would lower the CTC. Using the example of the three large utilities, the PX charges could increase by around $4 billion over the next five years. The stranded costs of these utilities are around $26 billion to $30 billion, substantially greater than the $4 billion. It seems that capacity payments would not penalize California's retail consumers in the short term.
Now, what about the long term? Could capacity payments lead to higher retail bills after the CTC charges are phased out? The California PUC expects CTCs to be completed by the year 2002, so one might ask if the retail consumer would experience higher bills after the year 2002. But the short-term impacts of capacity payments are largely eliminated by the year 2002. By the start of 2003, for example, the total PX price is almost identical in model simulations (Ford 1999).
1 Selected models used in the United Kingdom, in California, and in the Northwest are reviewed in the detailed version of this article (Ford 1999).
2 The observation on capacity markets appears in (CERA 1997). According to Rose (1995), "much has been written about the changing market for economy electric energy. In contrast, prices for wholesale electric generating capacity appear to be hidden."
3 The constant capacity payment was used here to confine the analysis within the reach of a highly aggregated model, which simulates the average annual price of energy. The constant payment was also adopted because of my instinctive aversion to adopting variable capacity payments as a supplement to variable payments for energy. My aversion is based, in part, on the findings by Bunn and Larsen (1992). They used computer simulation to demonstrate the destabilizing effect of the variable price for capacity originally used in the United Kingdom.
But setting a price for capacity is only one of several approaches to encourage investment in adequate generating capacity. A second approach would impose an obligation for installed capacity. This approach has been used by ISOs in New England, New York and Pennsylvania-New Jersey. According to Singh and Jacobs (1999), the capacity requirement approach "really boils down to administratively determined capacity payments."
A third approach relies on market prices as the primary signals for long-term investment while assigning short-term responsibility to the ISO. This is the approach in California where the ISO payments for ancillary services (AS) can be substantial. Sing and Jacobs (1999) report that California's AS costs "have been the highest observed thus far." Grix (2000) reports that AS payments amounted to 5.6 percent of total market energy costs in 1999. The value of AS payments can be important to investors, especially if their plant is to be equipped with automatic generation control. The value of AS payments is demonstrated by Deb (2000) using a multi-product, multi-area optimal power model. His illustrative calculation shows a 39 percent increase in net present value of earnings if generators optimize their operations across multiple markets.
This article is adapted from a detailed description of the simulation study, which is reprinted from , Vol. 27, Andrew Ford, "Cycles in Competitive Electricity Markets: A Simulation Study of the Western United States," pp. 637-658, 1999, with permission from Elsevier Science. Additional readings are listed below.
Bunn, Derek, and Larsen, Erik, "Sensitivity of Reserve Margin to Factors Influencing Investment Behaviour in the Electricity Market of England and Wales," , May 1992.
Borenstein, Severin, "Price Convergence in California's Deregulated Wholesale Electricity Markets," Fifth Annual Research Conference on Electric Industry Restructuring, University of California Energy Institute, Berkeley, Calif., March 17, 2000.
California Independent System Operator, Market Surveillance Unit, "Annual Report on Market Issues and Performance," Folsom, Calif., June 1999.
California Public Utilities Commission, "Electric Restructuring Resource Notebook," San Francisco, Calif., 1999.
Cambridge Energy Research Associates, , Summer 1997.
Deb, Rajat, "Rethinking Asset Values in a Competitive Environment," , Feb. 1, 2000.
DePasquale, Denise, and Wheaton, William, , Prentice Hall, 1996.
Ford, Andrew, "System Dynamics and the Electric Power Industry," , Vol. 13, Num. 1, Spring 1997.
Graves, Frank, "How Competitive Market Dynamics Affect Coal, Nuclear and Gas Generation and Fuel Use_A 10-Year Look Ahead," Final Report TR-111506 by the Brattle Group to the Electric Power Research Institute, Palo Alto, Calif., 1999.
Hoyt, Homer, , University of Chicago Press, 1933.
Grix, Richard, "Market Clearing Prices Under Alternative Resource Scenarios 2000-2010," California Energy Commission, Sacramento, Calif., February 2000.
Klein, Joel, "Interim Staff Market Clearing Price Forecast for the California Energy Market," California Energy Commission, Sacramento, Calif., December 1997.
Levesque, Carl, "Merchant Mania: Regional Markets Draw Gen Plant Projects," , Jan. 1, 2000.
Moore, Taylor, "Merchant Plants Drive Market Competition," , Vol. 24, Num. 2, Summer 1999.
Northwest Power Planning Council, "Analysis of the Bonneville Power Administration's Potential Future Costs and Market Revenues," Portland, Ore., 1998.
Northwest Power Planning Council, "Regional Power Supply Adequacy/Reliability Study," Phase I Report, Portland, Ore., December 1999.
O'Donnell, Arthur, "Exceptions to the Rule: Bypassing the California Transition Charge," , Nov. 15, 1996.
Richard, Dan, and Lavinson, Melissa, "Something for Everyone: The Politics of California's New Law on Electric Restructuring," , Nov. 15, 1996.
Rose, Judah, and Mann, Charles, "Unbundling the Electric Capacity Price in a Deregulated Commodity Market," , Nov. 15, 1995.
Rose, Judah, "Last Summer's Pure Capacity Prices," December 1997.
Singh, Harry, and Jacobs, Jonathan, "Capacity Products and ISO Markets," Fifth Annual Research Conference on Electric Industry Restructuring, University of California Energy Institute, Berkeley, Calif., March 17, 2000.
Sterman, John, , Irwin McGraw-Hill, 2000.
Thurston, Charles, "Merchant Power: Promise or Reality?" , Jan. 1, 1999.
Technical Appendix: A Description of the Model
Electricity demand is represented by the annual average demand for electric energy in the western United States. Demand grows at a constant rate year after year, thus eliminating variability in demand as a confounding factor. Demand is satisfied by a power pool following the rules adopted by the Power Exchange (PX) and the Independent System Operator (ISO) in California. The combination of the PX day-ahead price and the ISO real-time price are combined into a single, market price. Since the PX day-ahead price and the ISO real-time prices seem to have converged, for the sake of simplicity the article refers to this combined price as simply the "PX price."*
The model assumes that all of the demand in the U.S. West is bid without a price, so the PX price will be set at the cost of the most expensive generating unit needed to meet the total demand. If the region is short of resources, the PX will be forced to impose an administrative limit, and the model imposes a "circuit breaker" mechanism to interrupt the normal operations.
Nuclear and Hydro. There are around 2,000 generating units in the West, and several models are available to simulate each and every unit. For our purposes, it makes more sense to combine the units into broad categories like hydro and nuclear. These generators have low variable costs or operating constraints that require them to operate in a "must-run" mode. Nuclear capacity is held constant at just over 9,000 megawatts; the hydroelectric generation is constant at just over 20,000 MW. Of course, it is well known that hydro will change from year to year due to changes in the weather. We assume constant output from hydro generators to maintain focus on the construction cycle. If we see waves of construction and periods of over-/under-supply, we will be able to trace the waves back to their underlying cause without becoming lost in a myriad of "external" factors like changes in the weather.
Coal and Gas. There are around 36,000 MW of coal capacity and 33,000 MW of gas capacity in the West. Their generation is controlled by the PX price. The price is measured in dollars per megawatt-hour, and represents an annual average price for electric energy. We calculate the fraction of coal and gas capacity in operation by finding the heat rates of the marginal units. The model uses a simple search heuristic to find the PX price that will bring forth the needed generation. The demand is represented by an annual average demand for electric energy, so the model does not simulate the hourly or seasonal variations in the PX price. Instead, we use the annual average price to summarize a complex distribution of prices over the many hours in a year. We rely on a single price because it is believed that the average price over the entire year is likely to be studied by investors contemplating construction of new units.
New Plant Construction
We assume that combined-cycle, combustion turbines (CCCTs) will be the most popular technology for "merchant plants" constructed in the coming decades. The simulations envision CCCTs with:
Construction Cost 650 $/kW ('97$) Average Heat Rate 6,842 Btu/kWh Fixed O&M Costs 2.87 mills/kWh Availability Factor 90%
If the investors' fixed charge rate is 15 percent per year, the levelized fixed cost would be around $16 per megawatt-hour. If natural gas is priced at $1.50 per million Btu at the start of the simulation, the variable cost would be only $10 per megawatt, and the total levelized cost would be around $26 per megawatt-hour (in '97 $). The total cost will increase over time with increases in the price of natural gas. With natural gas prices at $2.20 per million Btu in the year 2003, for example, the total levelized cost would be around $27.5 per megawatt-hour (in '97 $), matching the 1997 price forecast of the California Energy Commission.#
We assume that new CCCTs will be constructed based on investors' forecasts of the future PX price. If investors look into the future and "see" a price exceeding $27.5 per megawatt-hour, they will apply for construction permits and receive permission to build 12 months later. Investors will not apply for a construction permit until they "see" a profitable PX price in the future. Investors are continuously monitoring the PX price. If they believe that profitability has dropped during the 12 months required to obtain the construction permit, they may cancel the project. But if CCCTs still appear profitable, investors will initiate construction, and the new units will come on line 12 months later.
This article focuses on the stability of construction over time, so it is important to appreciate the principal feedback mechanisms in the model. The figure below shows the two balancing loops that dominate the simulated system over time. The inner loop is highlighted with bold arrows. It represents the rapid adjustments in generation from existing gas-fired plants responding to changes in the PX price. This loop acts without delay to keep the supply and demand on the PX in balance. (The "seesaw" in the middle of the loop reminds us of the balancing function of the loop.) This loop can do its job as long as there is sufficient generating capacity in the region.
The upper portion of the figure shows the collection of assumptions on investors' behavior. To trace the cause and effect around the outer loop, suppose that growth in demand leads to higher PX prices over time. The increase in PX price leads to an increase in investors' forecast of future PX price, an increase in investors' forecast of profitability, and an increase in the number of CCCTs starting the permitting process. After a 12-month delay, there will be an increase in the number of CCCTs starting construction, and after another 12-month delay, an increase in the number of CCCTs coming on line. The increase in capacity will reduce the demand imposed on the PX and lead to a downward adjustment in the PX price. This example reveals that the closed chain of cause and effect acts to counter the increase in the PX price. Once again, we have negative feedback.
These feedback mechanisms may prove sufficient to maintain a relatively stable pattern of price changes over time. But there is no guarantee that this system will be stable just because we have negative feedback. It is certainly possible that delays in the actions of the construction loop could lead to periods of under and over supply. To sort out the possible patterns, we turn to computer simulation. (See main text.)
Notes to Appendix
* The California ISO market surveillance unit reports that PX day-ahead prices and ISO real-time prices were similar for most load conditions during the early months of 1998. In a more recent assessment, Borenstein reports that the prices have "gravitated toward one another and have been statistically indistinguishable during most, though not all of 1999."
# The estimate of $28.5 per megawatt-hour is explained by Klein (1997). In a more recent report, the California Energy Commission estimates the fully levelized cost of a CCCT at $31.3 per megawatt-hour (Grix 2000).
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.