|
...Adoption increases average load decreases operating profits with the largest decrease for oil-fired generation (59% when all customers adopt). Consumer surplus and welfare gains are modest (2.5% and 0.24% of the energy bill), and emissions of S[O.sub.2] and N[O.sub.x] increase but C[O.sub.2] emissions decrease. Approximately 30% of these efficiency gains could be captured by varying flat rates monthly instead of annually. Monthly flat rate adjustment has many of the same effects as RTP adoption, captures more of the deadweight loss than time of use (TOU) rates, and requires no new metering technology.
1. INTRODUCTION
As electricity markets were restructured in the last decade, sophisticated auction mechanisms were developed to trade wholesale electricity and reveal wholesale prices. However, much less attention was paid to retail pricing of electricity, (1) and the opening of retail markets to competition has lagged. (2) Economic theory describing efficient retail pricing of electricity is based on the well-known theory of peak-load pricing--known as real-time pricing (RTP) in current electricity policy debates. (3) Despite clear efficiency gains in theory, real-time pricing has encountered resistance from many quarters. To help understand this resistance, we analyze the short-run effects of time varying prices in the Mid-Atlantic electricity market known as PJM by constructing a simulation model of competitive wholesale and retail markets. (4,5) Using the model, we analyze the changes in surplus to different customers and producers and the environmental effects of RTP adoption.
The basic economic intuition of RTP adoption is straightforward. For customers not on RTP, retail service providers (whether public service utilities or competitive retailers) must procure sufficient power to cover retail demand at the predetermined flat retail price. With predetermined retail prices, demand in the wholesale market is very inelastic if no customers are on RTP. For customers on RTP, retail service providers pass through the wholesale price. If the wholesale price is high in a given hour, the RTP customers will conserve electricity and reduce the amount of electricity that must be procured. Conversely, if the wholesale price is low in a given hour, the RTP customers will increase their electricity consumption. Thus, RTP adoption by more customers increases the elasticity of the wholesale demand by rotating the demand around the flat retail rate.
The long-run theoretical effects of the increased demand elasticity from RTP adoption are described by Borenstein and Holland (2005). They show that off-peak quantities demanded and prices increase while peak quantities and prices decrease. This implies that the long-run equilibrium flat rate falls with RTP adoption, but that the effects on average loads and capacity are ambiguous. Borenstein and Holland also calculate long-run efficiency gains of three percent to 11% of the energy bill with RTP adoption. We find much smaller efficiency gains in the short run, which may help to explain why these efficiency gains have not been realized. While much of the discussion of time-varying prices has focused on real-time pricing, prices could vary in other ways as well. In particular, flat rates for all hours can vary more or less frequently, or rates could vary by time of use. We compare the benefits of monthly flat rate adjustment with traditional time-of-use (TOU) rates. Surprisingly, we find that the former is superior.
We also examine the environmental impacts of RTP adoption. More specifically, we model how emissions of sulfur dioxide, S[O.sub.2], nitrogen oxides, N[O.sub.x], and carbon dioxide, C[O.sub.2], change in the PJM market as more and more customers adopt RTP. This analysis is complementary to Holland and Mansur (2004), which econometrically estimates the environmental effects of RTP adoption. For each NERC region in the U.S., their paper estimates how a reduction in demand variance (a likely outcome of RTP adoption) will affect the emissions of S[O.sub.2], N[O.sub.x] and C[O.sub.2]. They find that the impact differs depending on the generation technology characteristics of the region.
Section 2 presents the theoretical model, which incorporates the effects of RTP adoption on the retail rates paid by customers not on RTP. Section 3 discusses the data, and describes the simulation model. Section 4 presents the simulation results. With RTP adoption, we find in the short run that: (i) the distributions of loads and prices are compressed, (ii) all rates decrease, (iii) average loads increase, (iv) profits decrease for all generating sectors, (v) consumers surplus increases for all consumers, (vi) efficiency gains are modest, and (vii) emissions of S[O.sub.2] and N[O.sub.x] increase, but emissions of C[O.sub.2] decrease. The robustness of the results to assumptions about demand, imports, generator outages, elasticity of peak demand, and homogenous customers is addressed in the appendix. Section 5 analyzes flat rates that vary by month or by time of use and compares both policies to annually varying flat rates. Section 6 concludes.
2. MODEL
To estimate the short-run effects of RTP adoption, we first model pricing in competitive electricity markets where some proportion of customers are on real-time pricing. A similar model is analyzed carefully in Borenstein and Holland so the model is only outlined here. (6)
Since electricity cannot be stored economically, demand must equal supply at all times. We assume there are T hours with retail demand in hour t given by [D.sub.t](p) where D't<0. (7) A fraction, [alpha], of the customers pays real-time prices, i.e., retail prices that vary hour to hour. (8) The remaining fraction of customers, 1- [alpha], pay a flat retail price, [[bar.p].sub.t], that is the same for every hour in a given period, i.e., [[bar.p].sub.t] = [[bar.p].sub.t]' if t and t' are in the same period. (9) We assume that [alpha] is exogenous and that customers on real-time pricing do not differ systematically from those on flat-rate pricing. (10) Aggregate (wholesale) demand from all customers is then [[??].sub.t]([p.sub.t],[[bar.p].sub.t]) = [alpha][D.sub.t]([p.sub.t]) + (1-[alpha])[D.sub.t]([[bar.p].sub.t]) which implies that [[??].sub.t] is decreasing in [[??].sub.p].sub.t] and [p.sub.t]. When [alpha] = 0, wholesale demand is perfectly inelastic. The larger the share of customers on RTP, the more elastic is wholesale demand. (11) Note that wholesale demand rotates around the point (D([[bar.p].sub.t]),[[bar.p].sub.t]) with RTP adoption.
Each of N generating units supplies electricity to the wholesale market based on its installed technology. We assume that generator n can produce up to capacity [q.sub.n] at constant marginal cost, [c.sub.nt]. Since marginal costs depend on fuel and other input prices, we allow the marginal cost for each unit to vary over the course of the year. A competitive generator would produce at capacity if the wholesale price, [w.sub.t], were above its marginal cost and would produce nothing if the wholesale price were below its marginal cost. Therefore, the supply curve from each generating unit is inverse-L shaped. The industry supply curve, [S.sub.t], is found by aggregating the supply from each generating unit for hour t.
The retail sector purchases electricity from the wholesale sector and Distributes it to the final customers. (12) We assume the identical, competing retailers have transmission and distribution costs of [c.sup.d] per MWh. The profits of the retail sector are then
[pi].sup.r] = [T.summation over (t=1)]([[bar.p].sub.t] - [w.sub.t] - [c.sup.d]) (1-[alpha]) [D.sub.t]([[bar.p].sub.t]) + ([p.sub.t] - [w.sub.t] - [c.sup.d]) [alpha] [D.sub.t] ([p.sub.t]) (1)
The first term is the retail profit from serving the flat-rate customers and the Second term is the retail profit from serving the RTP customers.
If there are no costs of switching retailers, Bertrand competition in the retail sector implies zero retail profits in equilibrium, i.e., [[pi].sup.r] = 0. Competition over real-time prices implies that each real-time price equals the wholesale price plus distribution costs, i.e., [[bar.p].sub.t] = [w.sub.t] + [c.sup.d], and competition over the flat rates implies that each annual flat rate is the distribution cost plus the weighted average of wholesale prices for that period where the weights are the quantities demanded by the flat-rate customers, i.e.:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (2)
for each t where [PHI].sub.t] is the set containing all hours in the same period as hour t. (13) Equating supply and demand in the wholesale market for each t, i.e., [S.sub.t] ([w.sub.t]) = [[??].sub.t]([p.sub.t],[[bar.p].sub.t]), completes the characterization of the equilibrium.
3. DATA AND SIMULATION MODEL
The study period covers two years from April to March beginning in April of 1998 and ending in March of 2000. We apply the model to the PJM electricity market, which covered parts of Pennsylvania, New Jersey, Maryland, and Delaware at that time. (14) Unless otherwise noted, all data are from various government and industry sources as detailed in Mansur (forthcoming). The 392 modeled fossil generating units, which range in capacity from 0.6 MW to 850 MW, account for approximately 60% of the electricity generated in PJM with the remainder being supplied primarily by nuclear power. (15) The largest fossil units are powered by coal (46% of fossil capacity) with the remainder powered by oil and natural gas (19% and 35% of fossil capacity).
Since the efficiency of each unit (measured by the heat rate in BTU per kWh) is publicly available, we can estimate the daily marginal cost of each unit from the costs of fuel and other inputs. Key input prices used in calculating the daily supply curves include prices of natural gas, heating oil, S[O.sub.2] permits, and N[O.sub.x] permits. (16) Coal prices are assumed constant throughout the study period.
Demand is based on the electricity load reported by PJM. The load averaged 29,400 MWh across the study period with a minimum load of 17,461 MWh and a maximum load of 51,714 MWh. Wholesale electricity prices ranged from slightly negative to the price cap of $999 with an average wholesale price of $25.80 per MWh.
The simulation model uses the data from PJM to estimate the effect of more customers adopting RTP in competitive markets. In the simulation, we make a further assumption of identical, constant demand elasticities, but allow demand for each hour to have a different scale parameter. (17) The scale parameters are calculated from the observed hourly loads and rates in PJM.
The supply side of the model includes generation from fossil, nuclear, and hydropower. (18) We assume the fossil supply curve for each coal-, oil-, and gas-fired generation unit is inverse-L shaped where the marginal cost is calculated from the unit's heat rate and daily fuel prices. The capacity of each unit is derated by its expected outage factor. (19) Supply from nuclear and hydropower is assumed to be perfectly inelastic at its observed hourly levels throughout the simulations.
Nuclear power stations have very low marginal costs and, thus, run whenever possible. Hydropower in PJM is mostly from run-of-river dams which do not vary their output based on market conditions. The supply curve is found by aggregating the supply from each source. Figure 1a illustrates the supply curve for a given day. Of the fossil units, the coal-fired units have the lowest marginal costs, the gas-fired units are the mid-merit technology, and oil-fired units have the highest marginal costs.
[FIGURE 1 OMITTED]
For a given flat retail price, the wholesale demand curve in any hour is completely determined, and a candidate wholesale market equilibrium can be calculated from the intersection of the wholesale supply and demand for each hour. These wholesale prices can be used to calculate profits to the retail sector for each year. If the retail profit for a year is positive (negative), the equilibrium flat rate for that year must be lower (higher) than the assumed flat rate. The flat rate...
NOTE: All illustrations and photos
have been removed from this article.
More articles from The Energy Journal
Energy and Sustainable Development in Mexico.(Book review), October 01, 2006
Electricity Deregulation: Choices and Challenges.(Book review), October 01, 2006
Energy prices and turning points: the relationship between income and ..., October 01, 2006
Looking for additional articles?
Search our database of over 3 million articles.
Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication
name or publication date.
About Goliath
Whether you're looking for sales prospects, competitive information, company
analysis or best practices in managing your organization,
Goliath can help you meet your business needs.
Our extensive business information databases empower business
professionals with both the breadth and depth of credible,
authoritative information they need to support their business
goals. Whether it be strategic planning, sales prospecting,
company research or defining management best practices -
Goliath is your leading source for accurate information.
|
