EnergyPolicy 1994 22 (7) 555-570
Electric vehicles and the electric utility company Andrew Ford
This article describes the probable impacts of electric vehicles on the utilities which will provide their electricity. It draws on the results of a recent case study of the Southern California Edison Company. The impacts on the system if millions of electric vehicles were to appear in southern California over the next 20 years are described, and the conditions under which the findings may be transferred to utilities outside southern California are explained. The paper concludes by discussing the effect of incentives that the utility company or the state of California might provide to promote the sale of electric vehicles. Keywords:Electric vehicles:electricityutilities
The proceedings of the 1992 Stockholm Conference on The Urban Electric Vehicle opens with the observation that: l The cities of the world are increasingly affected by automobile traffic and its associated pollution. This phenomenon is about to create important markets for electric vehicles (EVs), and electric hybrid vehicles, which do not pollute at their point of use and may often have lower atmospheric emissions than conventional vehicles on a regional and global basis. Industrialized countries and cities throughout the world are embarking on policies which will encourage the use of tens or hundreds of thousands of electric vehicles within the next decade, and most major automobile manufacturers are expected to begin producing electric car models in the next few years. Although much of the advanced research on EVs takes place in the USA, commercialization is said to be closer in Europe. Indeed, one speaker at Stockholm expects to see 500 000 EVs in Europe before the year 1999. Japan is also promoting EVs, and its government/industry plan calls for 200 000 EVs by the year 2000. ~ The author is with the Program in Environmental Science, Washington State University, Pullman,WA 99164-4430, USA.
0301-4215/94/07 0555-16 © 1994 Bunerworth-HeinemanLtd
In the USA a crucial factor is air pollution regulation. John Dabels, from the General Motors Corporation, estimates that air pollution legislation in the USA could translate into a production requirement of around 300 000 to 400 000 EVs per year once requirements in states like California call for 10% zero emission vehicles or ZEVs. The ZEV goals raise interesting questions for policy makers in California and other states that choose to follow the California example. The key issues revolve around the question of incentives versus penalties, and the various viewpoints were summarized by John Dabels at the Stockholm Conference as follows: This type of legislation requires General Motors and others to sell electric vehicles, but it does not require the public to buy them. It remains to be seen if these cars will be popular enough for us to sell this many. The current interpretation of the regulation in California is that those companies who do not comply will not be allowed to sell any cars in California the following year, which is a very severe penalty. In the meantime, we are trying hard to develop a car that the public will like and that will sell well. When you ask different people what will happen if the cars are good products but simply not popular, the answer you get depends on who you ask. Environmentalists say 'the car companies will have to lower the price.' Car people say 'the Federal and State governments will have to provide subsidies or reduce the percentage.' Regulators are generally silent. Time will tell.
Purpose and organization The purpose of this article is to examine EVs from the point of view of the electric utility company. The first objective is to examine the impacts on the utility if EVs were to sell at or above the levels called for by the regulators. I begin with a review of the literature on EVs. I then explain the assumptions adopted for eight EV scenarios selected in the recent case study of the Southern California Edison (SCE) system. I summarize the probable impacts on electric utility loads, system operations, revenue requirements and the average electric rate. I argue that EVs can lead to improved load shapes,
555
Electric vehicles and electric utilities." A Ford
40 35"
/"~''"'~
Breakthrough
30.
25
RFF,,, oucat
~ 2o :~
15
~ ,'n
10
/
o.o.... S'e
CECg0
EV
5. 0 ~
0 Number of electric vehicles (million)
Figure 1. Comparison of EV scenarios in previous studies.
improved efficiency of power plant operations, and a possible reduction in the overall electric rate. Of course, EVs also contribute to an improvement in air quality in southern California. But the emissions benefits 3 and the subsequent air quality benefits 4 are not the focus of the article. The various benefits of EVs raise the question of whether the utility should contribute financial incentives to promote the sale of EVs. I address this question with special attention to the impact on the average ratepayer. I show that it is difficult for the utility to finance an incentive program without raising the average electric rate. I conclude with an examination of the feasibility of a state operated feebate program to promote EV sales.
Literature In a review of the 'modern history of electric vehicles', DeLuchi observes that interest in EVs has peaked three times in the past few decades. 5 The first peak occurred in the mid-1960s; it coincided with growing concern over urban air pollution. The second peak arrived in the late 1970s as a reaction to problems from oil imports. The third surge started in the mid-1980s, and it continues to the present day as evidenced by the many important presentations at the 1992 Stockholm Conference. The current interest is attributed to a combination of concerns over energy security and urban air pollution. Southern California is a focal point of interest in EVs because of the serious smog problem. Vehicles are a major contributor to smog, and EVs have been the subject of detailed studies. Each of the studies employ a
556
scenario approach which calls for the analyst to prepare a script or story about the future of EVs in southern California. 6 Six scenarios from various southern California studies are arranged for comparison in Figure 1. Each of the scenarios was developed at the end of a 20 year planning period, so they correspond to conditions envisioned for the time period around 2010. Each scenario is located on the horizontal axis by the number of EVs and on the vertical axis by the annual sales of electric energy to charge the EV batteries. The three rays in Figure 1 allow the scenarios to be positioned relative to their daily energy requirement. For example, both the breakthrough and steady advance scenarios envision EVs using 20kWh/day. The breakthrough scenario shows the highest electricity sales to EVs; it envisions 5 million EVs which would amount to nearly half of the vehicles operating in southern California. The wide variation in conditions shown in Figure 1 is typical of scenario studies. The investigators' intent is not to converge on a forecast of the number of EVs likely to appear in southern California. 7 Rather, the previous investigators wanted to explore several widely varying futures. The previous studies lead to several common conclusions. First, the previous investigators all concluded that the power generation needed to supply EVs would come largely from natural gas fired generating units. Next, they concluded that EV demands would tend to improve the overall shape of the electricity demand and allow southern California utilities to operate more efficiently. One study estimated that SCE could accommodate around 0.6 million large electric vans within its existing resource plan, and the electric vans would lead to more efficient operation and a 3.5%
Energy Poliey 1994 Volume 22 Number 7
Electric vehicles and electric utilities: A Ford
Table 1. Electric vehicle scenarios. Number of EVs
Night-time charging
Day-time charging
Batteries, range and small car efficiency
I
2 000 000
2010
Customer convenience
Minimal
Na/S, 250 miles 4.17 miles/kWh
2
2 000 000
2010
Customer
Minimal
Na/S, 250 miles 4.17 miles/kWh
incentive 3
2 000 000
2010
Smart
Minimal
4
2 000 000
2010
Customer
Na/S, 250 miles 4.17 miles/kWh
control Some
Ni/Fe and Pb/acid 150 miles
convenience
4.17 miles/kWh 5
1 000 000
2010
Customer
Minima[
6
I 000 000
2010
Smart
Minimal
50 000
2000
Customer
Na/S, 250 miles 4.17 miles/kWh
control 7
Na/S, 250 miles 4.17 miles/kWh
convenience
More
Ni/Fe and Pb/acid 15(I miles
convenience
4.17 miles/kWh 8
500 000
2000
Customer convenience
More
Ni/Fe and Pb/acid 150 miles 4.17 miles/kWh
reduction in SCE's average electric rate. The previous studies also showed that EVs would reduce emissions of the air pollutants that lead to high ozone concentrations in southern California. My findings reinforce the general findings from previous investigations. I found that most of the energy to power EVs in southern California will come from generating units fired by natural gas, and I conclude that EVs could lead to improved load shapes, more efficient operation of the SCE system and a reduction in the average electric rate. But there are two unexpected conclusions from the SCE case study. First, I found that SCE may be able to accommodate an unusually large number of EVs without having to add new generating units to its resource plan. And secondly, I found smaller rate benefits than would have been expected from previous studies.
Electric vehicle scenarios for SCE Table 1 lists eight EV scenarios from the case study. We focused on the year 2010, the end of SCE's long-term planning interval and a time frame when advanced EVs might be seen in large numbers in southern California. Table 1 shows scenarios with one million or two million in the SCE service area. The one million EV scenarios are especially interesting because they correspond to recent forecasts presented to the California Energy Commission (CEC). The forecasts assume that car manufacturers will comply with the recently issued
Energy Policy 1994 Volume 22 Number 7
requirements by the California Air Resources Board (CARB) for the sale of ZEVs in California. The CARB calls for 10% of vehicle sales to be ZEVs by the year 2003. With present technologies, the only way to meet the ZEV requirement is through the sale of EVs. Forecasters assume that EV sales would be concentrated in southern California, and the CARB targets translate into approximately 1.7 million EVs - roughly 17% of the population of vehicles expected by the year 2010. The 2 million EV scenarios imply 3.3 million in southem California. With this vision, every third vehicle would be electric. Table 1 shows that the scenarios differ in the total number of EVs, their batteries, the range from an overnight charge, the pattern of night time charging, and the extent of day time or opportunity charging. The three scenarios with two million vehicles were used to examine different night time charging strategies. These scenarios are located in Figure 1 by expanding to 3.3 million vehicles in the entire southern California region. These vehicles would consume around 20 billion kWh/year, so their position would be slightly to the right of the steady advance scenario in Figure 1. The population of EVs was broken down into small vans, large vans, small cars, large cars and light trucks. The general assumption is that EVs will first penetrate the market for vans, followed several years later in the small car market. EVs are assumed to penetrate more slowly in the market for large cars and light trucks.
557
Electric vehicles and electric utilities. A Ford
#_ s 8
Noon
'
'
4pn~
'
'
I~pm
'
'Midnight
"
,4am
8am
o Q_
g_
Figure 2. Electricity demand in the first scenario: two million EVs with advanced batteries? a There is minimal day-time charging and night-time charging is at the convenience of the EV owner.
Their need for electric energy was estimated with a simple spread sheet program. The spread sheet inputs include the number of vehicles in each class, the high/low power settings on the charger, the total energy requirement and the extent of opportunity charging. We then experimented with different assumptions on the timing of night time charging. The spreadsheet calculates the combined demand for electric power from the five classes of EVs and stacks the demand on top of the demand from SCE's other customers. Figure 2 shows the demands expected in the year 2010 from the first scenario. The EV characteristics were taken from a recent study for the CARB. s The small car EV, for example, was assumed to travel 40 miles/day with an efficiency of 4.17 miles/kWh based on the four seater passenger car described in the CARB study. Thus, the energy requirement was 9.6 kWh per day. This energy was assumed to be provided at a
~8
constant power charge of 4.8 kW over a two hour interval. Short charging interval were also characteristic of the remaining classes of E V s . 9 These short intervals are to be expected in EVs with advanced batteries which are used for daily travel that is far below the maximum range of the vehicle. Day time or opportunity charging may be used to provide extra energy to supplement the night time recharging of EV batteries. Because of the extremely long ranges assumed in the first scenario it was assumed that the need for supplemental energy would be minimal. Only 4% of the electric energy for EVs would be provided during day time hours.l° The EVs contribution at 2 pm adds only 130MW to the 26 000 MW peak load, so the EV day time loads are not discernible in Figure 2. The main lesson from Figure 2 involves the night time EV loads. The bar charts suggest that EV demands would land in the evening hours if the timing of
Energy Polio' 1994 Volume 22 Number 7
the nightly charging were left to the convenience of the customer. The upper figure shows that the EV demands would lead to a second peak around 9 pm and that very little of the demand would land in the deep valley available on the SCE system after midnight. This general pattern is a logical result of the short charging intervals. The short intervals, in turn, are the logical result of a scenario in which EVs have a long range of 250 miles but are driven only around 40--60 miles/day.
later in the morning hours and obtain a better blending of the EV loads with SCE's regular loads. Figure 4 shows one possible example for a typical summer day in the year 2010. This pattern was found through iterative search with the spreadsheet program looking for a set of starting times that would allow the night time EV loads to blend nicely with SCE's regular loads. Figure 4 shows a marked improvement in the overall load shape for a typical summer day. Similar improvements were found for a typical winter day on the SCE system.
Managing the loads from EVs The Figure 2 results suggest a need for EV load management, even if the day time charging is minimal. Without load management EV loads may lead to secondary peaks and do little to raise minimum loads appearing in the early morning hours. II Load management options include off peak rates and direct control systems. As an example of an off-peak rate, we examined the impact of a 10 pm EV rate that would encourage EV owners to delay the onset of charging until 10 pm. The effect of this incentive was estimated with the spread sheet program by assuming that EV owners who would normally want to start charging earlier in the evening would delay turning on their chargers until 10 pro. The impact of this simple incentive was a large, secondary peak occurring around 11 pm. This program made the load shape worse, not better. More sophisticated financial incentives might be considered, but planners will wonder whether greater sophistication will lead to improved electricity demand shapes or to confusion on the part of the customer. If rate incentives prove difficult to design, the utility could turn to more direct means of controlling the timing of EV loads. Two broad categories of control schemes were defined for purposes of the scenario study: blind control and smart control. Blind control implies one way communication. The utility would send signals to start the EV's night time charging cycle. But the control system would not receive information on the status of the EVs. Not knowing how long the EV's charging cycle would last would force the controller to send signals early in the evening (ie by 11 pm) in order to ensure that all EVs are charged in time for the morning commute. Figure 3 demonstrates one of the better load shapes that could be obtained when operating under this constraint. This figure shows that the I 1 pm restriction prevents the control system from placing much of the EV load in the deep valley available in the early morning hours. Figure 4 shows the impact of a smart control system in which two way communication is possible between the controller and the charger. The utility can send signals to start charging, and information on the status of the EV can be monitored by the controller. With such a system, 12 the controller can start the charging of EVs
Energy Policy 1994 Volume 22 Number 7
The Southern California Edison system The impact of EV loads like those shown in Figures 2-4 were examined in the context of SCE, an investor owned electric system serving around 4 million customers spread over a 50 000 square mile service territory. The customers and their demands comprise approximately 60% of southern California. SCE uses a variety of resources including nuclear, coal, gas fired units, hydro-electric units, imports from other utilities, demand-side management programs and purchases from qualifying facilities (QFs). The QFs, in turn, employ a variety of generating technologies including cogeneration, wind, biomass, geothermal and solar. The crucial features of the SCE system for this analysis are a heavy dependence on existing, gas fired units and the attractiveness of new, gas fired, combined cycle units in the long term resource plan. 13 The large role of QFs proves to be an important feature in determining the overall impact of EVs. An important question in the case study was whether the EV loads could be served with the resources called for in SCE's existing resource plan. This question was addressed with the help of ELFIN, a computer simulation model for analysis of utility operations and financing. 14 A reference case was established in which ELFIN was used to simulate the impact of SCE's 1990 resource plan. A second case was then considered in which the EV demands were treated as load modifiers within ELFIN's portrayal of system operations. The ELFIN projections of reserve margins and loss of load probabilities are shown in Table 2 to see if EV loads could be served without additional generating resources. The accommodation levels are surprisingly high. For example, Table 2 shows that two million EVs could be accommodated within the existing resource plan if the vehicles were subject to smart control. Table 2 also shows that one million EVs could be accommodated regardless of whether the night time charging was left to the customers' convenience or subject to smart control. The large number of EVs that could be served without the need for additional generating resources may be attributed to a combination of advanced EVs, smart control systems and the reserves built into SCE's 1990 resource plan.
559
Electric vehicles and electric utilities." A Ford
I~
Demandwilhoul EVs ~
Demandfrom EVs
]
20~15 o ~ 10-
illlfIltl
iiili
ii!ii........... noon
' ~.~'
&o,~
',~ni~ht'
4ar~
' ' ~r~~'
Figure 3. Electricity demand from two million EVs with minimal d a y - t i m e charging and night-time charging subject to blind control.
30-
[~
Demar~:::lwknout EVs ~
Demandfrom EVs I
25-
il -
20-
.9
15-
iiiti. . .ii. .iii. i
"
'~I
lO-
...........
~ !i~ . . . . . . . . . .
....... ...... ....... ...... i .........
B
noon
4pro
8pro
midnight
4am
8am
Figure 4. Electricity demand from two million EVs with minimal day-time charging and with night-time charging subject to smart control.
I m p a c t on S C E o p e r a t i o n s ELFIN was also used to estimate which of SCEs generating resources would end up providing the extra energy needed to charge the EVs. The simulations confirmed what previous investigators had found: roughly 90% of the extra energy would come from natural gas fired units. A rough rule of thumb is that two-thirds of the gas fired generation will come from SCE's new, combined cycle units or the repowered units. The other third would come from the older, less efficient gas burning units. The remaining 10% of the electricity generation comes from electricity purchased from utilities in the Pacific North-west and SCE's coal fired units. The small contribution from coal is owing to the fact that the resource plan calls for almost full use of the coal units
560
without any EV loads. Similarly, SCE's nuclear units are expected to be in full operation without EV loads, so they do not contribute to the fueling of EVs. ELFIN was also used to estimate the impact of EV loads on the annual costs to operate the SCE system. In the scenario with two million EVs subject to smart control, for example, annual operating costs increase by around US$1.8 billion or 21%. The cost of higher deliveries of natural gas accounts for the majority of the cost increase. Higher payments to QFs are second to natural gas in accounting for the higher operating costs. The higher QF costs may surprise some readers since QF energy is treated as a load modifier in the production costing calculations. Thus, QF energy cannot be increased or shaped to take advantage of the change in
Energy Policy 1994 Volume ~2 Number 7
Electric vehicles and electric utilities: A Ford Table 2. S u m m a r y of ELFIN results pertaining to accommodation of large EV loads on the SCE system. 2009 No EVs
EVI
EV2
EV3
EV4
EV5
EV6
2000 No EVs
EV 7
EV 8
Number of EVs (million) Strategy for control of night-time charging % energy from day-time charging
2 2 Customer 10 pm convenience incentive
2 2 1 1 0 Smart Customer Customer Smart control convenience convenience control
0.050 0.500 Customer Customer convenience convenience
4
4
4
10
4
4
15
15
EV demand, day peak (MW) EV load, day peak (MW) Peak capacity (MW) 31 726 Peak load 25 875
121 131
121 131
121 131
358 387
61 66
61 66
26 006
26 006
26 006 26 262
Reserve margin (%) Operating reserve margin (%)
18 134
124
26 941
19 27 822 25 941 21 993
22 012
22 127
22.6
22.0
22.0
22.0
20.8
22.3
22.3
26.5
26.4
25.7
9.1
8.6
8.6
8.6
7.6
8.8
8.8
9.1
9.0
8.5
11.64
436.45
0.35
3.79
0.20
0.06
0.03
(I.03
0.06
0.044
0.359
0.031
0.010
0.006
0.006
0.009
Unserved energy (GWh/yr) 0.07 Loss of load probability LOLP (days/year) 0.0t0 Is the LOLP less than 0.10 days/year? yes Is the LOLP less than 0.05 days/year? yes
0.971
13.200
no
no
yes
no
yes
yes
yes
yes
yes
no
no
yes
no
yes
yes
yes
yes
yes
Table 3. Estimated impact of EV loads on the SCE average electric rate (measured in mills/kWh in nominal dollars). Base case 1 ELFIN projection of annual costs for fuel and purchases (million US$/year) 2 Other costs (million US$/year) 3 Total revenue requirement (million US$/year) 4 Electricity sales with no EVs (GWh/year) 5 EV Sales (GWh/year) 6 Total Electricity Sales (GWh/year) 7 Average electric rate (mills/kWh) EVs impact on the average electric rate (%)
Scenario 3 (two million EVs smart control)
8 607
10 434
10 141
10 141
18 748
20 575
9 479
21.2%
872
9 545
10.1% 10 141
1 827 9.7%
Scenario 6 (one million EVs smart control)
19 620
938 10.9%
10 141 872 4.7%
19 686
938 5.0%
95 104
95 104
12 343 107 447
197.1
191.5
load shape created by EV loads. Even though QFs are not projected to supply any of the energy needed for EVs, total payments increase due to QF energy sold to SCE on marginal cost contracts. 15 Marginal costs are driven upward in the EV scenarios, particularly if smart control permits the utilities to remove coal plants from the margin. The higher payments to QFs turns out to be especially important in the scenarios with one million EVs. The increase in QF payments is greatest when EV loads drive coal plants completely off the margin. Table 3 shows the estimated impact on SCE's average electric rate from three of the scenarios. These scenarios were selected for closer examination because the extra
Energy Policy 1994 Volume 22 Number 7
1 827
Scenario 5 (one million EVs customer convenience)
12 343 13.0% -5.6
-2.9
5 956 101 060 194.1
5 956 6.3% 3.0
-1.5
5 956 101 060 194.8
5 956 6.3% -2.3
-I.2
EV loads could be accommodated without having to add new generating units to the existing resource plan. Table 3 begins with the estimated impact on the annual costs of fuel and purchased electricity. The two million smart controlled EVs would cause a 21% increase in this component of the annual revenue requirement. With one million EVs, the annual costs would increase by around 10-11%. The other costs in Table 3 refer to SCE's fixed costs. Since we were working with a fixed resource plan, we assumed that EVs do not change the fixed costs. Thus, this rate calculation ignores utility spending programs that may be needed to provide the charging and distribution infrastructure for EVs. Total revenue
561
Electric vehicles and electric utilities: A Ford
requirements are simply the sum of variable and fixed costs. Table 3 shows that two million EVs would increase the revenue requirement by just under 10%. But EVs are shown to increase electricity sales by 13%, and Table 3 shows the impact on the average electric rate as a 3% reduction. The scenarios with one million EVs show a reduction of 1.5% when charging is left to the convenience of the customer and 1.2% when the charging is subject to smart control. These rate benefits are smaller than what we would expect from previous studies. ~6. The reduction in rate benefits is attributed largely to the impact of EV loads on marginal costs and the subsequent increase in payments to QFs. The higher QF payments are particularly important in the two scenarios with one million EVs. When looking at the final impact on the average electric rate, planners might well consider whether it is better to let EV owners charge the vehicles at their own convenience. The unusual results in Table 3 arise from the large amount of energy expected from QF resources and the assumption that EVs are superimposed on a fixed resource plan.
Transferability of results It is important to remember three key features of the SCE case study when thinking about the transferability of these results. First, and most importantly, we should remember that the SCE analysis is dominated by natural gas. Gas fired units are projected to provide 90% of the electricity needed to charge EVs. The large role of gas stems from the large number of gas fired units on SCE's current system, SCE's commitment to gas fired CCs as the long-term marginal resource and the relative position of gas burning units in the dispatching order. The reader should remember that gas fired units may not supply the EV's electricity in other systems. Thus, the large air pollution benefits expected from EVs in southern California may not apply for other utilities. ~7 A second important feature is SCE's dependence on electricity purchases from QFs. SCE expects QFs and self generation to account for around 38% of energy generation in the year 2010. The SCE study assumed that two-thirds of the QF generation would be purchased on marginal cost contracts. This assumption wiped out a significant portion of the rate benefit expected from EVs. 18 Other utilities which do not expect QFs to be such a large supplier would expect larger rate benefits from EVs. The final caveat to be mentioned here is that distribution system impacts of EVs were not addressed in the study. Work under way at several California utilities suggests that distribution system costs associated with EVs might become sufficiently large to erase the rate benefits identified in this study. I should also mention that the rate impacts identified in the SCE study do not
562
count possible utility spending on infrastructure to help make the EV easier to charge. 19
Summary and the question of incentives The case study shows that SCE could accommodate a large number of EVs within its existing resource plan, especially if the vehicle charging is subject to a smart control system. The high accommodation levels arise from the advanced design assumed for EVs in the year 2010. Specifically, the vehicles are envisioned to travel around 30-60 miles/day, but their range is 250 miles. This combination leads to rather short charging intervals, and smart control systems can allow the EV loads to be stacked on top of SCE's regular loads to improve the overall load shape. Around 90% of the electric energy needed to charge the EVs would come from natural gas fired units. The case study shows that the higher costs to operate the SCE system would be outweighed by the increase in electricity sales. The overall impact of EVs could be summarized as a reduction in the average rate of around 1.5% (for one million EVs) or 3% (for two million EVs). These potential benefits raise an important policy question for utility managers and regulators. Should the utility company contribute financial incentives to promote the sale and use of EVs? This question is reminiscent of questions first raised by Thomas Edison when he suggested that the electric utilities of his day get into the garage building business to promote the sale of EVsfl°. The question of incentives is under intensive study, particularly in the state of California. 21 One of the interesting questions is whether a utility company could offer significant financial incentives to reduce the purchase price of an EV. It might then be possible to finance the incentive program from the reductions in the average electric rate to be expected when EVs lead to flatter loads.
Discussion of incentives When an EV displaces a conventional vehicle (CV), the public may usually expect both clean air benefits and energy security benefits. But the first wave of EVs are likely to be expensive and suffer from limited range. Thus, the vehicle owner must pay the high initial cost and tolerate the limited range if the public is to enjoy the benefits of cleaner air and a more secure energy system. Incentives are usually suggested as a way to assist EV sales during the early years - they may overcome the large private cost to the owner in order to allow the public to enjoy the wider benefits. Incentives may take several forms, including purchase price incentives, electricity rate incentives, infrastructure incentives (ie quick charging stations), and use incentives (ie parking
Energy Policy 1994 Vohmw 22 Number 7
Electric vehicles and electric utilities: A Ford
privileges or dedicated lanes). 22 This article focuses on purchase price incentives. Purchase price incentives might be provided by the car manufacturer, the federal government, the state government, the air management district, the electric utility company, or some combination of the above. The vehicle manufacturers are in the best position to estimate vehicle production costs and how the costs might decline with increased production. The manufacturers are also in the best position to market the vehicles and to estimate future changes in production volumes. For these reasons, the manufacturers are usually viewed as the first source of incentives. The incentives would take the form of differential pricing; the manufacturer would lower the price of EVs and spread the cost across other vehicle types. This approach is described by Jannane S h a r p l e s s Y former chair of the CARB as follows: But there is nothing to prevent the cost of these cars [EVs] from being spread across other model line . . . if car makers can do that now for existing model lines, they can do it for electrics. John Wallace, 24 of the Ford Motor Company, agrees that the car manufacturers can provide incentives: ' A t Ford [Motor Company] we are willing to subsidize - but we would like to have the other stakeholders get involved.' In his presentation to the 1992 Stockholm Conference on the Urban Electric Vehicle, Wallace argues: Since EVs will be more expensive than gasoline powered vehicles, and have inferior driving range (at least until battery technology improves), incentives are needed to assist EVs with the transition into the marketplace. Governments can offer both purchase & use incentives to make EVs an attractive choice of transportation. A possible source of funding for incentives is higher taxes on conventional fuels. Ultimately, EVs must become market-competitive without the support of incentives.
Analysis of incentives The remainder of this article is devoted to an analysis of several proposals for promoting EV sales through purchase price incentives. I begin by describing a modeling approach for simulating the possible impact of incentives over time. I use the model to show the impact of incentives on EV market shares and to anticipate the likely rate impacts if the electric utility were to contribute a significant share of the incentive. I show that the ufility's ability to finance incentives is severely limited if the company wishes to avoid an increase in the average electric rate. I then turn to the state of California as a source of incentives. I focus on one of the more interesting proposals at the state level, a so-called feebate
Energy Policy 1994 Volume 22 Number 7
Alternative vehicles incentive simulation system
Vehicle sector: EVs and CVs compete for market share
Fund manager: CV fees, EV rebates, and a gasoline tax
Utility sector: user specifies a resource plan deterministic dispatch under a rotated load duration curve checked against phase I study EV Costs expensed or captialized Revenue requirement over sales gives the average electric rate
Figure 5. Design of AVISS. program in which EV sales are promoted through purchase price rebates while CV sales are discouraged by imposition of a lee. l show that it would be possible for the state to manage a feebate program in a prudent manner provided the state official is given sufficient latitude in setting the size of EV rebates. I conclude with an analysis of the likely impacts of a cooperative effort by the electric utility and the state to promote EV sales.
Analytical approach Figure 5 depicts a computer model developed to simulate the impacts of incentives to promote the sale of EVs. The model acronym is AVISS - the alternative vehicles incentive simulation system. Figure 5 represents AVISS as a combination of three sectors which operate together over time to simulate the likely impacts of incentives that may be offered to encourage the sale of EVs in southern California. The utility sector simulates the likely impact on SCE based on the user specified resource plan and the number of EVs appearing in the vehicle sector. The vehicle sector represents the decisions by the almost one million southern Californians that purchase a new vehicle each year. Their decision making is based on the findings from a recent, stated
563
Electric vehicles and electric utilities: A Ford 0.4
............................................................................
. . . . . . . . . . . . . . . . . . . . . . . . .
i
,. . . . . . . . . . . . . .
,. . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . .
i..,
............. i
1 990
1995
=
2000
CARB target
[]
2005
2010
No incentive
,
201 5
$4K incentive
2020
o
$8K incentive
F i g u r e 6. Three projections o f E V m a r k e t shares compared w i t h the C A R B goals.
2
. . . . . . . . . . . . . . . . . . . . . . . . . .
........................
4F-~- ,
/
1.5 . . . . . . . . . . . . . . . . . . . . . . . .
.........
/' / ~,"
*'" 1
.
.
.
.
.
.
.
.
.
.
.
.
.
*
. ~:"
0.5
i ! i
i .....
! ......
,
,
1990 --u
•
~r~-~
......
~cr ..............
~
:i
.r.ml'"
L]
.W
-~-,: : r_m7 . . . . . . . . . . . . . . . . .
~ i
,e, m-~3 ~11~ ~ , F ~ --m~" :
:
o .o~-'_~-~-~
, .o~'!-~r-
LT Z
: [y~[]/
~3
/
[] ~
!.--"-- ',
1995 No incentive
I
2000
2005 E3
$4K incentive
2010
2015 •
2020
$8K incentive
Figure 7. Three projections of the number of EVs (in millions) in the SCE area.
preference survey. The third sector in Figure 5 represents the operation of a special state fund which may be used to encourage the sale of EVs through rebates. The fund could be built up from fees imposed on the sale of CVs and/or a tax imposed on the sale of gasoline. Further information on AVISS is given in a research report to the CIEE. 25 The important point to emphasize here is that the flow of information among the sectors is continuous as the model steps through a simulated portrayal of the southern California system over time. The
564
flow of information among the three sectors in Figure 5 leads to the closure of several information feedback loops in the system. 26
Simulated impacts of incentives Figures 6 and 7 show the simulated impact of purchase price incentives by arranging three simulations from AVISS for comparison. This figure shows the fraction of new vehicle sales in southern California captured by a
Energy Policy 1994 Volume 22 Number 7
Electric vehicles and electric utilities: A Ford
US$21 000 EV competing against a US$ 16 000 CV. The EV fuel cost is around 5 cents/mile in the year 2000, but it declines to 4 cents/mile by the year 2010 as a result of improved efficiency. The EV range is 120 miles in the year 2000, and it improves to 160 miles by the year 2010. The CV, on the other hand, will cost 4.8 cents/mile for fuel and have a range of 450 miles. The CV delivers 121 horsepower whereas the EV can deliver only 60 horsepower. These and other attributes of the competing vehicles are explained in the research report to the CIEE. The reference case assumes that no incentives are provided to increase the sale of EVs, so the customer must pay about 30% more for the EV (almost all of which goes to pay for the expensive batteries). The EVs are assumed to first appear on the market in the year 1998, and Figure 6 projects that they would capture about 5% of new vehicle sales in the initial year. Their market share increases steadily over time because improved batteries allow for steady increase in efficiency (miles/kWh) and range (miles per charge) with no increase in the purchase price. Figure 6 overlays the reference case simulation on the CARB target for ZEVs. This comparison shows that EV sales would fall short of the CARB targets once the target is elevated to 10% of sales. 27 One would conclude that incentives are needed if the industry is to comply with the ZEV goals. Figure 6 shows that a US$4000 purchase price incentive would be sufficient to boost EV market shares above the CARB targets throughout the simulation period. With a US$8000 incentive, EV market shares would be increased still further. All three simulations show a steady increase in EV market shares due to the assumption of steady improvements in range and efficiency. By the year 2010, for example, the projected market shares are: reference case: 10%; US$4000 price incentive: 16%; US$8000 price incentive: 24%. These precise estimates are not useful because of the great uncertainty surrounding EV characteristics and consumer attitudes. But it is useful to compare their relative values to learn the extent of the free rider fraction. In this example, the US$4000 incentive program has a free rider fraction of 10/16 or 62%. In other words, 16 out of 100 buyers buy an EV and collect the incentive. But 10 of the 16 would buy the EV without the incentive. This high fraction raises the effective cost of the incentive far above the nominal cost of US$4000 per vehicle. When measured in terms of extra EVs, the cost turns out to be over US$10 000 per vehicle. Figure 7 shows the AVISS projections of the number of EVs that would be operating in SCE's service territory in the three simulations reported in Figure 6. The population of EVs is estimated by combining the Figure 6 market shares with the assumptions that southern Californians will operate their EVs for 15 years and
Energy Policy 1994 Volume 22 Number 7
their CVs for 10 years. Figure 7 shows that SCE would sell electricity to around 1 million EVs by the year 2010 in the simulation with an US$8000 incentive. We could anticipate from Table 2 that SCE would be able to accommodate these vehicles without having to add new resources to its resource plan. In addition, we might expect from Table 3 that SCE could lower its average electric rate with so many EVs on its system. An important question for utility regulators is whether a utility could contribute a significant share of a purchase price incentive and finance their program from the expected reduction in the average electric rate. This question was addressed with a variety of AVISS simulations summarized in the research report to the CIEE. Several simulations were needed to learn the impact of changing the fraction of a purchase price incentive paid by the utility and of changing the duration of the incentive program. I also examined the impact of incentive programs whose costs were recovered from the ratepayer immediately (expensed) or recovered over the presumed life of the EVs (capitalized). The various examples were sufficient to show that it could be quite difficult for an electric utility to provide a major share of purchase price incentives without facing a penalty on the average electric rate. The simulations suggest that rate penalties appear regardless of whether the incentives are large or small, if they are permanent or temporary, if they are expensed or capitalized, or if a socalled band wagon effect is included or ignored in the analysis. I conclude, therefore, that a utility cannot finance a purchase price incentive program from the expected improvement in system operations that will be achieved when EVs flatten utility loads. To strengthen this conclusion, I emphasize that the penalties on the average electric rate appear in simulations that ignore the costs to administer the program and the costs to upgrade the distribution system to accommodate the EV loads.
Discussion of utility incentives The simulated rate penalties lead to the question of whether it is good public policy to have utilities charge their customers more for electricity in order to obtain the public benefits of EVs. It might be argued that almost everyone in the southern California air basin consumes electricity, and almost everyone is exposed to the harmful effects of ozone. We could argue that regulators should allow for penalties in the average electric rate if such penalties were required to achieve the public benefits from EV sales. These penalties could be viewed as a tax on southern Californians who would benefit from cleaner air, and the utility company would play the role of the tax collector. But this line of reasoning leads naturally to the ques-
565
Electric vehicles and electric utilities." A Ford 10000
~ . . . . . . . . . . . . . . . . . .
.~-
i
--: ................................
\
:
5oo
i \:
5000
..................
%
, ..........
,--- ~.~ - - ................
,...........
..........................................................
2500
i " ....
,,,.
1990
"
-
•
i
i
1995
2000
0
$500 fee per CV
:
2005
~
2010
2015
EV rebate
•
--
~..
2020
maximum prudent rebate
Figure 8. Fees and rebates in an AVISS test of a feebate program.
0.3
0.2
~
.
.
.
.
-
.
.
-
.
.
, .
.
~ . . . . . .
.
.
: .
.
.
.Hl--!--,H~C--~
i
,
~ L ~ q - q " "7
I--.~lHI--m--i--m--.H.--I--m
/ E~ ~"
O-;n
,
/
/ ~ W - i
i
.
,
'-i
! ...............
.
.
.0
0.1
.
!
i
!
:
j
1990
2000
1995 -~
--
CARB
target
re--O--
2005
No incentive
2010
2015
•
2020
- with fee/rebate
Figure 9. EV market shares from two AVISS simulations compared to the CARB goals.
tion of whether an electricity tax is the best way to raise money for an EV incentive program. After all, this tax raises the price of an energy form that is clean at the point of use. Would it not make more sense to impose the tax on gasoline or gasoline powered vehicles which are major contributors to the poor air quality in southern California? Imposing the tax on the CV is a particularly
566
interesting idea because it provides car buyers with carrot and stick incentives. The carrot is a rebate on the EV purchase price; the stick is a fee imposed on the CV. Such proposals have been called feebates and feebates are under consideration in the states of Maryland and California as well as at the federal level. 28 Feebates are appealing because the fees and rebates are targeted on
Ene "gy Policy 1994 Volume 22 Number 7
Electric vehi¢'les and electric utilities: A Ford
the vehicle purchase price where one expects the most response from the consumer. But feebate programs might pose interesting management problems for the state official asked to run the program in a financially prudent manner.
Simulating the impact of a state feebate program Figures 8 and 9 show one AVISS simulation which tests the feasibility of a state feebate program. I assume that the program is funded from a US$500 fee imposed on the sale of CVs. Thus, the purchase price of a CV is raised from US$16 000 to US$16 500, and the US$500 collected from southern Californians who buy a CV goes into a state fund. Figure 8 shows that the EV rebate is set at US$8000 initially, so southern Californians would pay US$13 000 rather than US$21 000 for an EV. The combined carrot and stick incentive is US$8500 which turns out to be approximately the same as my recent estimate of the environmental benefit of an EV sold in the 1990s. Figure 9 shows the EV market shares under this particular feebate program. EVs capture well over 10% of the market when they are first introduced in 1998, and their market share continues to grow as long as the rebate is maintained at US$8000. But Figure 8 shows that the EV rebate must be reduced from the US$8000 value around the year 2005 if the state is to maintain an adequate balance in the fund. The maximum prudent rebate shown in Figure 8 first appears on the figure when EV market shares are sufficiently high to trigger a reduction in the size of the rebate. 29 By the year 2005, the rebate must be reduced to US$6000; by 2011, it is down to US$3000. The decline in rebates leads to the loss of EV market shares in Figure 9. But as battery improvements allow [br greater range and efficiency, EV market shares begin to increase again after the year 2005. By the end of the simulation, EVs are capturing 20% of the new vehicle sales; EVs on the SCE system have grown to around 800 000; and the balance in the state fund has been brought down to zero. Since around 400 000 EVs are projected to appear on the SCE system in the reference case, the cumulative impact of the feebate program is a doubling of the population of EVs operating in southern California by the end of the 30 year simulation period. These results demonstrate that the state could operate a feebate program to promote EV sales in a prudent manner. But for this to happen, the fund manager must be given the flexibility of changing the size of the rebates as EVs achieve significant market shares. The simulation demonstrates that the fund could be managed in a prudent manner despite the potential volatility from
Energy Policy 1994 Volume 22 Number 7
relying on CV sales fees as the funding source. If, however, the state does not want to give the fund manager the latitude to rapidly reduce rebates (as shown in Figure 9 for example), the state could increase the size of the CV fee or turn to a tax on gasoline) °
Combining incentive programs California policy makers are looking for an appropriate combination of partners to support the EV market, so the impacts from combining utility incentives with incentives from federal and state governments are important to consider. The question of partnerships is especially important to utility regulators since it is difficult for the utility company to provide purchase price incentives without penalizing the average ratepayer. To help anticipate what might happen when a combination of incentives are implemented in southern California, I used AVISS to simulate an example in which the fund manager follows the US$500/US$8000 feebate discussed previously. The CV fee is set at US$500, and the fund is allowed to build up until the year 1998 when EV sales begin. The initial rebate will be US$8000, but the manager will not allow the rebate to exceed a prudent value. As in the previous example, the prudent value is estimated within AVISS based on the size of the fund, the size of the fee, and recent values of EV market shares. Meanwhile, a US$4000 incentive is assumed to be available for EVs buyers when EVs hit the market in 1998. This incentive is available until the year 2005, and then phased out by the year 2010. SCE contributes half of the US$4000 incentive (the other half could be provided by the federal government through an income tax credit.) When thinking about how the utility and state programs would work together, one should expect their initial impacts to be complementary. In the early years of the programs, for example, the combination of the US$4000 incentive and the fund manager's US$500/US$8000 feebate will provide a total incentive of US$12 500 in favor of the EV. This should lead to larger EV market shares than seen in any of the previous simulations. Also, when the combination of utility, state and federal incentives is viewed as a package, it could be possible that the combination would not lead to rate penalties on the utility system. But the two policies do not necessarily complement each other over the long term. By contributing to higher EV sales around the years 1998-2000, the extra US$4000 incentive will change the situation faced by the fund manager. Higher EV sales will cause a greater drain on the fund, and the fund manager will be forced to reduce the size of the EV rebate somewhat sooner. The AVISS simulations of the package of combined
567
Electric vehicles and electric utilities." A Ford
incentives demonstrate that it would be possible for the utility to contribute to the package without imposing a penalty on the average rate payer. (I should emphasize again that this analysis does not count the cost to administer the program or to upgrade the distribution system to accommodate the extra EVs.) But the AVISS simulations also demonstrate that the longer-term impacts of a combination of incentives would be disappointing. The disappointment arises when the increase in EV market shares induced by the utility incentive forces the state fund manager to cut back on the size of the EV rebate. In other words, the addition of a utility incentive to a state feebate program would end up erasing part of the benefits expected from operating a feebate program alone.
Final conclusions and transferability of results The general conclusions to be drawn from my analysis of incentives are that: (i)
(ii)
(iii)
(iv)
(v)
incentives appear to be needed if EV sales are to reach the goals set by air pollution regulators in California; the effective cost of an incentive program can be much higher than the nominal cost due to a significant free rider fraction; it is difficult for an electric utility to contribute a major share of an EV purchase price incentive without imposing a penalty on the average rate payer; it is possible for the state to boost EV sales through a feebate program and to operate the program in a financially prudent manner; and finally the long-run impact from combining state feebates and utility purchase price incentives can be disappointing because the utility incentive would end up erasing part of the benefits of the state program.
These conclusions are quite general. They depend on the overall structure of the AVISS model, not on precise parameters used in the simulation analysis. Thus, they could well be transferred to the situation faced by other utilities or states that are studying the California example. When thinking about the transferability of these findings, it is important to remember three important aspects of the analysis. First, the analysis focuses on the SCE system, and the lessons from SCE's system may not transfer. For example, other utilities may not rely as heavily on purchases from QFs based on marginal cost contracts as SCE. If so, the rate benefits of EV loads would be significantly larger than the benefits reported here. Thus, it might be possible for such a utility to contribute to EV purchase price incentives without the rate penalties
568
mentioned here. In addition, other utilities may not rely as heavily on natural gas fired generation as SCE. If so, they would expect different air pollution benefits from EVs. Indeed, for coal dependent utilities, the idea of promoting EV sales to improve regional air quality might not make sense. Second, the analysis summarized here is based on a simplified picture of the market in which EVs compete one to one against CVs. A more complete portrayal would include other choices like vehicles fueled by compressed natural gas or methanol. With an expanded picture of the market, we would expect to see a decline in EV market shares relative to the results shown here. Thus, we would expect a greater need for incentives if EV market shares are to reach the CARB goals. And third, we should remember that the vehicle sector coefficients are taken t¥om the early results of the stated preference survey. These coefficients should be treated with caution because of the relatively small sample size and the inevitable problems respondents face when answering questions about vehicles for which they have little practical expertence: . ~l
The case study findings are documented in report to the research sponsor, the California Institute for Energy Efficiency (CIEE): Andrew Ford, The Impact of Electric Vehicles on the Southern Califi,'nia Edison System, available from the NTIS, 5285 Port Royal Road, Springfield, VA 22161, USA (July 1992). The report is one of several reports from a project on the Assessment of Natural Gas and Electric Vehicles. Further information on the overall project may be obtained from Professor Dan Sperling, Director, Institute of Transportation Studies, University of California, Davis, CA 95616, USA. The case study results and the views given in these articles are my own. They do not represent a position taken by the CIEE or by the gas and electric utilities which support the CIEE.
IOrganization for Economic Co-Operation and Development and the Swedish National Board for Industrial and Technical Development, The Urban Electric Vehicle, Proceedings of an International Conference, Stockholm, Sweden, 25-27 May 1992. 2 According to Donald Saxman, editor of the Battery and EV Technology newsletter, EVs are especially close to commercialization in the UK, Germany, Denmark and the Netherlands. These countries are said to have sophisticated commercial EV manufacturing and supply infrastructure. The automobile market in Europe is also better suited for EVs because of higher gasoline prices and governmental subsidies (Donald Saxman, '1990-2000: decade of the electric vehicle?', Energy, August, 1992). Japan's EV program is described in Masakazu Iguchi's, 'Market expansion programme of electric vehicles planned by the Ministry of International Trade and Industry, Japan' at the Stockholm Conference. 3Emissions benefits for the regulated pollutants are estimated in Chapter 7 of the SCE case study. An EV sold in 1994, for example, is estimated to eliminate over 100 lb of hydrocarbons, over 100 lb of nitrogen oxides and over 1000 lb of carbon monoxide during its 15 year operating life. Assigning monetary values to reduced emissions in the South Coast Air Quality Management District leads to a total value of just under US$9000 per EV. The US$9000 figure will decline with EVs built further in the future as a result of expected improvements in the conventional vehicles to be displaced. The US$9000 figure is specific to the South Coast Air Quality Management District and is highly sensitive to the method used to assign monetary values.
Energy Policy 1994 Volume 22 Number 7
Electric vehicles and electric utgities: A Ford 4The most difficult air quality problem in southern California is ozone. The SCE case study confirms what previous investigators have found that EVs, when taken alone, do not contribute to a major reduction in peak ozone concentration. For example, the probable impacts in the first year of operation of one million EVs sold in the early 1990s is a 1% reduction in the emission of hydrocarbons and nitrogen oxides. The peak ozone concentration would be reduced by around 0.7%. 5Mark DeLuchi, Quanlu Wang and Daniel Sperling, 'Electric vehicles: performance, life cycle costs, emissions, and recharging requirements', Transportation Research, Vol 23A, No 3, 1989. 6The term scenario originated in the field of drama and was then borrowed for war gaming and large scale simulations. Scenario analyses are becoming widely used in the energy field as planners search for more effective means to plan for a highly uncertain future (see Peter Wack, 'Scenarios: uncharted waters ahead', Harvard Business Review, September 1985; Southern California Edison Company, Strategies for an Uncertain Future, March 1988). 7As interest in EVs builds over time, however, utility forecasting departments have become interested in the best way to include EV loads in electricity demand forecasts. Some of the forecasting issues are raised in the California Energy Commission's Demand Issues for the 1992 Electricity Report Plvceedings, August 1991. SReport to the California Air Resources Board, 1990 Electric Vehicle Systems Update, draft copy dated 12 April 1990, cited with the permission of the CARB. 9The large electric van, for example, was powered by a Ni/Fe battery, was used to travel 55 miles/day, required 33 kWh in the night time charge, and was assumed to receive this energy at I1 kW over a three-hour interval. 1°The need for supplemental energy will depend on the daily travel requirements and the range of the vehicle, Travel requirements, in turn, differ widely from one family to another, and planners might expect EVs to sell predominantly to families which expect to assign the EV to daily travel of a limited and predictable nature. The minimal assumption for day time charging was implemented by assuming that 1% of the EVs would be drawing an opportunity charge during the hours from 8 am to 5 pro. ~Leaving charging to the convenience of the customer can lead to even worse problems in winter days, especially if the customer begins the nightly charging cycle shortly after returning home from the daily commute. This possible problem is highlighted in a staff report from the California Energy Commission on the Analysis of the Potential -
Electricity Demand, Electricity Supply and Emissions Impacts of Electric Vehicles, 10 February 1992. The need for EV load management is not unique to California utilities. For example, the benefits of EV load management is explained in the context of the Dutch electric system in Pier Boonekamp's Stockholm Conference presentation on 'The electric car and the public power system'. ~=A smart control system is easiest to envision as part of a two way radio communication system. Such systems are expensive, but they offer multiple benefits such as load monitoring, load management and automation of service on the distribution system (see the SCE Research Newsletter article on 'NetComm matures as advanced communication and metering system', Vol 19, No 4, 1990). But smart control might also be implemented directly through the design of a sophisticated charger which can be programmed for off peak charging. Specifications for the chargers preferred by General Motors are given in an address by Kenneth Baker on 'The importance of infrastructure,
Proceedings of the Electric Vehicle Policy and Technology Con/erenee, Los Angeles, California, 5 ~ March 1992. And finally, limited control of the EV loads might be achieved by installing a low cost sensor/switch on the distribution line. When the sensor measures high current flows (ie when an air conditioner turns on), the switch would cut power to the EV charger. For further information on charging options, see 'Charging up for electric vehicles', EPRI Journal, June 1993. ~3SCE's resource plan calls for major investments to repower existing gas fired units as well as investments in new, combined cycle generating units fueled by gas. Their reliance on new gas fired generation is typical of utility plans across the country. The question of whether gas pipeline companies can deliver the gas needed for electricity genera-
Energy Poli~'y 1994 Volume 22 Number 7
tion is addressed in "Natural gas for utility generation', EPRI Journal, Vol 17, No 1, 1992. 14ELFIN was originally developed by the Environmental Defense Fund (EDF). It is widely used in California, especially in proceedings before the California Energy Commission and the California Public Utility Commission. ELFIN is explained in an EDF report ELFIN Dispatch Algorithms and Simulation Methods, March 1988 and in a side by side comparison with a more complicated model of utility production costing, CEC Consultant Report, Production Cost Modelling of the Southern California Edison System, 1985. ~SThe cost calculations assume that two-thirds of the QF energy would be purchased on marginal cost contracts while one-third would be based on rates that do not vary with changes in SCEs marginal cost. 16The best previous study to estimate rate impacts was Glenn Ducatt's SCE report Electric Vehicle Deployment on the Southern CaliJornia Edison System, 15 March 1989. Ducatt estimated that around 600 000 large electric vans could be accommodated with the previous resource plan, and the EV loads could lead to a 3.5% reduction in the average electric rate. ~7Utilities with heavy use of coal-fired power plants should expect quite different air pollution benefits from EVs. Two studies by the Environmental Defense Fund (EDF) show the differences. First, Dianne Fisher has analyzed greenhouse gas emissions fi'om coal, natural gas and other sources of electricity generation to serve the EVs. She estimates that an EV charged from gas fired power plants would reduce greenhouse gas emissions by 36%. But if the EV's electricity came from coal fired plants, the EV would lead to a 5% increase in greenhouse impacts (Dianne Fisher, Reduein~ Greenhouse Gas Emissions With Alternative Transportation Fuels, April 1991 report available from the Environmental Defense Fund, 5655 College Ave, Oakland, CA 94618). In more recent analysis, the EDF is examining the emissions associated with EVs operated in Los Angeles and served by the Los Angeles Department of Water and Power. The Los Angeles system makes greater use of coal fired plants, so the EDF results provide an interesting comparison with the SCE results reported here (Francis Chapman, Chris Calwell and Diane Fisher, What's the change?, June 1994 report available from the Environmental Defense Fund). ~SThe impact of marginal cost contracts was most dramatic in the third EV scenario, the case with two million, smart controlled EVs on the SCE system by the year 2010, The average rate benefit in the base case was estimated at just under 3% in the year 2010. But if all the PURPA purchases were based on fixed contracts, the EV's rate benefit would be around 4.5%. ~9.Fitzgerald estimates that transmission and distribution costs could amount to 30-45% of EVs impacts on investments by the Pacific Gas and Electric Company. Daniel Fitzgerald, 'Making EVs PG&E's business', presentation to the Electric Vehicle Policy and Technology Conference, Westwood Marquis Hotel, Los Angeles, California, 5 March 1992. Contact Joshua Newman and Associates, 3110 4th Street, Santa Monica, CA 904(15, USA for information on proceedings. For further information on how utilities might anticipate distribution system impacts, see op eit, Ref 12, EPRI Journal. 2°Edison's contributions to the early history of EVs is described in Rudi Volti's 'Why internal combustion'?', Invention and Technology, Fall 1990. Volti explains how greater speed, acceleration and range of internal combustion vehicles won out in the competition with EVs and with steam vehicles. Volti also explains that EVs were less popular because the electric utility companies were not as aggressive as the oil companies in setting up charging stations to deliver their fuel. 2tThe role of the gas and electric utilities in promoting the use of low emission vehicles is under intensive study in by the California Public Utilities Commission. The California legislature has found that utility ratepayer funds may be used to promote the use of alternative transportation fuels and has called on the Commission to determine the extent of such funding. The California utilities, in turn, have submitted concrete proposals based on a six-year funding cycle which include financial incentives to lower the purchase price of EVs. For example, SCE has proposed a US$1500 EV battery incentive to promote EV sales (see SCE, Proposed EV Programs, Exhibit No. SCE-2, November 1993, Proceeding 1.91-10-029/R.91-10-028 before the Public Utilities Commission of the State of California).
569
Electric vehicles and electric utilities: A Ford 22A March 1992 conference on electric vehicle policy was held in Los Angeles in March of 1992, and incentives were the subject of several round table discussions. The conference organizers have summarized these discussions in their presentation to the Stockholm Conference (Derek Chart and Gerald Mader, 'Conclusions from the Electric Vehicle Policy and Technology Conference'). 23jananne Sharpless's remarks are published in the Proceedings ~?fthe Electric Vehicle Policy and Technology Conference in Los Angeles, 1992. 24john Wallace's remarks also appear in the 1992 Los Angeles conference proceedings. 2-~Andrew Ford, Design and Testing of the Alternative Vehicle Incentive Simulation System, draft report to the California Institute for Energy Efficiency, March 1993. 26 For example, the price of electricity is influenced by the utility production costs; production costs are influenced by the size of the EV loads; EV loads are influenced by EV sales; and EV sales are influenced by the price of electricity. 27The Figure 6 market shares arise in a simple model where EVs are the only competitor to conventional vehicles. We are expanding AVISS to include a variety of competitors (ie hybrid electric vehicles, methanol vehicles, compressed natural gas, and flexible fueled vehicles). The preliminary results from the expanded model suggest that EV market shares would be even lower than shown in Figure 6. 2aTbe state of Maryland enacted a program where a fee is imposed on gasoline cars with poor fuel efficiency in order to provide rebates for cars with higher efficiency. But the Maryland bill has been blocked by a Department of Transportation ruling that their feebate program is federally preempted. In California, a feebate program called DRlVE+ passed the legislature in 1990 but was vetoed by the governor. DRIVE+ focuses on improvements in vehicle emissions through fees on vehicles with high emissions of hydrocarbons, nitrous oxides, carbon monoxide as well as carbon dioxide. These fees then provide the revenue for rebates on vehicles with lower emissions. Amory Lovins endorses feebates for their general properties (market driven, technology pull and simple to administer). He thinks feebates hold the ultimate promise of replacing prescriptive regulations, and he states that 'The feebate concept embodied in Drive+ is the single most important energy policy initiative now being considered at any level of government anywhere in the United States' (Lovins's views on
570
Supercars: The Coming Light Vehicle Revolution are available from the Rocky Mountain Institute, 1739 Snowmass Creek Road, Snowmass, CO 81654-9199, USA). His views on feebates are explained in a 6 July 1991 memo.) The DRIVE+ proposal is described in a July 1989 report from the Lawrence Berkeley Laboratory by Deborah Gordon and Leo Levenson entitled DRIVE+. And a comprehensive discussion of feebates aimed at promoting fuel efficiency is given in a September 1992 report by John DeCicco from the American Council for an Energy Efficient Economy entitled Feebates for Fuel
E~'onomy. 29The maximum prudent value of the rebate is derived internally in AVISS to allow the state fund manager to bring the balance in the fund to zero over the life of the program. I do not assume that the fund manager has to balance the cash flows into and out of the fund in each and every year. This would require forecasting the EV and CV market shares with a p~ecision that is not possible. Rather, I assume that the fund manager will be allowed to build up a balance in the fund by charging fees on CV sales prior to the start of an EV rebate. Then the fund may be drawn down gradually over the life of the program in a prudent manner. A key feature of this simulation is that the simulated fund manager is not required to have a crystal ball that predicts future EV sales with perfect accuracy. Rather, the maximum prudent rebate may be estimated from past EV sales information that would be readily available. 3°A gasoline tax is a more stable source of funds since the tax applies to all the vehicles in California that are consuming gasoline rather than just the new vehicles sold each year. I used AVISS to test a 10 cents/gallon gasoline tax to support EV rebates, and I found that EV market shares could be increased in a manner similar to the US$500/US$8000 feebate program. To phase out the rebates over time, state officials would have to remove the gasoline tax and then lower the rebates to bring the balance in the state fund to zero. ~tThe survey of attitudes toward alternative vehicles is continuing under funding provided by California utilities and the California Energy Commission. The initial coefficients are reported by D Bunch, M Bradley, T Golob, R Kitamura and G Occhiuzzo, 'Demand for clean fuel personal vehicles in California: a discrete choice stated preference survey', to be published in Transportation Research. For information on the more recent survey results, contact Thomas Golob, Institute of Transportation Studies, University of California, Irvine, CA 92717, USA.
Energy Policy 1994 Volume 22 Number 7