ML20073D727
| ML20073D727 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 04/04/1983 |
| From: | Taylor V UNION OF CONCERNED SCIENTISTS |
| To: | |
| References | |
| ISSUANCES-SP, NUDOCS 8304140168 | |
| Download: ML20073D727 (148) | |
Text
...
r e
UNITED STATES OF AMERICA NUCLEAR REGUI.ATORY COMMISSION before the ATOMIC SAFETY AND LICENSING BOARD In the Matter of CONSOLIDATED EDISON OF NEW YORK, INC.
(Indian Point Unit Number 2) 50-247-SP Docket Nos.
)
50-286-SP POWER AUTHORITY OF THE STATE OF NEW YORK (Indian Point Unit Number 3)
TESTIMONY ON THE ECONOMIC COSTS OF CLOSING INDIAN POINT Testimony of Vince Taylor April 4, 1983 i
Union of Concerned Scientists 1384 Massachusetts Avenue j
Cambridge, Massach2setts 02238 l
B304140168 830404 PDR ADDCK 05000247 T
Q Please state your name, occupation, and business address.
A.
Hy name is Vince Taylor.
I am a senior economist for the Union of Concerned Scientists. My business address is 21 Elm Ave., Richford, vermont, 05476.
Q.
Please state your prof essional qualifications, t
A.
I have attached a summary of my professional experience and publi-cations as Exhibit I.
Q.
What are the purposes of your testimony?
A.
The purposes of my testimony arer
- 1) to explain how the economic costs of closing the Indian Point nuclear plants should be calculated,
- 2) to provide an informed estimate of these costs, and 3) to review and comment upon previously made estimates of these costs, with particular anphasis on a study of the Rand Corporation.
Q.
Would you briefly summarize the previous studies tt.at you will consider in your testimony.
A.
There are three studies:
- 1) In a November 1980 report, the General Accounting Offica (CAO) estimated that the near-term costs of shatting j
down Indian Point 2 and 3 nuclear plants would exceed $600 million in the j
first year and that the 15 year cost to New York City ratepayers of a shutdown could be over $18 billion.
The near-term cost represented the i
CAO's estimate of the cost of replacing the lost nuclear generation with oil-fired generation, based upon information supplied by Consolidated Edison of New York (Con-Ed). The $18 Sillion fif teen-year figure represented estimates of the incremental revenue requirements (those above requirements 1
with Indian Point open) of Con-Ed.
The incremental revenue estimates were from a study performed for Con-Ed by Stone and Webster Closing of Indian Point would also raise costs for the Power Authority of the State of New York (PASNY) and the CAO estimated that incremental revenue requirements for PASNY could amount "to as much as $600 million annually." (CAO Report, p. i.)
No duration for this incremental cost was specified.
- 2) In a response to the CAO report (Exhibit II, attached). prepared for the Union of Concerned Scientists (UCS), co-authored with Charles Komanoff, I documented that the near-term costs of replacing the lost nuclear generation would be approximately $340 million per year and i
..7,.,,.-,
e
. t ha t the upper limit of these costs in the longer-run would be $292 million per year (costs in 1980 dollars) if new coal capacity were built to replace the closed nuclear capacity.
Closing Indian Point would also avoid prospective major one-time expenditures to meet new saf ety require-ments and for repairs to steam generators. I estimated these avoided expenses (savings from closure) to be $375-575 million, with a significant chance that the upper end of this range would be exceeded.
Because the CAO report did not make a comprehensive estimate of the total expected costs of clocure, I did not attempt to make such an estimate in the UCS response to the CAO report.
If my upper-limit estimates of replatement generating costs were summed for 15 years and the one-time savings subtracted, the figure would be $4.4 billion (1980 dollars),
compared to the $18 billion plus figure cited by the CAO.
I wish to stress, howev er, that $4.4 billion does not represent an accurate estimate of the economic costs of shutting Indian Point. For a number of reasons to be explained later, the economic cost would be substantially less than $4.4 billion.
The CAO has made no public response to the UCS critical evaluation.
3)
Subsequently, PASt:Y hired the Rand Corporation to make a new estimate of the costs of closing Indian Point.
Rand concluded that the costs "may lie anywher; between $7.7 billion and $17.4 billion" (in d iscounted 1980 d'ollars).
The authors state that their estimate can be compared to " implicit total cost estimates of $9.3 billion.
, and
$2.5 billion" by the CAO and UCS, respectively.*
The Rand report also attributes an estimate of $4.0 billion to the Congressional Research Service (CRS), based on its published response to the original CAO report.' The CRS analysis focused on the question of how ratepayers and stockholders of Con-Ed should share any costs incurred by a shutdown.
It did not purport to be a comprehensive analysis of the costs of shutdown and, therefore, did not consider many questions relevant to such an analysis. Rand produced the "CRS estimate" cited in its report by translating the fragmentary, qualitative criticisms of CRS into quanti-tative adjustments to CAO's estimates. The resulting figure is,not a meaningful, independent estimate of the total costs of closure, as Rand attanpts to suggest. To include it in the comparisons provides a false sense of comprehensiveness.
I therefore omit the CRS estimate in the foilowing discussion.
. Q.
Please describe Rand's methodology more fully.
A.
The Rand estimates were intended to be comprehensive estimates of all of the additional costs (af ter netting out savings) that would be borne by the U.S. economy if the Indian Point units were " shut down bmaed-intely rather than at the end of their normally scheduled lives." (p.3.)*
Rand terms these added costs incremental costs and distinguishes them from expenses that have already been incurred (sunk costs) or that would be incurred whether or not Indian Point is closed (unavoidable costs). Rand points out, correctly, that only incremental costs are relevant to the decision on whether or not to close the units.
Rand presents estimates of the total sum of the incremental costs of closure in terms of disdounted 1980 dollars. By using 1980 dollars Rand eliminates inflation from the analysis. All future costs are calculated in dollars of the same purchasing power (that of 1980). Add itionally, future costs are discounced back to 1980 at the compound interest rate of 5 percent per year. Thus a dollar of cost incurred in 1990 is discounted to $.61 ($1/1.0510) in calculating the discounted cost of closure.
The rationale for discounting when summing costs and savings that occur at different times is that dollars in hand today are worth more (even aside from inflation) than those to be received (or paid) in the future.
Fundamentally, this is because a dollar productively invested today will produce over its lif etime more than a dollar of additional output, that is, it will earn a positive rate of return.
If cociety can invest and earn a 5 percent rate of return per year, it would be equally well of f with an extra $.61 today or an extra $1.00 ($.61 x 1.0510) ten years hence.
The rate of return on investment (af ter abstracting from inflation) is the appropriate rate to use in discounting. Because the economy has been so volatile during the past decade (and the future so uncertain),
the choice of discount rate to use in the present case is not a simple matter. Rand considers a range of discount rates and uses 5 percent per All page ref erences are to Reference 3 unless otherwise f.oted.
i i l
year in its final estimates and comparisons.
I accept the 5 percent figure as a reasonable basis for comparing the various estimates of the cost of closure.
Rand considered four categories of costs in making its estimates of the total cost of closure:
1.
Generating costs include all of the extra costs directly resulting from generating the required electricity at another, more costly, facility af ter Indian Point is shu t down.
2.
One-thne costs include all of the one-time costs of decommissioning Indian Point and disposing of the spent fuel as well as all of the one-time savings associated with the now-planned activities and back-fits that will not be needed if Indian Point is shut down.
3.
Business costs include all of the increased financial costs, the increased returns that will be demanded in both the bond market and the stock market if Coned and PASNY attempt to remain viable and to continue to supply electri-city to their service area af ter the Indian Point units are shu t down.
4.
Secondary costs include the indirect costs--the second, third, and later-round effects that are induced, or flow from, the imposition of the other costs.
(p.4.)
High and low estimates were made for each category and combined to yield total high and low estimates.
Q.
Did Rand accurately represent the estimates of the original GAO and UCS estimates and compare them f airly to the Rand estimates?
A.
No, Rand did not.
It attributed " implicit" total cost estimates of closure of $9.3 billion to the CAO report and $2.5 billion to the UCS repcrt and compared them to its own range of $7.7 billion to $17.4 billion.
This comparison can be interpreted as lending strong support to the competence of the original CAO estimate and to repudiate the UCS estLmate.
The estimate attributed to the CAO appears prudently conservative compared to those of Rand, whereas the one attributed to UCS lies far below the bottom of Rand's range. The superficial impression conveyed by the summary comparison is not supported by the body of the document.
A.carsful reading of the entire report reveals that on almost every point of dif f erence between the CAO and UCS on fact and economics. Rand sides with UCS.
In making its own estimates of the cost of closure, Rand essentially 1
accepted the UCS estimates of the annual cost of replacement power and the l
l o
f net, one-time savings f rom closure. These were the major items considered l
1 in the UCS analysis. These provide, as stated in the UCS report, an upper i
limit to the costs of closing Indian Point.
{
The Rand summary misleads the reader by representing that the CAO, UCS, and Rand estimates given there are comparable (arrived at by " applying a common set of assumptions and converting to common dollars," p.3).
l They are not.
Rand 's estimates include a major item, " secondary costs" (with a range of $2.3 to $6.9 billion), not even considered by the other 4
estimators, and Rand's total cost estimate spans 25 years, whereas the figures cited for the CAO and UCS over only 12 years. Furthermore, the figure cited as the "CAO estimate" is not based on the estimates of the cost of replacement generation originally reported
- by the CAO.
These l
estimates were first revised sharply downward by Rand (ironically enough, j
by making corrections suggested by UCS) before the summary figure was calcula ted. These elements of non-comparability expand the Rand estimate t
and shrink the CAO estimate from its original value so that the two appear in reasonable agreement.
l Q.
Do you accept as accurate the $2.5 billion estimate of the total I
costs of closure attributed to UCS by Rand?
A.
No.
This estimate was generated by Rand from estimates of replace-ment generating costs and one-time savings presented in the original UCS report. These were labeled in that report as upper limit costs because they did not take into account consumer responses to higher electricity prices, responses that would reduce the total gest to the economy.
a Taking these responses into account would produce a much lower estimate of total cost of closure, Q.
Have you prepared a comprehensive estimate of the total costs of closing Indian Point, taking into account the cost-reducing responses of consumers to higher electricity prices? If so, what is its value?
A.
I have prepared a comprehensive estimate of costs of closure that j
is directly comparable to Rand's estimated range of $7.7 to $17.4 billion (discounted 1980 dollars). My estimate is $0.8 to 0.9 billion j
(discounted 1980 dollars). The basis for this calculation is described fully later in my testimony.
Q.
Could you explain why your estimate is' so much lower than Rand 's?
A.
My estimate is based on the same assumptions as Rand 's with respect
l to plant lif etime, treatment of inflation, and discount rate. There is little dif f erence between my and Rand's estimates of replacement generating costs and one-time savings. The great diff erence arises in Rand's categories of business and secondary costs. Whereas Rand assigns large positive values to these categories ($3.3 to $12.8 billion for the two together), I find no evidence or reason to believe " business costs" (beyond those accounted for under replacement generating costs) would be significant, and I estimate that rather than resulting in added costs, the secondary eff ects (responses of consumers) would off set much of the initial cost im pac t. Thus, while consumers would initially bear the costs of replacing all of the generation lost by closing Indian Point, they would, over time, red 0ce their consumption of the higher-priced electricity, reducing the amount (and cost) of replacement generation. The reductions in electricity use would occur both because consumers would use electricity more ef ficiently and because they would shif t some expenditure away from electricity to other goods (pure conservat' ion). My c t hna t e, the details of which are presented later, is that secondary responses will off set $3.2 billion of the estimated replacement generation costs of $4.4 billion (in discounted 1980 dollars). Taking into account the discounted value of one-thne savings ($0.3 to $0.4 billion), I arrive at the comprehensive estimate of the costs of closure of $0. 8 to 0.9 billion.
Q.
Do you believe that actual closure costs could f all below $0.8 billion?
A.
Yes. To make my estimate closely comparable to Rand's, except in the treatment of business and secondary costs, I accepted several assumptions that are not realistic and that bias my cost estimate upwards:
1)
I used Rand's assumption that both Indian Point reactors would, in the absence of a regulatory order to shut down, operate for 25 more years (through 2005). This would bnply lifespans of 32 years for Indian Point 3.
No commercial reactors have operated *his long, and several early power reactors (including Indian Point 1) have been shut down perma-nently well before this age by a combination of rising safety requirements and declining operating reliability; thus this is a highly quesedonable a ssumpt ion.
i 2)
I considered only near-term, one-time savings from closure in the original UCS report, and I did not revise this estimate of one-time savings.
More distant major repairs, such as steam-generator replacement or treatment
I for reactor-vessel embrittlement, and the imposition of major new saf ety requirements would be nearly certain to occur if the plants operated for another 25 years. These more distant one-time costs would be avoided by c lo sure.
- 3) 1 used the historical average capacity factor of Indian Point reported in the original UCS report (57 percent) in calculating replacement genera t ion. This average was based on data through December 1980.*
Adding data for 1981 reduces the cumulative average for both reactors to 52 percent. Moreover, Indian Point 3 was shut down in late March 1982 for repairs to steam generators and will not return to service until September, at the earliest; thus it will be hard pressed in 1982 to match the dismal 36 percent capacity factors of 1980 and 1981. Thus the outlook is for a still lower cumulative capacity factor at the end of 1982. Using the cumulative capacity f actors through 1981 in the calculations would reduce the estimated incremental costs of closure (and secondary off sets to these costs) by 9 percent.
Q.
Can you explain more fully the diff erences between your and Rand's est ima tes?
A.
Yes. I would like to explain these dif f erences as part of an overall review of the Rand report. In the course of this review, I will:
- 1) note Rand's disclaimer of any independent expertise with respect to the data and methodologies underlying its estimates of the a
cost of closing Indian Point, 2) explain why Rand's comparisons of estimates of incremental generating costs are misleading, 3) show that Rand rejected most of the CAO cost estbmates and accepted many of the UCS estimates, 4) show that Rand's estimates of " business costs" are without foun-d a t ion, 5) explain how Rand erred in its treatment of " secondary costs,"
6) explain and document the economically correct way to calculate the ef fects of secondary responses by consumers on the compre-hensive costs of closure, and 7) present a detailed calculation of the correctly estimated total costs of closing Indian Point.
Excluding initial operation for Unit 2 in late 1973 and for Unit 3 in late 1976, during which the two units performed poorly and well, respectively.
o
?
-8 l
)
Rand's Disclaimer of Competence 1
j At the outset, Rand disclaims any independent expertise with respect
]
to the data or methodologies that underlie its est.imates of the cost of j
closing Indian Point:
l At the request of the Power Authority (of the State of New i
York), all materials used in constructing the cost estimations were obtained from the open literature. The authors of the report reviewed the available literature on the economic' impact of closing the Indian Point facilities.
In so doing. they made no attempt to verify either the underlying facts or the methodologies used in any of the studies cited herein. and they do not comment on the accuracy or reliability of those f acts or methodologies.
(Preface, p. iii, emphasis added.)
Having so restricted thanselves, the authors would seem to have disqualified themselves from making any meaningful contribution to the debate over the costs of closing Indian Point, since the" accuracy and
}
reliability" of the competing faces and methodologies are the essence of the matter at hand.
If Rand cannot attest to the accuracy of the facts on which it bases its estimates, of what value are its estimates?
i Although, in fact, the Rand researchers do exercise their own judgement in choosing among competing " facts" and estimates, they retreat to their I
"non-expert" position at crucial points.
For example, in discussing business costs, they note that the only analysis of such costs in the reports reviewed were provided te the CAO by " Stone and Webster, a manage-ment consulting firm specializing in utility finance *," and that "the CAO 2
j documents that analysis strictly as a ' black box ' exercise," meaning that 1
i no supporting evidence or methodology were given (p.35).
They continue,
)
"[B]ecause of the limited documentation of that (Stone and Webster) analysis (of business costs), we have little basis for developing a reasonable and defensible estimate of their magnitude at this. time."
(p.36.)
This does not, however, deter Rand:
"We do believe these costs could easily l
Not mentioned but certainly relevant, Stone and Webster is also one of the nation's largest architect-engineering firms engaged in constructing nuclear plants, including the two nuclear plants still under construction in New York State.
4
.,,.a
.-w n..
e
--e.
-w n
a
.r-.-
.s
-9_
amoant to $1 billion.
So we use that as our lower estimat e." (p.36, anphasis added.)
Bel,ief is the only basis given for this lower estimate' l
The Rand researchers base their upper estimate of business costs on the undocumented, unsubstantiated Stone and Webster estimate (of $5.0 billion),
i which they adjust upward
- to $5.9 billion.
(p.3 6.)
In this instance, Rand's selected use of a "non-expert" position allows it to accept and use an unsubstantiated figure that constitutes one-third of its high estimate of the total cost of clo. ora.
Throughou t the report, the restrictive ground rules set out by the Rand researchers are used to justif y the " inability" to go beyond the CAO, UCS, and CRS ' reports for additional information 3 use to choose among competing assertions.
For example, in examining estimates of replacement generating costs, Kand notes that, "The CAO does not explain or question the CE model it uses. Thus the causes of the large cost [ estimate used by the CA0] for substitute generation remain obscure." (p.18.)
Rather than pursuing the matter further, however, Rand states "In the absence of other information, our estimates must rely on the CE model output and the ir. dependent USS estimates." (p.23.)
The UCS estimates were backed up by extensive, published, publicly available documentation. Rand could have pursued this documentation and used it to provide an independent, expert judgement of whether the CAO or UCS estimates were more realistic. This hand authors chose not to do.
- Rather, they accepted the undocumented CAO estimate as the initial basis for their "high estimate" of replacement generating costs ** and the UCS estimates for their own " low estimate."
On extremely tenuous grounds.
Rand's high estimate of replacement generating costs was 18 percent higher than the (Rand-corrected)
CAO estimate of such costs, (reflecting the combined eff ects of Rand's longer time horizon and its discounting of future costs).
The Rand authors' " thought it might be more realistic to relace the business costs to the generating costs." Thus they -increased the
$5.0 billion estimate of Stone and Webster by 18 percent to $5.9 billion.
But then, inconsistently, adjusted them downward to almost the level of the UCS estimates.
l l
. By disqualif ying themselves f rom judging the accuracy or reliability of the competing estimates, the Rand researchers have equally disqualified the accuracy and reliability of their own estimates, which are derived from the competing estimates.
Rand presents a range of estLaates of the cost of closure, $7.7 to $17.4 billion, and attributes the large width of this range to "uncer-tainty." In reality, the width reflects the unwillingness of the Rand researchers to pursue an independent assessment of the facts in order to make an informed judgement of probable costs. The vide range of Rand estimates is more accurately attributable to ignorance -- ignorance that is primarily self-imposed.
141sleading Comparisons of Generating Costs A major part of the cost of closing the Indian Point plants would be the cost of replacing the foregone nuclear generation.
Rand terms this the incremental generating cost.
Incremental costs would be incurred throughou t the lif espan the plants would have in the absence of a decision to close them.
The Rand report states that:
To construct an estimate of the total incremental generating costs that would be required by the closure of the Indian Point nuclear units we must sum the individual yearly estimates, and to do that we must discount the costs that would be incurred in years later than 1981 so as to properly represent the present value of the cost stream. (p.21.)
The discounted (present) value of future, annual incremental costs provides a meaningful estimate of the recurring costs associated with closure of Indian Point. Summing over the next 25 years may be, as Rand recommends, an appropriate time horizon, if_ the plants would remain operable for that i
long in the absence of a regulatory decision to close them. As noted previously, this assumption is questionable.
Further, Rand does not consider the possibility of future major expenditures,1f the nuclear plants continue to operate for 25 more years. co meet unknown but probable future regulatory f
upgrades and to provide for major overhauls of the plants.
Rand estimates, thus, are biased upwards.
4 But even accepting the Rand assumptions as valid basis for comparison, they are not applied uniformly to the original CAO and UCS estimates to provide a f air means of comparing them.
_Rev,isions of the original CAO estimates The CAO estimates of replacement generating costs for the years 1982-92, as originally published, are not used by Rand. Rand arbitrarily deflated 4
them by 9 percent per year, removing a major cause of the high estimate of closure costs originally reported by the GAO and a major objection of UCS to the. original CAO estimate. Rand recognizes the Lapropriety of the original CAO methodology when discussing the af f ects of discounting 4
on the calculated costs of closure:
It [the GA01 assumes nearly a 10 percent inflation rate and does not discount. That assumption has deservedly exposed the CAO report to much criticism.
(pp.24 and 25, anphasis added.)
Rand also subtracts the amount of taxes included in the original CAO estimates ($607 million for the 12 year period considered), stating "we agree with the UCS report that taxes paid on increased fuel inven-tories are not, properly speaking, resource costs; they simply represent transf ers from the utilities to the state." (p.13.)
Finally, Rand changes the CAO assumption of a 69 percent future capacity f actor for Indian Point to 60 percent, reducing the GAO estimates of incremental generating costs by an additional 13 percent.*
The Rand adjustments reduce the CAO's estimate of 12 years of closure costs f rom the orig'insi figure of $9.8 oillion to $4.2 billion.
I do not object to Rand's raaking these adjustments to the CAO figures, in f act I applaud tLen but I do consider it misleading and objectionable to represent the adjusted figures as "CAO estimates," as the Rand report does in its Summary, in all of its figures that compare the j
.various estimates, and in much of the discussion of the alternative est ima t es.
But still not reducing it to a level that can be justified on the basis of historical operating experience.
i w
. Use of unequal t ime horizon Rand compounds the misleading nature of its comparisons of total incremental costs by summing the CAO and UCS estimates of annual costs for 12 years but summing its own estimates for 25 years. This creates an enormous upward bias in the comparative Rand estimate, ess entialy doubling it on an undiscounted basis from the value it would have had if calculated identically to the CAO and UCS estimatas.
The Rand researchers do not try to hide this most peculiar and unnecessary divergence from comparability, but only the careful reader will detect it, especially since the Rand report mentions it only in the section on incremental costs and even there treats the 12 and 25 year estimates as though they were strictly comparable. The concluding paragraph in the section on incremental generating costs read:
The Rand estimates, since they are based on 25 years of costs rather than only 12, are significantly larger than the others, especially at the lower rates of net discount. At the 5 percent net discount rate, however, our estimates are only slightly above the original [ sic'] CAO estimate. Ours are
$4.6 billion and $4.8 billion, while the CAO estimate is right at $4.1 billion. (p.25.)
Implic it in this discussion is the assumption that the 12 and 25 year estimates are comparable, for otherwise the discussion would be meaningless--
Ohich of course it is.
Treating the 12 and 25 year estimates as though they are strictly comparable causes Figures 4 and 5 of the Rand report (reproduced here) to be extremely misleading. Exactly how misleading can be judged by the f act that the low Rand estimate and the UCS estimate of incremental generating costs that underlie these figures are identical, except 3dl ear UCS figure has be n carried forward for 13 more years t ha t the 12 y
to produce the Rand estimate.
Rand has no quarrel with the UCS estimate of incremental generating costs, stating "An our lower estimate'of incr emental generating costs, we adopt the UCS estimate. The UCS report contends that the CAO estimates are too high and it presents reasonably well documented alterna-tives."* (p.22.) Further, Rand modified CAO's estimates, based on UCS Rand does make an unwarranted but re'latively minor upward adjustment in the post-1987 UCS estimates, increasing them over our original values by $24 million per year.
Evidently Rand did not understand that the higher, long-run nuclear 0&M estimate we used reflects an empirical trend of rising, real 0&M coser
1 18 16
$ Generating costs
, 14 o8 12 10 5
8 6
4
~
2 O
GAO ERS UCS Rand
-2 NOTE: Estitaates are based on S percent aanvar rate of discovat and at:vme that future iallation rate for these costs is eoval to the general rate (cr the U.S.
economy.
Fig. 4-Estimates of incremental generating costs 14 12 10 8
ci 8
~
e
.95 GAO D G UCS 64 2
CRS 1
l I
I l
I l
g.
-10
-5 0
5 10 15 20 Net discount rate Fig. 5-Sensitivity orgenerating costs to the net discount rate l
1 I
l i
. c r it ic ism s, so that its high estimate is virtually identical to the UCS-based low est ima te.
- Yet although the Rand and UCS estimates are virtually identical and the original, unadjusted CAO estimates were much higher than Rand's es t ima t es, Rand's Figures 4 and 5 convey exactly the opposite impression.
The CAO and Rand estimates are represented as being close together, and the UCS estimates substantially smaller.
Assertion of comparability I have dwelt at length on the misleading character of Rand 's supposedly comparable estimates becasue the Rand report explicitly states that:
Working primarily with their (UCS, CAO, and CRS) cost estimates, but applying a common set of assumptions and converting all data to common dollars, we estimate the total cost implic it in each of the three reports, and we synthesize from those three studies and from other generic material a range of total costs for closing Indian Point that we believe to be reasonably accurate. (p.3.)
The reader is told that the estimates generated by Rand are based on common assumptions and reported in common dollars; they are represented as being strictly comparable. Thu s, what is the reader to think when shortly thereaf ter he or she is presented with Figure 3 (reproduced here) of the Rand report?
18 r
O roi i cosi 14 12
.! 10
=3 8
I'.
6 0
4 2
0 GAO CRS UCS Rand NOTE: Estimaies are based on 5 percent aanval rate of eincovat and assume that future inflation rate for these costs is equas to the general rate for the U.S.
economv.
I Fig. 3-Estimates of total costs of closure See Table 6, p.23.
What else except tha t the UCS estimates are unreasonably low, and that the CAO and Rand are in relatively good agreement?
)
Rand Criticizes the CAO and Agrees with UCS on Many Points Because the summary figures of the Fand report convey a superficial impression of close agreement between the CAO and Rand, I want to anphasize that in many of the details, Rand was critical of the CAO and accepted the UCS estimates. A number of these points were quoted earlier. Consider also the following:
1
(
We question the CAO's estimates of generating costs primarily because they rely on an undocumented costing model, unrealistic proj ections of IP-2 and IP-3's production capacity, an inadequate treatment of operation and maintenance costs, and an abbreviated time horizon. (p.18) 4 The CAO does not explain or question the CE model it uses.
Thus the causes of the large cost for substitute generation 4
j remain obscure.
. We need to obtain and review the full output of the CE model before we can give full credibility 1
to its cutput. (p.18.)
The 69 percent production capacity used in the CE simulations l
and accepted by the CAO appears unjustified. The GAO never attempts to justify this figure; it just adopts it.
We believe the national average of 60 percent wvuld be a much more j
reasonable estimate for Indian Point's capacity over the next 12 years, particularly as the IP units have seldom achieved ev en t ha t level.
- (pp.18-19.)
i The CAO assumes that Indian Point O&M costs would continue, i
apparently through 1992, even though the plant would be dismantled over a six-year period. We agree with the CRS and UCS that
( p. 2 0. )
this is not proper.
We do not count the write-offs (of past investments] as part of closure costs. We f eel they are expenses that have already been incurred. We agree with the UCS that they are sunk costs; they represent items that have already been used, or at least firmly contracted for, and thus will not be affected by the decision on the early closure petition. (pp.31-32) i The historical cumulative average through December 1980 was 57 percent, the figure used by UCS in its analysis. By the end of 1981, i
the cumulative capacity factor had dropped to $2 percent (see prior d iscu ssion).
-16s I
Concerning the one-time savings, we accept the ones suggested l
by the CAO and some portion of the additional saf ety-related costs discussed by the UCS. We believe there is no way to foretell the full total cost of required modifications. UCS places the estimate somewhere between $153 million and $353 million. We use their midpoint of $253 million although we realize that it may turn out to be much higher.* (p.32.)
Close Agreement of Rand and UCS on Incremental Generating Costs and One-Time Costs and Savings When Rand and UCS estimates are placed on a truly comparable basis, those for incremental costs and one-time costs and savings are in close agreement. This can be demonstrated by modif ying the entries in Rand's Table 12 to make the estimates cover the same 25-year time ho r izon. I have done this, while simultaneously making several changes in the UCS estimates to restore them to the values originally reported by UCS.**
In the original report, I did not make an estimate of decom-missioning costs because I was not attempting a present-value calculation, to which they are relevant.
For purposes of the present comparison, I accept the CAO-Rand point estimate, although I want to emphasize that I consider it almost pure speculation. Reliable estimates will not be available until af ter dismantlement of a large reactor has been attempted.
Table A presents both the original and modified estLnates for Rand and UCS. As can be seen, when placed on a truly conparable basis, the Rand and UCS estimates are very close together. This closeness reflects Rand's acceptance of UCS's estimates in almost all cases.
Except that Rand uses the mid-point rather than the range of UCS estimates, Rand's and UCS's estimates of one-time savings from closure are identical. See Tables 9 and 10, pp.30 & 32.
Note that the last two columns of Table 9 are mislabeled with "CRS" rather than "UCS".
These changes consist of 1) reducing the post-1987 incremental generating costs (see Table 3, Rand report) by $24 million per year,
- 2) reducing non-recurring writeoffs on fuel to $20 million, and
- 3) restoring the range of non-recurring savings.($375.$575.million) originally reported l
C i l
Table A Estimates of Selected Components of the Cost of Closing Indian Point (Billions of 1980 dollars)
=-UCS---------
-- Rand
-=
Original Modified (12 years)
(25 years)
(25 years)
Und iscounted Generating Costs 3.9 7.6 8.0-8.2 One-Time Costs 0.0 0.0 0.0 One-Time Savings (0.5)
(0.5)
(0.5)
Total 3.4 7.1 7. 5-7. 7 Discounted at 5%
Generating Costs 2.9 4.4 4. 6-4. 8 One-Time Costs 0.0 0.1 0.1 One-Time Savings
(.4)
(.3.4)
(.3)
To tal 2.5 4.1 -4. 2 4. 4 -4. 6 c.
i r
6
-y,
. Rand's Estimates of Business Costs Have No Foundation Rand defines the " business costs" of closure as those costs chac uould'be incurred in addition to costs of replacement generation and obvious one-time costs and savings. Business costs include "such things as construction programs, financing options, and dividend policies.
is They represent costs that will need to be incurred if Indian Point closed and the utilities are to remain viable.
. " (p. 35.)
Rand's derivation of business costs (from Stone and Webster's
" black Box" exercise) was reviewed in a previous section. The authors made high and low estimates in. spite of their explicit recognition that, "we have little basis for developing a reasonable and defensible estimate of their magnitude at this chuc." (p.47.)
The higher value they choose exceeds their high estimate of replacement generating costs ($10 billion versus $8.2 billion, undiscounted, Table 12, p.47).
Yet, they are able to provide no justification for this value. Indeed, they simply accept the Stone & Webster " black Fox" estimate.*
Rand 's method of estimating business costs f ail to meet normal, i
minimum standards of scholarship. Rand needs to determine to what extent the costs of Con-Ed and PASKY will be increased by the closure of Indian Po int beyond the incremental generation and one-time costs, which are calculated separately.
In the Stone and Webster cases used by Rand to determine " business costs," items are varied, as UCS pointed out in its critique of the CAO report, that have nothing to do with the closure of Indian Point. As compared to the base case, the Indian Point closure case assumes a) higher rates of return on investment for Con-Ed, b) improv ed cash flow (accounting for over $2 billion), c) higher dividends to Con-Ed stockhold ers. These assumptions, which imply closure of Indian Point will bring great benefit to Con-Ed stockholders (at great expense to ratepayers),
account for a major (but unspecified) portion of the cost increase over the is notable that the entire section on business costs is two and It a half pages, whereas 20 pages are devoted to incremental generating i
and one-time costs.
. base case that Rand assigns to business costs. There is simply no valid way of using the Stone and Webster analysis, as presented, to derive a l
l meaningful estimate of business costs as defined by Rand.
1 Hore fundamentally, aside from the absence of any reliable empirical estimate of the magnitude of business costs, there is no basis for arguing that " business costr," as defined by Rand, will be significant. These are supposed to represent costs in addition to all the costs treated elsewhere. But why should there be any significant other costs? Con-Ed and PASNY will continue in business, and they should not require substan-tial extra staf f over a long time to handle af f airs above and beyond close directly connected with shutting down Indian Point and providing replacement power, the costs of which are included in other categories.
There is the suggestion in Rand's definition of business costs that, for some reason, closing Indian Point will raise the costs of non-related construction programs, require Con-Ed to raise its dividends, or require Con-Ed and PASNY to pay higher interest rates to borrow money. But Rand presents no argument to support this vague suggestion.
It would be an argument dif ficult to make. The magnitude of the construction programs needed to replace Indian Point 2 and 3 (1500 megawatts of coal-fired capacity would suf fice) should not strain the combined management capabilities of Con-Ed and PASNY to the point where the ef ficiency of other construction programs suf f ers. Nor would closure of Indian Point create some "special financing cost" not included in the analysis of replacement generating capacity. This analysis already allows for interest, depreciation, and return on equity and capital that are standard for the utility industry. Thus financing costs are already includ ed.
It would be possible to argue that the regulatory decision to close Indian Point could raise investors' perceived risk of utility investment in general; thus raising the return required to attract funds and raining the overall cost of utility construction programs.
A similar argument has been made in the context of rupporting f ederal assistance in cleaning up Three Mile Island, in order to prevent GPU f rom going ba nkru p t.
The argument was that if CPU went bankru pt, all utilities would suf f er f rom higher risk premiums.
i
I
. Such arguments miss the point t ha t the risk premium attached to j
nuclear utilities represents investors' best estimate of the additional financial risk associated with nuclear activities. These risks are real, as THI snd WPPSS* have demonstrated. They are not an artifact of admini-strative or regulatory decisions.
To the extent that these risks are real, they cannot be reduced by regulatory permissiveness or federal su bsid y. Such actions may transfer the burden of risks from one party to another (for example, f rom stockholders to ratepayers or taxpayers),
but they will not af f ect the overall burden to society.
Because utilities do not operate in a competitive marketplace, regulation has the task of attempting to balance costs and risks at their optimum points. A decision to close Indian Point 'will change the balance of costs and risks in the utility sector, but the change will be in the direction of lesser societal risk.
Indian Point is at issue in these hearings because it poses a danger to the population of the greater New York metropolitan area. Closing it and replacing it with coal-fired plants will eliminate this danger. Since Con-Ed bears some (although only a small part) of the financial risks of an Indian Point nuclear catastrophe, the risk premium associated with Con-Ed would decrease if Indian Point were closed There is no theoretical or empirical justification for attributing any significant business costs, as defined by Rand, to a decision to close the Indian Point nuclear f acility. The estimates of this cost category by Rand are without foundation and should be eliminated from WPPSS stands for Washington Public Power Supply System.
It is now commonly pronounced WHOOPS in recognition of the financial debacle created by its ambitious 5-plant nuclear program started a decade ago and now in ruins.
Estimated costs grew from $4 billion in the early 1970's to $24 billion before the impossibility of obtaining the necessary financing forced WPPS to terminate two plants _between 20 and 30 percent complete (which will still incur termination expenses of about $340 million) and to suspend coustruction for 5 -years on another plant 57 percent complete (raising estimated completion costs f rom $2.5 to 4.5 billion). WPPSS now faces the need to get voter approval for future bond issues, since Washington State voters recently passed an initiative to this ef fect.
(Nucleonics Week, January 21, and April 22, and May 6,1982.)
consideration.**
1 Q.
Did Rand incorrectly estimate any other elements of closure cost?
A.
Yes.
Rand erred enormously in estimating what it terms secondary costs. The Rand authors stressed their belief in the importance of these costs:
"We are convinced that secondary costs need to be counted and have estimated that they will range from $2 billion to nearly $7 billion (d iscounted 1980 dollars]." (p.4 0. ) Secondary costs are meant j
i to measure the cost implications of the secondary effects of higher electricity prices on the economy, and they are conceptually very impor tan t.
In attempting to derive a methodology for measuirng these costs, houever, Rand confused income transf ers and fiscal eff ects with add ed cos'es.
Rather than raising total economic costs, consumer responsec to higher electricity prices must necessarily lower them.
Families and businesses will substitute less expensive resources for the higher-priced electricity, reducing electricity consumption and, thereby, the total costs of foregoing Indian Point nuclear generation.
The section of the Rand report on secondary costs is vague and confused, making it difficult to trace the sources of error in the Rand estimates, bu t I will do my best. Several pages are devoted to talking about the possible ef f ects of higher electricity prices on business, household s,
t income, and employment, but a precise definition of secondary costs is never given. The authors stress their interest in the micro-economic detail of the effects of higher prices on types of business and households, asserting the importance of distributional issues f or political decision-making. (p.39.)
But, the aggregate measures of closure costs made by Rand, UCS, and the CAO are not intended to (and cannot) measure distribu-tional effects. Thus Rand's criticism of aggregate economic models (p.39) because they don't capture such eff ects is misplaced. These effects do not need to be captured to accurately estimate secondary costs.
l It is noteworthy that the CAO, which included the Stone and Webster estimates of such costs in its Indian Point analysis, did not include any detectable allowance for added financing costs in its later study of the ec j
located near Chicago, gnomic costs of closing the Zion nuclear plant t
=
l l
l l
. Rand then reviews a study of a nuclear phasecut in California.
This study found that t he pha seou t, which would cause electricity prices to be about 15 percent higher in 1985 and 25 percent higher in 1995, would have almost no eff ect on the growth rate of the California economy between 1975 and 1995 (p.40.)
Evidently, the analysis showed that secondary consumer responses would almost entirely nullify the eff ects of higher electricity prices on the economy. The Rand authors dismiss the relevance of this study to Indian Point on the grounds that "elec-tricity prices in the (Indian Point] service area would increase substan-tially and almost immediately."(p. 40, enphasis in the original.)
Although true that the increase would be immediate, it would amount to only about 10 percent.* To term this increase substantial is stretching the meaning of the word.
Rand continues, "Even more hnportant, the increase would be limited (or at least focused) on a rather small, well defined r eg ion.
This could lead to changes in the location of firms and house-holds, consequences totally ignored by other studies." (p.40, emphasis i
in the original.) Again, it is stretching the usual meaning of a word to term Metropolitan New York small, especially when discussing its economy.
Fu r thermo re, the magnitude of the cost increase would not be large enough to cause many people or firms to consider relocating.
For the residential customer consuming 500 kWh per month (more than the average Con-Ed customer), the increase would amount to about $5 (1980 dollars) per eenth.
For the average New York manufacturer, a 10 percent increase in the cost of electricity would amount to an ir. crease of just 0.2 percent in the total cost of manufacture.**
The estimated benediate increase in costs is about $337 million per year.
Con-Ed's 1980 revenues were estimated to equal $3,867 million (CAO Report, Table 3-12, p.47).
Based on 1976 data for New York State for Value Added by Manufacture of $38.9 billion.(Statistical Abstract of the United States. 1979 Covernment Printing Office, Washington, D.C.) and total electric i
utility revenues from Large Light and Power Sales of $754 million (Statistical Yearbook of the Electric Utility Industry. 1976, Edison Electric Institute, Washington, D.C.,1977).
j Thu s, the Rand reasons for rejecting the California study do not stand up
]
to careful examination.
Wit hou t further discussion, Rand turns to its own method of estimating secundary costs:
"To arrive at our estimates, we reviewed 'over a dozen studies that used various types of economic multipliers to trace secondary costs."
( p.41. ) Here is where the authors seem to have gone astray.
Most of the studies reviewed treated the fiscal multiplier eff ects of an increase in expenditure.
If I spend an extra $10, the person or store that receives it will in turn buy something else with most of it, and the recipient of that payment will also spend most of it, and so forth.
T hu s, the total additional purchasing power created by my initial expen-diture is multiplied severalfold by the chain reaction. This fiscal-stimulus multiplier is the one taught in elementary macro-economic courses.
It does not measure economic costs but the total economic stimulus flowing from an initial expenditure.
When there are unemployed resources in an area, the fiscal multiplier is useful for calculating the stimulus to enployment of resources from a given investment or expenditure program. Such multipliers have no relevance to the measure of indirect costs associated with the closure of Indian Po in t. To u se t hem, as Rand does, to estimate its vaguely defined i
" secondary costs" is an error at once so fundamental and so elementary as to raise serious questions about the competency of the Rand researchers in the field of economics.
The studies reviewed by Rand have multipliers that range from a negative value (unspecified) to 5.5.
The study that produced a negative multiplier, implying consumar-response of f sets greater than the initial cost increase, examined the eff ects of an increase in electricity prices in Buffalo, New York.
Rand dismisses this without evidence, simply stating:
It is unusual to expect such a large reaction, and we view it as unrealistic, at least for a large metropoli-tan area such as the New York City /Westchester County service area of Coned and PASNY.
(p.47.)
The Rand review of multipliers also included a study that seems very relevant to Indian Point, although it treated a far more drastic case. This was a study of the eff ects of abolishing all nuclear power in Sweden by 1990, which would mean closing six operating nuclear plants
6
-2 4 -
that provided 25 percent of the nations electricity in 1980 and foregoing the construction of additional, planned plants that were expected to supply 40 percent of electrical consumption in 1990.9 This study was conducted in connection with the Swedish ref erendum on nuclear power held in 1980. The Rand report notes that this study "uses a multiplier i
of 0.7.
A multiplier that is positive but less than unity indicates that the total (or welfare) costs of a particular policy are actually less than the direct costs alone would suggest."
Rand dismisses the relevance of the Swedish study, stating:
It deals with an equilibrium situation. The markets always clear, unemployment is usually minimal, and prices and wages normally are quite flexible. Capital and labor easily substitute for all forms of energy.
In such a situation, economic theory says that direct costs closely approximate total costs.
( pp.33 -34. )
The ref erence cited for the Swedish study (Ref erence 9) makes no mention of such a theoretical economy, but rather goes into extended detail about the dislocations and disruptions that would be caused by a nuclear prohibition.
Further, the last sentence quoted above is in error.
Consumers will respond to higher prices by reducing electricity consump-t ion, thus bringing total costs below direct costs.
Throughout the section on secondary costs, the Rand authors reject the possibility that consumer responses could off set part of the direct costs. This blind spot is especially surprising because in the " Intro-duction," the authors explicitly recognize that:
If the increased costs (due to closure of Indian Point]
result in higher rates, and those rates reduce the use of electricity.
, the incremental costs will be reduced The reports we reviewed either ignore or confuse these demand ef fects. We suspect that they thus overestimate the incremental generating costs that would result from the closing of Indian Point by some unknown but perhaps significant amount.
(p.7, emphasis added.)
Having recognized the possible importance of secondary of f sets, the authors immediately go on to say that estimating their magnitude will require research beyond the scope of their report; thus they propose to ignore them.
As justificat ion, they present their personal beliefs:
We believe, howev er, that cost changes associated with
demand ef f ect s, while certainly important, are well within the range of uncertainty associated with the other major elements of cost, and thus should not affect our major conclu sion s. "
(p.7,)
But, it is not any great task to estimate the magnitude of the 5
demand responses and their eff ect on closure costs, as I shall show later.
Indeed, as has been noted, several of the studies reviewed by Rand per-formed this analysis for related cases.
In the f ace of all evidence to the contrary, much of which they marshal themselves, the Rand authors in estimating secondary costs rej ect those studies that have multipliers below one and focus on those that estimate (irrelevant) fiscal multipliers.
In their final estimates, the Rand authors use multipliers of 1.5 and 2.5* to yield estimates of secondary costs of $2.3 to $6.9 billion discounted 1980 dollars.
Q.
Could yoti explain how secondary costs should be calculated?
A.
What Rand sas evidently attempting in their section on secondary j
costs was commendable, that is to include economic costs not contained in the estimates of directly measured costs. Such costs (and savings) ought to be included in any comprehensive analysis. Where Rand erred was in carrying out this conceptually desirable step.
An unexpected rise in electricity prices may create temporary economic disruptions and perhaps induce some heavy users of electricity to move to lower-cost areas.
These are true costs, bu t in the case at fund seem likely to be relatively small. As previously noted, the initial 10 percent cost increase would raise manuf acturing costs by only a fraction of a percent. A rise of this amount would not disrupt the economy nor i
induce many f irms to move.**
7 The major effects of a rise in electricity prices are its longer-term ef f ects on consumer behavior.
Users will cut their use of electricity by increasing the efficiency with which they use electricity. (by, for example, installing more ef ficient appliances, air cor.ditioners, and machinery, 1
Which they term " conservative multipliers" (') and justify with the statement, "which we consider aost realistic," (p.45.) bu t with no supporting evidence.
Firms sensitive to the cost of electricity have long since lef t the Con-Ed service area, which has the highest electricity rates of any metropolitan area in the continental United States.
. and better in su la t ion), by shif ting some consumption to alternative fuels, and by simply consuming f ewer electrical services.
The literature on the economic ef fects of higher energy prices is O
now, in the af termath of the oil crisis, very large.
It is widely accepted among economists that over time, the secondary responses of consumers will of f set much of the cost increases that underly a rise in overall energy prices.11 I have shown elsewhere that, under assumptions appropriate to the U.S. economy, secondary responses would off set between 85 and 90 percent of the direct costs of a 10 percent increase in U.S.
primary energy prices.*
For a rise in the cost of electricity alone, the secondary offsets are a greater pcoportion of the initial cost in-crease because there are greater possibilities for substitution than for energy taken as a whole. Analyses of historical data in the U.S.
suggest that the long-run elasticity of demand for electricity is in the 12 range of -0.8 to -1.2, with -1.0 being a reasonable median estimate.
Thic means that a 10 percent increase in the price of electricity will, eventually, cause a 10 percent decrease in electricity consumption. The decrease will come about primarily because of substitution of other, cheaper resources for the higher-priced electricity.
Because the substitution possibilities for electricity (including increasing ef ficiency of use) are large, the final cost to society, after adjustments are made, will be only a small fraction of the initial, direct increase in cost.
The magnitude of indirect monetary saving s In t he attached Append ix, I demonstrate that under assumptions appropriate to New York State, secondary consumer responses would, af ter consumers have had time to adjust fully to higher prices, off set j
96 percent of the direct costs of a 10 percent rise in electricity prices.
This perhaps surprising result reflects the empirical evidence that there are many possibilities for substituting other resources for elec-t r ic it y, if a rise in prices makes such substitutions profitable.
See Append ix C, Table C-1, Ref erence 10.
. i 1
The above result applies only in the long-run, perhaps 10 to 20 yearn af ter the increase, when consumers have fully adjusted their stock of equipment, but a substantial part of the adjustment will take place quite quickly. A comprehensive econometric analysis of U.S. data found that 20 percent of the desired adjustment would take place in each year.13 This bnplies that 50 percent of the total will occur within 3 years, 67 percent within 5 years, and 90 percent within 10 years.
My own research supports the conclusion that demand responds quite quickly to changes in price.
I performed a detailed analysis of changes in electricity prices and sales of the Public Service Company of New Ilampshire (which provides over 90 percent of the electricity consumed in New Hampshire).
The analysis was based on data covering the period 1960-1981. Demand responses within each of 4 customer classes were analyzed separately.
Electricity prices were first adjusted for inflation; then changes tu yearly electricity sales were related to the prior 3-year average of adjusted price changes. The elasticity of demand with respect to (be 3-year-average price change was estimated to range f rom
.5 (for Industrial - Commercial and Service customers) to -1.0 (for Residential and General Service customers). For the company as a whole, the elasticity of demand with respect to the 3-year average price was
.82.
This means that over the three years following a 10 percent increase in price, demand will decrease by 8 percent, other things remain-ing the same.
As longer-term ef f ects of price changes were not considered in the New ilampshire analysis, the elasticity estimated there (of
.82 over 3 yea r s) is consistent with a long-run elasticity of
-1.0.
The adjustment rate implied by the New Hampshire analysis is more rapid than the 20 percent rate found by the authors of Ref erence 13.
The Cost of closing Indian Point I have calculated the annual incremental costs of closing Indian Point, including secondary of f sets.
The calculations are based on the i
i 1
-27A-model presented in the Appendix, with a long-run elasticity of demand o f -1. 0, a ssum ing t ha t 20 percent of the adjustment toward the desired long-run demand-level occurs each year.
I used the UCS values for annual incremental generating costs of $337 for 1981-87 and $292 million there-after.
The results are presented in Table B, which also shows the dis-counted values of the costs of closure. The total incremental costs of closure f or 25 years are $1.2 billion (discounted 1980 dollars).*
Including net one-time savings (f rom Table A), the discounted net cost of closing Indian Point is $.8 to $.9 billion.
The Benefit of Closing Indian Point In deciding whether er not to close Indian Point, the net economic costs of closure of approximately $1 billion need to be compared with the benefits f rom eliminating the risk of nuclear accidents at this facility.
If there is a severe accident, millions of people and many billions of dollars of property will be at risk.
There is no agreed upon method for placing a monetary value on human lif e, nor can the probability of a nuclear accident at Ind'ian Point Because most of the costs occur in the first years af ter shutdown, the dif f erence between discounted and undiscounted costs is only
$0.4 billion.
i I
-28 Table B Incremental Costs of Closing Indian l
Point Allowing for Secondary Responses (Millions of 1980 dollars)
Incremental Costs 2
Years Und iscounted Discounted 1981 273 259 1982 221 200 1983 179 155 1984 14 6 120 1985 119 94 1986 98 73 4
1987 81 57 1988 59 39 1989 49 32 1990 42 26 i
1991-95 14 0 74 1996-2000 86 36 2000-05 67 22 Totals 1560 1187 1.
Based on replacement generation costs of 337 mfilion during 1981-87 and $292 million thereaf ter.
2.
Discounted to 1980 at 5 percent per year.
I l
[
I l
l
-29 be estimated with any precision; thus no neat calculation is possible of
?
I the dollar benefits of eliminating accidents by closing Indian Point.
j Still, it is important to stress that closure would produce not only economic costs but economic benefits.
3 To illustrate the principles involved, but without attaching signi-ficance to the particular numbers used, consider the following. Suppose I
that each avoided death were porth $500,000 dollars and each avoided serious injury worth $50,000. Suppose further that the probability of a l
major accident producing 20,000 expected deaths, 50,000 serious injuries, and $40 billion in property damage and clean-up costs was one per thousand per year. Then, over the next 25 years, the probability of such an accident would be 25/1000 or 2.5 percent. The benefit of avoiding such an accident would be 20,000 x $500,000 + 50,000 x $50,000 + 40 billion =
52.5 billion. The expected value of the avoided accident would be 1
.025 x $52.5 billion = $1.3 billion. This accident would occur on average 1
12.5 years in the future. Discounting by 5 percent over 12.5 years, the approximate present value of the benefit is $.7 billion.
The $.7 billion benefit figure is slightly below the calculated cost i
of closure of $.8 to $.9 billion.
Rather modest (relative to the uncertain-ties involved) changes in assumptions would produce a benefit figure greater than the net cost. As all of the benefit assumptions were made nere1** for illustration, no special significance should be attached to the close. 'ss of tue resulc3 To t he ex t ent, houev er, tha t the illustrative figures can be considered to be within the range of reasonableness, the result indicates that the economic dimension of the Indian Point issue is not clear cut.
Whether one judges that the benefits of closure excced the costs depends upon the assessment probabilities and consequences of a major nuclear accident at Indian Point, an assessment impossible to conduct with any precision or confidence.
i f
a b
..-m.
.,4
. _ Appendix Equilibrium Economic Costs of an Increase in Electricity Costs Consider the following simple model:
8
~*
(1) Y = AE X
,where Y = Total output (CNP)
E = Resources used to produce electricity.
X = Resources used to produce other output.
This model is appropriate for considering changes in electricity price because it implies an clasticity of substitution of 1.0, which is approximately the historical value in the U.S.*
The share of electricity in total production in New York State (in 1979) wa s a bou t.036.
Given the properties of the simple production function defined in (1), this implies (2) a=.036 Withou t loss of generality, let Y = 1.0 and the prices of E and X(PE ""
X}
also equal 1.0, initially. Then, the value of inputs equals the value of outputs:
(3)
E, + X, = Y, = 1. 0 Also, a condition of equilibrium is that:
(4) PE=
a
.X; P
l-a E
X thus for P X"
~
E (5)
, g _,,
E a
and using (3), this can be solved to yield X, =.964 and E, =.036, and f rom (1), A = 1.1677.
1 Asst me the price of E increases by 10 percent, that is P
" l*l*
E 1
1 In the new eqt'ilibrium, P /P 1.1, and
=
E X
The elasticity of substitution is essentially equal to the absolute value of the elastic ity of demand. See Reference 12 for elasticity est ima t e s.
i
)
1 i
-31 1
b = 1.1.
(
1, I
Using (3), this yields X =.9672.and E =.0328.
1 y
The new value of Y, (l') Y, = A(E )" (X ) ~* =.99985.
y 7
The loss due to the 10 percent increase in electricity costs is
.015 percent of total output. The assumed initial increase in electricity costs was equal to.36 percent of total output. The final economic cost, af ter adjustments are made, is thus, only.015/.36, or 4 percent of the initial cost increase. The indirect savings are 96 percent of the initial cost increase.
4
1 o
t
" 5 2 '
Reference.
i l,
1.
Economic Impact of Closing the Indian Point Nuclear Facility, Report 2
by the Comptroller General of the United States, U.S. General Accounting
}
Of fice, November 7,1980.
Indian Point 1 is no longer licensed to operate and is not at issu e.
2.
Vince Taylor and Charles Komanof f, An Evaluation of " Economic Impact of Closing the Indian Point Nuclear Facility." A Report of the General i
Accounting Of f ice, The Union of Concerned Scientists, Cambridge, Mass.,
December 3,1980.
i 3.
James Stucker, Charles Batten, Kenneth Solomon, Werner Hirsch, Costs of Closing the Indian Point Nuclear Power Plant, Rand Corporation, i
Santa Monica, Calif orn ia, R-2857-NYO, November 1981.
4.
Carolyn Kay Brancato, "The Indian Point No. 2 Nuclear Facility,"
i Congressional Research Service, December 5,1980.
5.
Made in a series of "
Dear Colleague" letters from Allen Ertel (Dem.,
Pa.)
in support of his bill (H.R. 2512) to provide increased insurance for l
nuclear utilities and for assistance in cleaning up TMI. Nuclear News, Oc tober 1981, p.106.
6.
Economic Impact of Closing the Zion Nuclear Facility, Report of the i
Comptroller General, General Accounting Of f ice, Washington, D.C.,
EMD-82.3, Oc tober 21, 1981.
7.
Martin L. Baughman et al., Direct and Indirect Economic, Soc ial. and j
Environmental Impacts of the Passage of the Calif ornia Nuclear Power j
Plants Initiative, Center for Energy Studies, University of Texas at Austin, FEA/G-76/261, April 1976 6.
J. H. Savitt, Electr ic Energy Usage and Regional Economic Development, Final Report, EPRI, Palo Alto.. California, ES-187, August 1976.
I f
9.
Konsekvensutredningen, Suppose We Go Non-Nuclear....?
Effects on the Ec onomy. Em plo ym en t. and the Environment in Sweden, Departementens Of f setcentral, Stockholm, 1980.
i 10.
Vince Taylor, Energy: The Easy Path, Unica of Concerned Scientists, Cambridge, Mass., January 1,1979, explains in detail the potential for such responses and provides ref erences to other literature.
i 11.
William Hogan and Alan Hanne, " Energy - Economy Interactions: The
~
Fable of the Elephant and the Rabbit," in Modeling Energy - Economy Int erac t ions:
Five Approaches, Charles Hitch (Ed.), Resources for the Fu ture, Washington, D.C., R-5,1977, shows the basic economic reasoning behind this conclusion.
j i I a
i 12.
Lester Taylor "The Demand for Energy: A Survey of Price and Income Elasticities," in International Studies of the Demand for Energy, William Nordhaus (Ed.), North-Holland Publishing Co., New York,1977.
13.
Tim Mount and Tim Tyrell, Energy Demand: Conservation, Taxation, and Growth, Appendix to Chapter B, Report of the Panel on Energy Demand and Conservation of the Committee on Nuclear Energy and Alternative Energy Systems (CONAES) National Research Council, National Academy of Science, Washington, D.C.,
published as Cornell Agricultural Staff Paper No. 77-33, August 1977. See Table 2, p.18.
14.
Testimony of Vince Taylor, Investigation Into The Supply And Demand For Electricity, Public Service Compar,y of New Hampshire, New Hampshire Public Utilities Commission, Docket DE 81-312, October 8,1982 1
1
e
,_si AFFIDAVIT State of Vermont County of Franklin On this 42 day of April, 1983, before me personally appeared Vince Taylor to me known and known to me to be the individual who executed the foregoing testimony and he thereupon duly acknowledged to me that he executed the same.
o zb t',&pt ]Il l'
? bdA'c dm W
-.uL
~~
, (g ut Lu w flM'}
[
Ss I" ' U I j m.
, af n w z u.,
3 EXHIBIT I VINCENT D. TAYLOR PROFESSIONAL QUALIFICATIONS EDUCATION:
Bachelor of Science in Physics, California Institute of Technology,1958.
Doctor of Philosophy in Economics, Massachusetts Institute of Technology, 1964.
PROFESSIONAL EXPERIENCE:
Economics Department, Rand Corporation, Santa Monica, California,
1961-1969; consultant to Capital Research, Inc., Los Angeles, on security selection, 1970-1973; senior staff member of Pan Heuristics, Los Angeles (a division of Science Application, Inc., La Jolla, California),
1974-1978; energy consultant to the Union of Concerned Scientists,1979; senior staff of the Union of Concerned Scientists, 1980-82; economic consultant, 1982 to present.
PROFESSIONAL EXPERIENCE IN ENERGY-RELATED RESEARCH:
Beginning in 1974, my prof essional work has been exclusively on the economics of energy, with particular emphasis on the comparative economics of nuclear power and its alternatives. During the period 1974-1982, I performed research and wrote reports and articles on:
the comparative economics of nuclear and coal generated electricity, forecasts of the future demand for energy, electricity, and nuclear electricity, the comparative economics of the use of uranium and plutonium on fuel for nuclear reactors, the economics of the uranium market, the economics of reprocessing of spent nuclear fuel, the comparative future energy potentials of nuclear power, synt'hetic fuels and improvements in the ef ficiency of energy use, the economic eff ects of the oil crisis, the contribution of electric utilities to oil consumption, and the economic implications of closing a nuclear power plant.
ENERGY-RELATED CONSULTING:
During the period 1974 -1983, I have provided consulting services, research reports, or expert testimo.y for:
- The United States Arms Control and Disarmament Agency
- The Energy Research and Development Administration
- The California Energy Commission The Council on Env.ronmental Quality
- The Nuclear Regulatory Commission
- The Pennsylvania Public Utility Commission
- The Vermont Public Service Board
- The Of fice of Technology Assessment
- The New Hampshire Public Utilities Commission
3 o
ENERGY-RELATED PUBLICATIONS:
The Uncertain Future of Nuclear Power, with Dennis Holliday, California Seminar on Arms Control and Foreign Policy, P.O. Box 925, Santa Monica, California, August 1975 Is Plutonium Really Necessary?
Pan Heuristics, Los Angeles, Sept., 1976 (Revised)
The Myth of Uranium Scarcity, Pan Heuristics, Los Angeles, April 25, 1977 How the U.S. Government Created the Uranium Crisis, Pan Heuristics, Los Angeles, June 1977 (Revised)
The Economics of Uranium and Plutonium, in " Moving Toward Life in a Nuclear Armed Crowd?", Minerva, Volume XV, Numbers 3 and 4 (combined issue), Autumn-Winter, 1977 Prepared Testimony of Dr. Vince Taylor in the Matter of GESMO, prepared for the California Energy Resources and Development Commission, Pan Heuristics, Los Angeles, March 4, 1977, Chapter A.
Energy:
The Easy Path, prepared for the U.S. Arms Control and Disarmament Agency, January, 1979 (published by the Union of Concerned Scientists, Cambridge, Mass.)
The Easy Path Energy Plan, Union of Concerned Scientists, Cambridge, Massachusetts, May, 1979
" Science and Subjectivity," Technology Review, February, 1979.
"A Warning:
E. F.
Schumacher on the Energy Crisis," MANAS September 3, 1980.
"The End of the Oil Age," The Ecologist, October-November, 1980.
Swords from Plowshares, with Albert Wohlstetter, et. al.,
University of Chicago Press, Chicago and London, 1979.
Conservation, Equity, and Efficiency, testimony before the Vermont Public Service Board, Docket 4475, November 5, 1980 Electric Utilities:
The Transition from-Oil, testimony before the Subcommittee on Oversight and Investigations of the Commit-tee on Interstate and Foreign Commerce Committee of the United States House of Representatives, December 9, 1980.
" Electric Utilities:
A Time of Transition," Environment, Volume 23, No.
4, May 1981.
Testimony on the Economic Costs of Closing Indian Point, testimony before the Atomic Safety and Licensing Board of the Nuclear Regulatory Commission, Docket Nos. 50-247-SP and 50-286-SP (pending).
a s
.p ENERGY-RELATED PUBLICATIONS (CONTINUED) :
Testimony of Vince Taylor, Investigation Into The Supply And Demand For Electricity, Publ,ic Service Company of New Hampshire, New Hampshire Public Utilities Commission, Docket DE 81-312, October 8, 1982.
l "Living Without Nuclear Energy," Nuclear Power:
Both sides, M.
Kaku and J. Trainer (Editors),
W.
W. Norton, New York, 1982.
l t
l l
l
~
t.
Exhibit II Ill0ll 0 Comcensso SCisNTISTS AN EVALUATION OF " ECONOMIC IMPACT OF CLOSING THE INDIAN POINT NUCLEAR FACILITY,"
A REPORT OF THE GENERAL ACCOUNTING OFFICE Vince Taylor Charles Komanoff December 3, 1980 Union of Concerned Scientists 1384 Massachusetts Avenue Cambridge, Mass., 02238 (617) 547-5552 or 1725 I Street, N.W.
Washington, D.C.,
20006 (202) 296-5600
i l
l AN EVALUATION OF " ECONOMIC IMPACT OF j
CLOSING THE INDIAN POINT NUCLEAR FACILITY,"
A REPORT OF THE GENE,RAL ACCOUNTING OFFICE The GAO report does not provide an accurate assessment of the costs of closing Indian Point 2'and 3 nuclear plants.
Due to a number of errors, it exaggerates both the short and long run costs of a shutdown.
SHORT-RUN COSTS The GAO estimates that first year costs of a shutdown would exceed $600 million.
This estimate assumes unrealistically l
high operating rates for the nuclear plants (a capacity factor l
of 69 percent versus a historical average of 57 percent).* It also assumes that nuclear fuel costs and operating and maintenance costs are the same whether or not the Indian Point i
plants are closed--an obvious error.
Correcting for these factors, the near-term, net costs of closing Indian Point appear likely to be about S337 million per year (see Table 1, attached).
MAJOR EXPENDITURES SAVED BY CLOSURE The GAO estimates that closing Indian Point would save future expenditures on major plant repairs and safety improve-ments of S220 million.
This estimate, however, includes only a portion of likely future safety related expenses and does not even reflect all of the items costed in their report.
In particular, costs of responding to NRC directives issued in response to TMI appear substantially underestimated, and no allowance is made for safety improvements now being considered by the NRC because of Indian Point's proximity to New York City.
A more realistic assessment of future, non-recurring expenses that might be avoided by closing Indian Point is S374-575+
million, with a significant chance that the upper end of this range could be exceeded (see Table 3, attached).
ERRONEOUS GAO " COSTS" OF CLOSURE Rather than the figure of S431-million.given by the GAO
(" Digest," page v) as the cost of closing Indian Point, the actual added cost of early closure seems likely to be only a few tens-of millions of dollars.
The GAO cites a figure of S233 million as the cost of decommissioning the plant and moving and temporarily storing l
the spent fuel off-site.** Costs of these activities are i
erroneously counted as " costs of closing the units."
- See addendum
- Page v of the GAO repor t " Digest" mistakenly says that the cost is to " dispose'of the waste fuel," but the text refers
.to moving it to temporary storage.
[.
. l Decommissioning costs will be incurred whether the plants are closed now or later.
If past trends toward rapid escalation of costs of activities involving radioactive materials continue, i
early closing would reduce rather than add to decommissioning costs.
The cost of moving the spent fuel could be avoided by keeping the fuel on-site until permanent storage becomes l
available.
The GAO report also counts a loss of $198 million on fuel as a cost of closing.
This is also an error.
About one-half of the S198 million represents fuel already in the l
core, an equivalent amount of which must remain in the core throughout the lifetime of the plant.
It represents part of the investment cost of the plant which has already been made and should not be counted as a new cost of closing the plant.
l The remainder of the S198 million is attributed to losses on fuel being fabricated; but since this fuel could be used in other Westinghouse reactors (perhaps with some modification) actual losses after deducting salvage value should be relatively small.
LONG-RUN COSTS OF REPLACING INDIAN POINT GENERATION Erroneous GAO Estimates The GAO estimate that closing Indian Point will cost Con-Ed ratepayers $18 billion over the next 15 years,
(" Digest,"
p.
i) is a gross exaggeration.
The figure was generated by a model that included as " costs of closing Indian Point" such items as a) higher rates of return on investment for Con-Ed, b) improved cash flow (accounting for over $2 billion of the S18 billion), and c) higher dividends to Con-Ed stockholders.
These factors account for a major (but unspeci-fied) portion of the total.
Further, high rates of inflation were assumed, and costs were reported in inflated dollars.
The GAO report estimates that closing Indian Point will cost the Power Authority of the State of New York (PASNY),
which owns Indian Point 3, "as much as $600 million annually"
(" Digest," page i).
Not mentioned in the " Digest" is that this was the upper limit of a series of estimates provided by PASNY to the GAO.
The lower limit was $23 million annually (after an initial first-year cost of $226 million; GAO Report, page 52), and the GAO provides no basis for. suggesting where in this range of $23 million to $600 million per year the probable, correct answer might lie.
Since the total short-run cost (which is higher than the long-run cost) to both Con-Ed and PASNY is about S325 million per year (Table
- 1), it is obvious that the cited figure of $600 million per year for PASNY alone is well beyond the bounds of possibility.
l i
l l
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ - l The True Upper Limit of Long-Run Costs The true upper limit of long run costs of closing Indian Point is provided by the costs of constructing and operating equivalent new coal-fired capacity less the savings in fuel, operating and maintenance, and future investment costs from closing Indian Point.
An estimate of this upper-limit cost, based on a comprehensive analysis of trends in the cost of constructing new coal-fired facilities, is $250 million per year (in 1980 dollars; Table 2, attached).
If Indian Point were to be closed, the added costs of replacement power would, initially, approximate $337 million per year.
The costs of both fuel and replacement capacity would eventually fall to $250 million per year if new coal capacity were to be built to replace the capacity.
l Coal-Conversion versus New Capacity The GAO report does not consider the possibility of converting existing oil-fired capacity to coal to replace Indian Point.
This may offer a quicker, cheaper alternative to building new coal-fired capacity.
Costs of meeting stringent air pollution standards would not be a barrier to such conversions.
At the current cost differential between oil and coal, fuel savings would pay for a technically advanced pollution control system, including high-efficiency precipi-tators and a sulfur-dioxide scrubber, in two and one-half years.
The Jamesport Coal Plant--An Early Possibility The Long Island Lighting Company (LILCO) has received permission to build an 800 MW coal-fired plant at Jamesport.
Reportedly, LILCO will not build this plant if it decides to go ahead with federally ordered coal conversions of existing oil-fired plants.
If this is the case, the Jamesport plant could be built specifically to replace (about one-half of)
Indian Point generating capacity.
It could be available between 1986 and 1988, depending on the length of time required to obtain permits from the Environmental Protection Agency and, possibly, local governments.
Potentials for Conservation The GAO report assumes that demand for electricity in the Con-Ed service area will increase by 7.8 percent between 1981 and 1988 (Table 3-5, p. 36).
Between 1973 and 1979, demand in this area declined by 4 percent.
Recent sharp increases in rates together with further increases that can be expected would cut demand further, especially if an effective, utility financed conservation program were instituted simultaneously.
. A 10 percent reduction in electricity consumption would offset 40 percent of the loss in generation caused by closing Indian Point.
t GAO EXAGGERATED EFFECTS ON PASNY CUSTOMERS The GAO estimates that closing Indian Point would raise.
PASNY's average cost per kilowatt by 4.19 cents in 1981 to 6.24 cents in 1992
(" Digest," p.
iv).
The repart states (page 45) that PASNY rate increases in 1981 would range from 45 to 95 percent of 1980 rates.
These estimates are based, first, on inflated estimates of replacement power costs and, i
second, on the assumption that all costs of closing Indian Point 2 would be loaded onto customers in the Con-Ed franchise area, which accounts for only 40 percent of PASNY kilowatt sales.
If added costs are shared equally by all PASNY customers,
[
the short-run estimate of costs given in Table 1 implies added costs for PASNY of 0.85 cents per KWH.
This would represent y
an average increase of 14 percent for the 5 largest PASNY I
customers (Table 2-5, page 24).
t i
t t
?
k r
f I
i
! ~i 1
1 I
Table 1 t
Short-Run Net Costs of Closing the Indian Point Nuclear Plants (1980 dollars) 1.
Plant Capacity 1838 MW 2.
Capacity factor" 57%
3.
Annual generation 9.168 billion KWH Oil required for replacement generation Low High Total Sulfur Sulfur b
4.
Millions of barrels 11.36 4.64 16.0 5.
Price / barrel
$31.15
$26.77 6.
Fuel Cost (S million) 353.9 124.2
$478.1 million Savings from closing Indian Point 7.
Fuel savings 0$.008/KWH"
$73.3 millicn 8.
O and M savings'
$67.9 million 9.
Total savings
$141.2 million 10.
Net Costs:
(6) minus (9)
$337 million a.
Lifetime cumulative factor through June 1980.
b.
From GAO report Table 3-7, page 38, with quantities shown. for 3 981 reduced by 17.4% to reflect a capacity factor of.57 percent for :
Indian Point rather than the 69 percent used by the CAO.
Average U.S. retail prices for first 6 months of 1980, Monthly Energy c.
Review, Department of Energy, DOE /EIA 0035/80 (10). Price for high sulfur oil is that given for.3 to 1.0 percent sulfur. Actual consumption would include higher sulfur oil, which was $6.50 per barrel less than the cost used.
d.
Fuel costs are the bulk of additional cost associated with oil generation of replacement power, 'since the generation will be from existing, operating facilities. Operating and maintenance for these plants will not vary greatly whether or not extra generation l
is demanded of them. The GA0 report includes taxes as a. cost, but these are not true economic costs but a transf er to the state.
l l
l-r l
. I (Notes to Table 1 Continued) i e.
The CAO report (p.56) gives a Con-Ed estimate of, $46.7 million for i
a fuel reload for Indian Point 2.
Based on a 60 percent capacity factor for 18 months, the cost per KWH is $.0068. Adding $.0012/KWH for interest and waste disposal costs gives $.0080/KWH.
i.
f.
Operating and maintenance (0 & M) costs for 1979 were $59 million (GAO report, page 57). Applying a factor of 1.15 to account for inflation yield $67.9 million.
i i
l 4
1 1
1 I
I, t
4 l
l 4
f i
.-.,.__,.J.-.-__,,,_,.
.. j,..___ ~ _. - - -, _,. ~,. _.. _.. _,. _.. _,.
.. - = -
Table 2 i
i Net Long-Run Costs of Replacing Indian Point with New Coal-Fired Capacity' 3
(1979 dollars)
[
Indian Point Data i
i 1.
Capacity 1838 MW j
2.
Capacity factor 57%
3.
Annual Generation 9.168 billion KWH j
Savings from closing Indian Point b
l 4.
Non-Recurring
$ 375 million Recurring' 0&M S.0071/KWH Fuel
.0109/KWH 5.
Total recurring
.0180/KWH l
Coal Plant Data i
f 6.
Annual generation 9.168 billion KWH d
7.
Capacity factor 70 percent 8.
Capacity required 1500 MW i
9.
Capital cost per KW
$936 10.
Total capital cost
$1404 million e
Recurring O & }f S.0071/KWH Fuel Cost
.0241/KWH 11.
Total recurring S.0312/KWH i
i i
Net cost of replacement power I
i l
Coal capital cost (9)
$1404 million l
Non-Recurring savings (4)
-375 million j
12.
Net capital cost
$1029 million 13.
Real fixed charge rate
- 10.0% per year 14.
Net capital cost per year
$102.9 million 15.
Net capital cost per KWH 0114/KWH l
[
v.
7-y y,.
.,,-.mn
-,--.-y.
,,,.,,..e,,
,e
. Recurring coal costs (11)
S. 0312 /KWH Less recurring nuclear costs (5)
.0180/KWH 16.
Net cost per KWH S.0246/KWH 17.
Total annual net cost (3) x (16): $225 million/ year 18.
Cost in 1980 dollars (11% inflation): S250 million/ year a.
All data on coal and future nuclear fuel costs are from Power Plant Escalation by Charles Komanof f (f orthcoming 1981).
Coal plant is assumed to be in operation in 1988.
b.
See Table 3--Minimum estimates used here. Actual savings could exceed $575 million.
c.
Estimate by Charles Komanoff for levelized 1988-2017 costs, expressed in 1979 dollars.
d.
Appropriate for plants of approximately 300 MW, the preferred size of unit, c.
Fixed charge in a non-inflationary environment, corresponding to the use of constant dollars in the analysis.
Current dollar fixed charge rate equals the real rate plus the expected long-term rate of inflation.
. i Table 3 Non-Recurring Cost Savings from Closing Indian Point (millions of dollars) 1.
Repairs to steam generators"
$145 2.
Budgeted saf ety improvements" 62 3.
Utility estimates of future budget additions for saf ety improvements
- 23 4.
Emergency response planning" Utility costs 8.8 State costs 4.5 County costs 1.5 r
5.
Study costs for IP cooling towers"'
10.0 Subtotal 254.8 6.
Additional costs of meeting post-TMI C
safety changes 120+
7.
Costs of improving saf ety of reactors d
located near large population centers 0--200+
Subtotal 375--575+
4 8.
Cost of meeting future NRC requirements to resolve currently Possibly unresolved saf ety problems and future very problems detected by operating experience
- large a.
From CAO report b.
Costs are cited at S3.5-4 million per year for Con-Ed (page 57);
no duration is given in the report. The study is assumed here to last for about 3 years.
Estimated costs included in the CAO report for Laplementing NRC c.
i safety improvements emerging f rom the post-Three Mile Island review are $26.9 million, plus some portion (unspecified) of the unbudgeted but estimated $23 million. The management of Rancho Seco recently _
l raised their initial estimate (of March 1980) of $39.2-53.5 million i
to $85.7 million, primarily due to expansion of NRC requirements (Nucleonics Week, November 20, 1980). The review process is by no means completed and, thus, further increases can "be expected. Even i
aside from these further increases, the Indian Point facility, which consists of two units of. approximately the same size as Rancho Seco, can be expected to incur costo twice as great as Rancho Seco, or
$170 million, $120 million more than included in the CAO report.
. (Notes to Table 3 Continued) d.
The NRC is currently studying the issue of whether additional safety modifications should be required for plants located near large population centers, e.g. Ind ian Foint. Design changes being considered are vented filter'ed containment pressure relief system, core retention devices, and hydrogen control (GAO report, page 17).
No estimates were made in the GAO report of the possible range of costs. Very preliminary estimates were provided by Jim Myer, NRC project manager for these measures by telephone (December 2,1980):
The filte' red containment system could range from $15-50 million; the hydrogen control system could range from small to $50 million; and the core rentention devices are not currently costed; thus the range is from zero (no requirements) to upwards of $100 million per reactor, or upwards of $200 million for the two Indian Point reactors.
e.
The NRC currently recognizes 14 " unresolved saf ety issues,"
some with potentially serious consequences, and the NF.C staf f has proposed adding 7 more to the list. Additionally future operating experience (such as that provided by Three Mile Island) can be expected to reveal new problems that will need correction.
No firm estimates are possible, but the upper limits are obviously very high and should not be completely ignored.
- ADDENDUM l
l Indian Point Capacity Factor l
4
(
There is no basis for the 69% capacity factor assumed for the Indian Point units and the GAO report ( page 34 ).
To date, Indian Point II and III have operated at lifetime average capacity factors of only 55% and 57%, respectively.
Moreover, their year-by year records fail to demonstrate the " operational maturity" that the report assumes.
Unit II performance has improved somewhat over time, but Unit III performance has deteriorated, as the following table indicates.
INDIAN POINT CAPACITY FACTORS Year Unit II (873-MW)
Unit III (965-MW) 1974 43%
1975 64%
1976 30%
1977 68%
65%
1978 57%
65%
1979 63%
57%
1980 56%
39%
Average 55%
57%
1 I
First 11 months.
l Source:
Net generation data from U.S.
NRC " Gray Books."
l Note:
Table excludes initial Unit II operation in late 1973 and l
initial Unit III operation in late 1976, in which the units performed poorly and well, respectively.
. l-Indian Point's poor generating performance is consistent with that-of other large nuclear units.
The 39 U.S.
reactors of 800 MW capacity or greater registered a lifetime average capacity factor of only 54% through mid-1980.
(See'Komanoff, "U.S. Nuclear Plant Performance," Bulletin of Atomic Scientists, November 1980.)
Contrary to nuclear industry assertions and government misp'erception, there has been little or no improvement in the performance of these large units over time.
Any putative future improvement in Indian Point's capacity f actors is likely to be' offset by prospective more stringent regulations of nuclear plant operations.
i r
i I
e o
STATE OF NEW HAMPSHIRE PUBLIC UTILITIES COMMISSION
)
Investigation Into The Supply And )
Demand For Electricity, Public
)
DE 81-312 Service Company of New Hampshire )
)
Testimony of Vince Taylor On Behalf Of THE CONSERVATION LAW FOUNDAIION OF NEW ENGLAND, INC.,
THE NEW HAMPSHIPI ENERGY COALITION, AND THE UNION OF CONCERNED SCIENTISTS October 8, 1982
~ d
....s-..*.
.. ~.
,, r.
,1..-
'1 AFFIDAVIT State of Vermont County of Franklin On this 28th day of September 1982, before me personally appeared Vince Taylor to me known and known to me to be the individual who executed the foregoing testimony and he thereupon duly acknowledged to me that he executed the same.
N.
NY
(
/
(Vlh,0 $ YW]N o
t aw, R
'/ M73
%m AJ/xw k
r t
~
\\
1 TABLE OF CONTENTS 1
TABLE OF CONTENTS..................................................
1 QUALIFICATIONS.....................................................
1 T h e P r ic e o f U ra n ium.......................................... 1 Gro wth o f Nuclea r Po wer....................................... 2 The Economics of Reprocessing Nuclear Fu el.................... 3 So lv ing t he Energ y Cr is is..................................... 4 Co n c lu s io n.................................................... 5 I
OBJECTIVES.........................................................
6 i
CONCLUSIONS ON SUPPLY, DEMAND AND C0ST............................. 7 FINDINGS ON SUPPLY, DEMAND AND C0ST................................ 9 The Importance of Seabrook....................................
9 The Influ enc e o f Pr ic e on D emand.............................. 9 Ex pla in ing D emand Chang es.................................... 10 Future Demand and Cost.......................................
11 Ad equac y o f Plann ed Genera tion............................... 12 Canc ella t ion o f. S eabroo k 1I.................................. 13 Unc ertainties Other Than S eabrook............................ 14 EXPLANATORY COMMENT ON THE SENSITIVITY OF DEMAND TO, SEABROOK COSTS.15 ECONOMIC AN ALYSIS OF ELECTRICITY DEMAND........................... 17 Methodology.................................................. 17 Historical Analysis of Electricity Demand.................... 20 R e s id en t ia l............................................. 2 0 General Serv 1ce......................................... 22 Indu s tr ial -- Manuf ac tu r ing............................. 2 9 Industrial -- Commercial and Service.................... 32 PROJECTING FUTURE DEMAND FOR ELECTRICITY.......................... 3 6 Und e rly in g A s sum p t io n s....................................... 3 6 The Ba s ic Econom ic Env iro nm ent.......................... 3 7 Proj ecting Future Elec tricity Co sts.......................... 37 Al t erna t iv e S eabroo k Ca s e s................................... 3 7 Proj ec t ed Sale s and Pr ic es................................... 3 9 Appendix A
)
PROJECTING ELECTRIC LOAD GROWTH FOR PSNH: ASSUMPTIONS AND METHODOLOGY....................................................A-1 Assumptions..................................................A-1 Proj ect ed Basic Economic Environment.................... A-1 Prices of Fuel and Interchanged Power................... A-2 Gen era t in g Ca pa b i11 t y................................... A-2
t
)
-11 S ea br oo k P erfo rmanc e................................... A-3 Lagg ed Respanse to Price Shif ts........................ A-3 Met hodology -- An Illu stra tive Example...................... A-3 Trial Valu es o f Requ ired Generation.................... A-3 Proj ect ed Production and Purchase Costs................ A-4 Non-Produ c tion Co s t s................................... A-4 Costs per KWH..........................................A-4 I
Price Changes by Cu stomer Cla ss........................ A-5 Pr oj ec t ed Reta il S al e s................................. A-5 P roj ec t ed Pr im e Sa1 e s.................................. A-5 Appendix B CAPITAL COST OF SEABROOK II TO P3NH RATEPAYERS FOR ROSEN'S B AS E -C AS E..................................................... B - 1 Appendix C PROFESSIONAL QUALIFICATIONS......................................C-1 Ed u c a t io n................................................... C-1 P ro f e s s ional Ex p er ienc e..................................... C-1 Prof essional Experience in Energy-Related Research..........C-1 En ergy-Rela t ed Con su lt ing................................... C-1 En ergy-R ela t ed Pu blica t ion s................................. C-2 l
i l
e
,. QUALIFICATIONS Q.
Please state your name, occupation, and business address.
A.
My name is Vince Taylor. I am an economic consultant. My business address is 21 Elm Ave., Richford, Vermont, 05476.
Q.
Please state your professional qualifications.
A.
I have attached a summary of my professional experience and publi-cations as Appendix C.
Briefly, my professional history includes a Ph.D.
in economics from M.I.T., ten years in the Economics Department of the Rand Corporation, and eight years of work on problems related to the economics of energy, with emphasis on the economics of nuclear power and its alternatives.
This history, however, provides little by which to judge the use-ruiness of my testimony in this case. We are attempting here to peer forward into a murky. future to discern what may happen to electricity demand. Foreseeing ' accurately what may occur requires more than mathema-tical models or economic analysis, although both are helpful tools. Most important is an ability to identify those factors that will be most instrumental in influencing the events of concern. No matter how complex and apparently sophisticated the analysis, or how credentialed the analyst, if influential daterminants of the future are overlooked or ignored, the results can be wildiv erroneous. One need look no further than the field of electricity demand forecasting to find examples of catastrophic failures.
The best credential for the present task is success at past s hni-lar efforts. I would like, therefore, to review my own record of forecasting in related areas.
The Price of Uranium In 1977, the price of uranium was over $40 per pound, up from $10 per pound a few years earlier.
Forecasts of future shortages and higher prices were nearly universal. Wall Street analysts were talking about
$100 per pound. A government report on the economics of reprocessing spent nuclear fuel projected a price of $60 per pound for uranium I
before the year 2000.* The concern of government was how to prevent l
Benefit Analysis of Reprocessing and Recycling Light Water Reactor Fuel, U.S. Energy Research and Development Administration, ERDA 7 6-121, December 1976.
I
0-B j l shortages that might impede the growth of nuclear power.
In 1977, I published two research reports that took issue with the conventional view.*
In the first report, I stated:
The major conclusion is that uranium needs in this century appear capable of being met from resources produci-ble at less than $20 per pound of uranium oxide, and prices under $10 per ' pound in the first quarter of the twenty-first century seem at least a reasonable possibility.
Current near-term shortagis and high spot market prices are largely the result of artificial demands created by past nuclear-fuel enrichment policies of the U.S. government The main danger, if policies are not revised (which they wer e no t, until too late], is another cycle of boom to bust in the uranium industry.
l My analysis of the uranium market was discussed in a Wall Street Journal editorial on April 13, 1977 and prompted heated denials of its accuracy from the government and several uranium producers.***
It took several years for the accuracy of what I was saying to be widely appreciated,**** but there is no longer any controversy. The i
market price recently reached $17 per pound, far below my price prediction for this century, reflecting the distress sales that are accompanying the present uranium glut.
I was able to foresee the coming uranium bust because I 1) under-stood the reason for the then-existing shortages, 2) determined that uranium supply would be very responsive to increases in uranium price, and 3) foresaw that prevailing predictions of nuclear-power growth were greatly exaggerated.
Growth of Nuclear Power As part of the analysis of uranium demand, I projected future U.S.
i l
The Myth of Uranium Scarcity, Pan Heuristics, Los Angeles, April 1977, and How the U.S. Gavernment Created the Uranium i
Crisis (and the Coming Uranium Bust), Pan Heuristics, Los Angeles, June,1977 (revised).
Prices are in 1976 dollars.
Inflating to 1981 dollars, the two prices become $32 and $16 per pound.
i Wall Street Journal, editorial page, May 19, 1977.
]
The spot-market price peaked in May 1978 at $43 per pound.
4
, nuclear growta. The projected amounts of installea nuclear capacity were 110 gigawatts in 1985 and 336 gigawatts by the year 2000.*
Industry figures projected 165 gigawatts by 1985,** and an official government mid-range projection for 2000 was 510 gigawatts.***
Present expectations have now f allen below my 1977 projection were meant to be conservatively high in order to assess the adequacy of uranium resources.
My more accurate projections were based on an appreciation of "l) the likelihood of lower electricity growth, given the probability that real electricity prices would increase in the future, 2) the then-existing excess of generating capacity, which had accumulated as electricity growth slowed and which made near-term def errals and can-cellations inevitable, and 3) the economic problems of nucl ear power, which made it likely that coal would be a serious competitor in the future.
The Economics of Reprocessing Nuclear Fuel In the mid-1970's, it was generally assumed that reprocessing of spent nuclear fuel from power reactors would be a normal part of the fuel cycle. A major argument for this process, which would introduce plutonium, a nuclear-bomb material, into civilian commerce, was its favorable economics. Four studies released by industry and government in 1976 showed it would be profitable.
In an analysis of these studies published in 1977,**** I concluded that reprocessing would be quite unprof itable. This conclusion was based on an appreciation of 1) the technical difficulties of reprocessing, which would lead to costs for reprocessing much higher than those assumed in the other studies, and
- 2) t ha t the price of uranium, with which reprocessed fuels must compete, would be lower than assumed by the others. Af ter my analysis was com-pleted, a European reprocessor announced contract prices for reprocessing that were more than triple the average cost used by the others and 50 The Myth, op. cit., Table 4, p.50.
Nuclear News, August 1977.
E. Hanrahan, et. al., "World Requirements and Supply of Uranium,"
Office of Planning Analysis and Evaluation, U.S. ERDA, presented at the Atomic Industrial Forum International Conf erence on Uranium, Geneva, Switzerland,14 September 1976.
The Economics of Uranium and Plutonium, in " Moving Toward Lif e in a Nuclear Armed Crowd?," Minerva, Volume XV, Numbers 3 and 4 (combined issue), Autumn-Winter,1977.
_4 _
percent higher than my estimate.* Uranium prices have f allen far below the value I used. Thus, there is at present no question about the profitability of reprocessing.
Solving the Energy Crisis In the years following the oil embargo, U.S. government policy emphasized the need to expand domestic supplies of energy, especially nuclear power and synthetic. fuels, as a solution to the energy crisis.
In two reports published in 1979, I took issue with this policy.**' In the first report, I concluded:
The maximum contribution of nuclear power, synthetic fuels, fusien, etc. would be far too small to off set the decreases in oil and gas production that, in the conventional view, are expected to occur near the turn of the century.
[but] improvements in the productivity of energy, without any contribution from nuclear power or other potentially danger-ous forms of energy supply, can suffice to extend the lifetime j
of remaining conventional fuels to make them the dominant source of world energy supply until well beyond 2025, provi-ding sufficient time for an unhurried transition to saf e, renewable energy resources.
The jury is, of course, still out on this prediction, but the odds certainly would be considered by most to be more f avorable now than when it was made.
In the second report, which was written before the Iranian revolution and the second round of oil price increases, I presented an alternative forecast to the Carter Administration's Second National Energy Plan (NEP II), which projected that U.S. energy consumption would increase by 18 percent from 1978 to 1985. I projected that a continuation of the existing trends toward Laproved energy productivity (efficiency) would hold the increase to 6 percent and predicted that a f ew relatively simple policy initiatives to increase movement toward increased energy efficiency could put 1985 energy consumption below that of 1978.
Ibid.
Energy: The Easy Path, January 1979, and The Easy Path Energy Plan, September 1979 (revised), Union of Concerned Scientist, Cambridge, Massachusetts t
D '
The oil price rises of 1979-81 provided more of an impetus toward improved efficiency of energy use than could any simple government j
policy initiatives.
U.S. energy consumption in 1981 was 6 percent below the 1978 level, and it is now generally accepted that 1985 energy con-sumption will be below or close to the 1978 level.
The relative accuracy of my forecasts of energy consumption was due to my appreciation of the many possibilities for saving energy by i
improving the efficiency of its use, together with an ability to use existing economic analyses to estimate the quantitative magnitude of these possibilities.
Conclu sion In sum, I believe my past record reflects an ability to identify and concentrate attention on the most important determinants of the future trends under examination. This ability will be a crucial ingre-dient of success in the present case, because future demand for electri-city within PSNH depends on very many factors. Only by identifying and focusing on the most important of these will it be possible to provide projections that accurately reflect future possibilities.
e l
1 l
OBJECTIVES I
Q.
Please describe the objectives of your testimony.
A.
The first objective of my testimony,is to provide an improved
{
methodology. for proj ecting future electricity demand for PSNH. The i
methodology to be presented explicitly relates changes in electricity demand to changes in the real (inflation-adjusted) price of electricity and to changes in economic activity.
i The inclusion of price changes in demand projections taproves upon the methodology of Robert Camfield which j
the Commission found in its Report on DE 80-47 (October 19, 1981) to t
produce the "most reasonable estimate of growth of demand for PSNH of l
3.00 percent annually through 1990," for a business-as-usual case. The model used by Mr. Canfield related economic activity and electricity demand i
but did not include electricity price as an explanatory variable. The i
Camfield model, thus, is unable to shed any light on how demand for i
electricity will be aff ected by changes in the cost of providing it, by i
j changes in the rate structure, or by conservation programs such as the one proposed by Paul Chernik in his testinony in the present docket. The need for an improved basis for exploring these questions underlay the i
Commission's order to open the present docket.
The second objective of my testimony is to use the improved demand-J ptojection model to illuminate the situation now facing PSNH. The dominant element of this situation is the Company's commitment to Seabrook. What are the implications for electricity demand and cost if the Company maintains its present ownership share and completes both units? How i
will variations in the cost of Seabrook aff ect future demand? What would happen to future demand and cost if Seabrook II were canceled?
These and related questions are explicitly addressed in my testimony.
A third objective is to convince the Commission that the demand model presented here, or a sinilar model that explicitly incorporates price, should be required in future demand forecasts presented to it by the Staff and the Company. Only in this way can the Commission be assured that the demand projections used to justify proposes senerating additions are consistent with the estimated costs of such additions.
_ _~
s
. CONCLUSIONS ON SUPPLY DEMAND AND COST l
Q.
On the basis of your findings, what do you conclude about proj ected i
demand and the generation plans of PSNH7 A.
One, future demand for electricity depends critically upon the cost of Seabrook to PSNH.
With the present PSNH ownership share of Seabrook, ten year projections 4
of demand mads with the same economic growth assumptions vary from plus l
3.5 percent to minus 5.3 percent per year as Seabrook costs vary between low i
and high estimates of $3.6 and 8.6 billion. It follows, therefore, that:
Two. in order to make a meaninaful determination of proiected l
' electricity growth for regulatorv purposes, the Commission must first make a determination on the range within which Seabrook costs are expected 4
j to fall.
Three, if Seabrook costs are near to or exceed $5.5 billion, cancel-lation of Seabrook II will reduce future electricity costs and improve the balance between future demand and generation capability.
For a middle estimate of $5.5 billion for Seabrook, sales are projected to decline over the next 10 years. As a consequence, Seabrook j
II generation would be entirely surplus to PSNH needs through the end of j
the ten-year planning period. Cancellation would eliminate this surplus, while simultaneously raducing future electricity prices.*
Four the present PSNH ownership of Seabrook exposes its customers i
to the risk of unprecedented rate increases.
The capital costs of Seabrook, which will not vary with electricity sal'es, will be the dominant element of electricity costs af ter Seabrook I comes on-line.
If costs of Seabrook approach the high estimate, this alone would suffice to cause real electricity prices to nearly triple in j
the next eight years. But even with.over Seabrook costs, similar increases could occur if other uncertainties turned out unfavorably. Demand might fall significantly below the projected levels because of national economic problems or greater-than-anticipated eff ects of past price increases.
Seabrook generation might fall far below expectations because of technical 4
problems or an accident. Given-the fixed nature of Seabrook costs, any Cancellation of Seabrook II would also be beneficial to ratepayers at costs less than the mid-range estimate. The minimum cost at i
which cancellation would be beneficial could be determined by i
projecting demand growth and electricity prices for different i
1 Seabrook costs.
i i
s of these would raise per unit electricity costs above projected levels.
Of course, uncertainties could turn out favorably and costs could i
be lower than anticipated.
But, the question that needs to be considered is whether it is prudent to expose ratepayers to even a modest risk of tripled electricity.ates.*
5 A risk that is far more than modest if the testimony on Seabrook costs and performance presented in this docket by Richard Rosen is accurate.
e 4
1
-c
. FINDINGS ON SUFPLY, DEMAND AND COST Q.
What are the major findings of your testimony with respect to the supply, demand, and cost of electricity of electricity over the next 10 years?
The Importance of Seabroo_k
~
A.
The most important factor, by far, in determining the future costs of electricity will be the cost of Seabrook to PSNH shareholders. This cost is uncertain, depending upon construction costs, financing charges, completion dates, inflation rates, and ownership share, but it is certain that unless PSNH radically alters its ownership plans Seabrook will consti-tute the largest element in the PSNH cost structure at the end of the decade.
The commitment to Seabrook means that there will be no business as usual in the PSNH future. Seabrook in New Hampshire is a Giant in Lilliput.
If Seabrook sneezes, the cost of electricity in New Hampshire will jump.
The present PSNH rate base is $500 million. On present plans, completing Seabrook I and II will raise the rate base to between two and three billion dollars, depending upon the outcome of various uncertainties.
The revenue charge on the rate base has already increased substantially because of 1igh-cost borrowing to finance Seabrook.
It will increase still more as this borrowing accumulates. Including taxes and depreciation,
' the charge against the rate base will exceed 30 percent per year when Seabrook I enters service.
The combination of a multiplied rate base and a high capital charge rate will transform the cost structure of PSNH from the present one, which is dominated by fuel costs, to one dominated by capital costs. Thu s,
trends in electricity costs in PSNH over the next decade will be determined largely by developments affecting the cost of the PSNH share of Seabrook.
The Influence of Price on Demand PSNH electricity sales are significantly and quickly aff ected by changes in real (inflation-adjusted) electricity prices. This general finding should come aIs no surprise, since all economic investigations of electricity demand have found price to be an important factor. Analyses j
l l
l I
i
' i of historical U.S. data indicate that the long-run elasticity of demand.
4 l
with respect to price is in the range of -0.8 to -1.2, with -1.0 being a l
reasonable median estimate.* This means that a 10 percent increase in the real price of electricity will, eventually, cause an approximate 10
~
percent decrease in electricity demand, and vice versa. What may be I
surprising is the extent of the short-run response to price found for PSNH l
electricity sales. Within 3 years of a 10 percent real price rise, l
Tsales were found to decrease by 8 percent, other things being unchanged 2
1
~ (a price elasticity of -0.8).
l Explaining Demand Changes 1
I The determination of the price sensitivity of demand was based on
{
an examination of the history of demand growth from 1960 to 1981. Each of four broad customer classes was analyzed separately. The influences of price and economic activity were studied using econometric and other techniques. This analysis produced four relatively simple equations, one for each customer class. These equations relate annual changes in demand to changes. in electricity price and economic activity.
l Except for the 1974-76 period, following the first oil shock, the quantitative response of demand to real price changes has been very consistent.** Since 1976, the relations 'between changes in sales, price, and economic activity have been very stable.
Table I (Exhibits, S-0, p.1) compares actual growth in electricity sales during 1977-81 with the growth estimated by the demand equations. Sales growth for a given year is estimated using actual changes in electricity price and economic activity for that year. Thus, Table I shows how well the demand equations would have predicted sales growth during the 1977-81 j
period if price and e:onomic data were known accurately in advance. Under this assumption, the demand equations do an excellent job of tracking j
year to year changes in sales growth for each customer class as well as i
total sales growth. For the period as a whole, estimated total ' sales growth Lester Taylor "The Demand for Energy: A survey of Price and Income Elasticities," in International Studies of the Demand for Energy, William Nordhaus (Ed.), North-Holland Publishing Co., New York,1977.
In the post-erabargo period, demand did not decrease as much as normal experience would hav,e suggested, given the size of the real price j
increases that occurred. I hypothesize that this was because of the widespread confusion and uncertainty about whether the increases, which were the first in the history of electricity, were permanent or temporary.
.~
~
~,,
_11 of 2.76 percent per year is almost identical to actual growth of 2.72 percent per year.
Of course, equations such as these always are better at predicting the past than the future. But, the ability of the equations to track year to year changes closely within each customer category over the entire 5-year period gives more than the usual degree of confidence in their future usefulness.
Futurs Demand and Cost The electricity demand equations were used to project Prime Sales for PSNH through 1991 for three different estimates of the cost of Seabrook.
' The present PSNH ownership. share (35.6%) was assumed to be maintained.
For all three cases, mid-range economic growth was assumed: growth in manufacturing of 4.7 percent per year (5.9 percent through 1986) and Personal Income growth of 4.4 percent per year. Real fuel prices were assumed to remain constant through 1985 and to increase at a few percent per year thereaf ter.
Inflation was assumed constant at 8 percent per year.
Results for the' three cases are displayed in Figures I and II (Exhibits, S-0, pp.2 and 3).
- 1) Low Seabrook Cost: The first case uses the Company projection of non-fuel costs based on its estimate of Seabrook ($3.6 billion).
Seabrook I comes on-line in February 1984 and Seabrook II in May 1988 (a two-year delay).
In this favorable cost scenario, demand for electricity falls below the 1981 level for the next 5 years, but begins to grow rapidly after 1986. Under the assumptions used by PSNH, declining equity costs, depreciation of Seabrook I, and continued rapid inflation combine to push down rapidly the Seabrook contribution to real electricity costs af ter Seabrook I enters the rate base. Costs per kWh decline even during the years when Seabrook II is phased into the rate base. Falling electricity prices cause growing demand in the last part of the decade. For the 10 years through 1991, Prime Sales increase by an average of 3.5 percent per year, with all of the increase occurring af ter 1986.
- 2) Mid-Range Seabrook Cost: In the second case, the costs of Seabrook are estimated at $5.5 billion, between the PSNH estimate ($3.6 billion) and the base-case estimate of Richard Rosen ($8.6 billion) l
. presented in this docket. In this case, electricity demand remains below the 1981 level for the entire 10-year forecast period. Average electricity prices (measured in 1981 dollars *) increase from 7.6c/kWh in 1981 to about 13c/kWh in 1988-89. The increase in price depresses demand approxi-mately 20 percent below the 1981 level. Prices decline and demand grows af ter 1989, but demand is still approximately 10 percent below the 1981 level in 1991.
i
- 3) High Seabrook Costs: The highest cost considered for Seabrook j
is the base-case estimate of $8.6 billion presented by Richard Rosen in this docket. If this estimate, based on his analysis of industry-wide 4
cost trends, proves accurate, New Hampshire ratepayers will experience an unprecedented rate explosion. Real electricity prices will nearly i
triple in the next eight years, reaching 22c per kWh.
(By comparison, they increased by 92 percent between 1973 and 1981.) The demand equations predict that price increases of this magnitude would push down electricity demand by over 40 percent, although these increases are so far outside of the range of experience on which the equations were based that this result must be treated cautiously. On the other hand, because these price increases would not necessarily be shared by neighboring states, the assumed manu-facturing growth that underlies these projections may fail to materialize, causing electricity demand to be lower than projected.
Adequacy of Planned Generation Whether the generation plans of PSNH are appropriate depends criti-cally on the cost of Seabrook. Planned generation capability was compared with projected required generation for the three estimates of Seabrook cost.
Results for the three cases are presented in Figures III-V (Exhibits, I
S-0, pp.4-6).
- 1) Low Seabrook Cost: In this case, planned additions to generation are sufficient to eliminate net purchases of interchanged power ** and almost j
all oil-fired generation for most of the 10 year planning period. Under the assumptions of PSNH, the substitution of nuclear power for interchanged and oil-fired generation would be economically beneficial to ratepayers.
All electricity prices, unless stated otherwise, are expressed in 1981 dollars.
Except for nuclear entitlements, which would continue to be purchased because of their low cost.
' Projected demand and generation capability appear well matched, given the cost assumptions of PSNH.
- 2) Mid-Range Seabrook Cost: Seabrook raises prices and depresses danand sufficiently to result in excess coal and nuclear generating capacity
, or 1987. By 1,989, when Seabrook II is operational for the whole year, l
excess coal and nuclear generation equal 75 percent of the PSNH share of S eabrook. PSNH cannot utilize even the capacity of Seabrook I.
Excess generation remains high until the end of the planning period. For the mid-range Seabrook cost ($5.5 billion), projected generation is not sufficient to justify acquisition of the PSNH share of Seabrook II.
All of the gener-ation from Seabrook II (or equivalent coal generation). would need to be sold to other power companies. Under present NEPOOL policies, with the i
fuel prices assumed, PSNH would receive about 4c per kWh, significantly less than the cost of Seabrook II power.*
i
- 3) High Seabrook Cost: In this case, the entire generating capacity of Seabrook is in excess of PSNH needs from 1988 forward, emphasizing how excessive would be the PSNH commitment to Seabrook if the high estimate were to prove correct.
Cancellation of Seabrook II If Seabrook costs equal or exceed the mid-range estimate, the outlook for demand and prices would be improved if PSNH were to divest itself of its share of Seabrook II.
Efforts of PSNH to find buyers for Seabrook shares have been unsuccessful, suggesting skepticism on the part of poten-tial buyers that investment in Seabrook provides an attractive alternative to continued dependence on oil-fired generation.
If this continues, can-cellation of Seabrook II may be the only means available to PSUH to reduce its share of Seabrook.
The projected demand and generation capability, assuming cancellation of Seabrook II and the mid-range cost for Seabrook I ($3.1 billion), are At $2.8 billion for Seabrook II, a levelized charge rate of 20 perc ent, a levelized capacity factor of 55 percent, and non-capital costs of 2c per kWh, the cost per kWh (in 1981 dollars) would be 7.8c/kWh.
- l. f l
shown in Figure VI (Exhibits, S-0, p.7). As compared to the case including i
Seabrook II; there is a much better balance between demand and capability. -
)
Purchases and oil-fired generation are e1Lainated by Seabrook I, except l
for the last two years when some oil-fired generation is again required.*
The improved balance between supply and demand for generation when j
Seabrook II is cancelled occurs not only because generation capability j
is reduced but because demand levels in the latter part of the decade are j
raised (Figure VIT,. Exhibits, S-0, p.8). This occurs because cancellation i
results in significantly lower prices af ter 1977 (Figure VIII, Exhibits, l
S-0, p.9).
1 i
Uncertainties Other Than Seabrook 1
l In addition to the major uncertainty about the cost of Seabrook, there are other significant. uncertainties that aff ect projections of demand and cost.
- 1) Economic growth: Plausible variations in future economic growth could cause growth rate of sales to vary by plus or minus one percent per I
year about the values based on mid-range economic assumptions.
f 2)
Inflation: The projections presented all assumed constant infla-tion of eight percent per year.
If present government policies are continued and succeed (an uncertainty), inflation will trend downward over the decade.
1 Real interest rates and rates of return on equity could also be expected to decline under this circumstance, developments that would be favorable to the cost of Seabrook. On the other hand, approximately 60 percent of l
Seabrook will be financed by debt, whose dollar cost is fixed at the tLae of issuance. Declining inflation will add to the real (inflation-adjusted) burden of this debt in the future.
At lower rates of inflation, real capital charges on Seabrook will; 1
i not decline as rapidly over thme, and the decline in real electricity prices
[
projected for the end of the decade will be less steep. This will imve an adverse aff ect on projected demand. Thus, the upturn in demand projected i
Under the assumed course of oil prices, the fuel cost o'f oil-fired generation is 7.4c per kWh in 1991, roughly comparable with the I
cost of power f rom Seabrook II in the mid-range case (see. previous footnote); thus there is no economic advantage to building.
Seabrook II to replace oil.
1 a
4
~ ~
~--.,r.
-14m-to occur nesr the end of the planning period for the mid and high Seabrook costs would be less steep if inflation slows.
- 3) Performance of Seabrook: The proj ections presented assume that Seabrook performance will conform to a statistical norm (operation at 55 percent of capacity from age 3 forward), but individual reactors may perform far from the norm. Ten percentage points diff erence in the capacity factor will cause about a 5 percent diff erence in electricity costs when both units are in operation.
4)
Inaccuracies in the Demand Equations: The coefficients of the equations used to project demand are themselves uncertain, being based on statistical analysis of data.
In particular, there is substantial uncertainty about the rate at which the electricity intensity of buildings and factory processes adjusts to major shif ts in electricity price trends, such as occurred in 1973. I believe it prudent to allow plus or minus one percent in the projected 10-year growth rates to reflect uncertainties in the equations.
In summary, there are numerous uncertainties, in addition to the uncertain cost of Seabrook, that aff ect our ability to project future demand accurately. If the estimated growth rate is consistently off by two-percent per year, the error will equal (about) 10 percent in 5 years and 20 percent in 10 years.
- i EXPLANATORY COMMENT ON THE SENSITIVITY OF DEMAND TO SEABROOK COSTS Q.
Would -you like to provide further explanation for the sensitivity of future demand to the cost of Seabrook?
A.
It may seen surprising to some that demand projections are so very sensitive to the estimated cost of Seabrook. In addition to the obvious ef f ect of higher Seabrook costs on electricity price and, thereby, demand,
~
there are several less obvious factors that contribute to this sensitivity:
High Proportion of Fixed Costs: When Seabrook II is in full opera-tion, cost itens unrelated to the level of sales (primarily capital charges and taxes associated with Seabrook) will constitute a high proportion of total costs of PSNH. If sales decrease, these fixed costs must be spread over fewer kWh, requiring that prices be higher. Higher prices cause demand to fall, which causes still higher prices. This interaction between sales and price multiplies the effect of changes in the cost of Seabrook.
The exact multiplier varies with the structure of costs and the sell-i ing price for surplus power.
For the mid-range cost case, af ter Seabrook II comes on-line an initial rise in the total cost of electricity of 1.0 percent will lead to a final increase in price of about 1.6 percent; thus the multiplier is about 1.6.
Sales will decrease by 1.25 percent, 1.6 times the decline that would have occurred in the absence of secondary sales-price interactions.
Real Seabrook Costs: The actual costs of Seabrook to ratepayers will be substantially higher than the estimates produced using the methodology approved by the Commission for rate purposes. This methodology, which uses embedded average rates on long-term debt to calculate AFUDC, was used by the Company and by Richard Rosen in preparing the estimates of Seabrook cost cited herein. Thus, the actual costs that will be borne by ratepayers (and that were implicitly used in the calculations of projected demand) will be much higher than these " accounting" estLnates.
Based on the embedded cost methodology, future AFUDC was calculated at 11.5 percent by PSNH and 12.5 percent by Rosen. Actual borrowing to finance Seabrook. had a cost to ratepayers of about 25 percent per year in 1981.*
The weighted cost of new money, calculated using 18 percent for borrow-ing and a cost of equity of 17 percent plus taxes (for a total cost of equity of 33 percent).
I r
. +
Actual interest during construction will be more than double the accounting AFUDC.
Appendix B presents a detailed calculation of the cost of Seabrook II using all of Rosen's base-case assumptions except for the 12.5 percent AFUDC rate. Instead, Rosen's projected costs of equity and new borrowing were used to calculate interest during construction,The resulting estimate to ratepayers for the PSNH share of Seabrook II is $2413 million, 48 percent higher than Rosen's accounting estimate.
The actual cost to ratepayers is what will determine future electricity costs. Understating interest during construction, as does the method used in New Hampshire, can hide some of the true costs of Seabrook but it cannot avert the necessity of paying them.
What embedded-cost AFUDC does is to change the timing of payment for borrowing. A portion of new borrowing costs enter PSNH rates as they i
are incurred ra~ther than being deferred until Seabrook enters commercial operation. The new borrowing does not enter the rate base, but the higher i
costs of new debt raise the average cost of the debt used in calculating the allowed return on the rate base. This was explicitly recognized by j
)
the Commission in a recent order:
" Rolled-in with the embedded cost rate of debt and preferred stock, recent issues tend to increase the earnings rate required to service the senior capital. It is this fact coupled with up-ward movements in short-term debt costs that has profoundly knpacted the overall return requirements of the Company.
Whereas PSNH had an overall return rate of 12.29% in mid-1980
)
(DR 79-187), current estimates of the overall return shown in exhibits filed in this docket range from 14.00% to nearly 16%.*
~
1 The rise in rate of return noted by the Commission is a cost of the
}
Seabrook construction program currently being paid by ratepayers.
In terms of the eff ect on future prices of electricity, the method mandated by the Commission causes electricity prices to rise before the completion of Seabrook, but to rise less at the time Seabrook enters the rate base than would be the case if true interest costs were booked to AFUDC.
The overall rise in electricity prices caused by Seabrook must reflect actual borrowing costs. The full impact of Seabrook on future prices, thus, will be substantially greater than the standard accounting estimates superficially suggest.
Report and Order Nos.15,24 and 15,425 (DR 81-87), NHPUC
. ECONOMIC ANALYSIS OF ELECTRICITY DEMAND Methodology Q.
Could you explain the methodology used by you to analyze growth in consumption of electricity?
A.
The approach used was to develop models that related annual changes in WWh use to changes in economic variables, such as the price of electri-city, manufacturing activity, the number of customers, per capita income, and the price of oil.
I deflated all data reported in current dollars by an appropriate price index; thus all price and income data used were in constant dollars.
In economics, these deflated data are termed "real" income and "real" price data. By using such real data, the distorting effects of varying rates of general price inflation are eliminated from the analysis.
To develop a basis for projecting future demand for electricity, I 1
examined the historical relations among electricity use and various economic variables. To assist in this examination, I used the econometric technique of multiple regression analysis, bu t I did not rely solely on it.
Q.
Can you explain multiple regression analysis and why ycu did not rely solely upon it?
A.
Multiple regression analysis is a technique for deriving the "best" statistical relation between one " dependent" variable (electricity consump-tion) and a number of " independent" variables (such as number of customers and per capita income). This technique involves many a'ssumptions that are highly unlikely to hold exactly in any real world problem, especially one that ranges over as turbulent a period as the 1970's. Thus, while regression analysis can assist in providing quantitative measures of the relations among variables that held historically, an intelligent assess-ment of future behavior requires that additional information, analysis, and judgement be brought to bear on the problem.
Q.
Are there advantages to the economic approach over "end-use" analysis, which breaks down electricity consumption into as many end-use categories as f easible?
A.
Because electricity price and economic activity are explicitly incorporated in the economic approach to forecasting, one can explicitly and quantitatively explore uncertainties in these variables. This is bnpossible with end-use analysis, since price and income elasticities are not available for detailed end-use categories.
Q.
What customer categories did you consider separately in your detailed analysis?
A.
I examined growth in electricity use separately for the following customer classes:
A.
Residential B.
General C.
Industrial 1.
Manufacturing I
2.
Large Commercial and Service These categories correspond to the major rate classes of PSNH customers.
I have separated the sales to large users (termed sanewhat misleadingly by PSNH " industrial" users) into two sub-categorier because, although both See, for example, the testimony of Robert Camfield before the Public Utilities Commission of New Hampshire, DR 80-47, May 1981, pp.4-6.
i I
l
. pay the same rates, they are aff ected quite diff erently by fluctuatAons in economic activity.
Q.
Please continue to explain your analysis.
A.
As a basis for projecting into the future, I examined the historical peric. from 1960 through 1981. The latter part of this period, af ter 1973, diff ers substantially from the former part. Some aspects, though, of the experiente of the earlier period may still be relevant for under-standing the future. I have looked at both sub-periods separately in an attempt to discern which of th'e factors aff ecting electricity growth have operated more or less onchanged throughout the entirc period (providing reasonable confidence that they will continue to do so in the future) and which have been altered substantially by the disruptions and uncertainties that have af flicted the economy (and the energy sector, especially) since 1973.
In conducting the analysis, I have consistently used annual rates of change (sometimes averaged over 3 years) of the variables of interest, calculated as diff erences in the natural logarithms of the variables in adjoining years.* It should be kept in mind that all growth rates cited in the text and tables refer to logarithmic rates.
Examining rates of change in electricity use in relation to rates of change in other variables greatly reduces a common problem in economic analysis: the problem of spurious correlation. Because many parts of the economy have grown together over time, time series for causally unrelated items will show a strong correlation. For example, cigarette sales and electricity sales have both expanded many times since the 1920's; thus a plot of annual electricity sales versus annual cigarette sales will For all practical purposes, these are equivalent to commpnly used arithmetic growth rates, but there are minor quantitative dif f erences for large growth rates (above 10 percent). Logarithmic rates are used because they have desirable properties for linear regression analysis.
. show a strong positive correlation..It would obviously be an error to suggest that growth in cigarette sales must go hand in hand with growth in electricity sales.
The correlation between electricity and cigarette sales reflects, in part, the eff ects on both of growing population and income, but it also reflects very importantly shif ts in taste toward cigarettes and changes in energy technology favorable to electricity that occurred more or less cot emporaneou sly. Because these shif ts were not causally related, however, they did not take place in lock-step; thus a plot of annual changes in
^
electricity sales versus annual changes in cigarette sales will show a much weaker (possibly zero) correlation.
Analysis of rates of change helps to distinguish meaningful from spurious correlations.
- Historical Analysis of Electricity Demand Resid ential Q.
Please present the details of your historical analysis of electricity use in the residential customer categories.
A.
My analysis focuses on annual electricity consumption (use) per residential customer and the average real (inflation-adjusted) price of electricity to residential customers. I have found that most of the changes that have occurred since 1960 in electricity use per residential customer (use per customer, for short) can be explained by changes in electricity price over time (with the exception of the changes occurring in years 1974-1976, an exception that will be discussed later).
Figure 1 (S-1, p.1, Exhibits) shows a plot of annual changes in electricity use per customer versus 3-year average annual changes in the real price of residential electricity.* This plot covers the years 1963-1973.**
Except for the two years 1967 and 1972, all of the data points lie remarkably close to the straight line (shown in the figure) with a slope of -1.0 and a zero price-change intercept of 0.015.
The real price change for a single year is defined as in(RPK(t)/
CPI (t)) - In(RPK(t-1) / CPI (t-1)), where RPK(t) is the revenue per kWh (for residential custorer$, CPI is the Consumer Price Index, and t indicates the year. The 3-year average price change is (1/3)
[ln(RPK(t) / CPI (t) - In(RPK(t-3)/ CPI (t-3)].
Price data go back to 1960, but the first year for which a 3-year average can be calculated is 1963.
. The implication of Figure 1 is that almost all of the remarkable growth in use per customer that occurred during the 1960's and early 1970's was caused by the substantial, persistent decline in real electricity prices. From 1960 to 1973, use per customer increased 6.8 percent per year, for a total increase of 140 percent. In the absence of the price decreases that were experienced, Figure 1 Laplies that the increase would have been only 1.5 percent per year, for a total increase for 13 years of 21 percent.
The persistent price decreases that underlay rapid growth in demand came to an abrupt halt in 1972 (surprisingly, two years before the full impact of OPEC pricing actions), and the rest of the 1970's was character-ized by price changes that fluctuated wildly between negative and positive values. What had been an apparently stable, predictable pattern of declining prices was replaced by a chaotic unpredictable one.
Figure 2 (S-1, p.2, Exhibits) plots annual increases in electricity consumption and annual decreases in real electricity prices.
(Note that both these are annual as opposed to 3-year-average changes.) The contrast in price behavior between the early and more recent periods is apparent.
Also apparent is the sharp reduction in electricity growth once prices stopped declining steadily and instead began to fluctuate erratically.
Not surprisingly, as shown in Figure 3 (.S-1, p.3, Exhibits), the 3-year average of annual price changes lost its predictive power during the first years af ter the 1973 oil crisis, when confusion over future prices was predominant. In a time of confusion, past price behavior does not provide people with a reliable indicator of future behavior. Most decisions that aff ect current electricity consumption, such as, for example, whether to install electric heat, buy a f reezer, install electric rather than gas or oil hot water heat, or buy a larger, frost-free refrigerator, will also aff ect consumption for a long time in the future.
Thu s, expectations about future prices are an important determinant of changes in current consumption.
Af ter experiencing a few years of the oil crisis, people appear to have accepted that a new and diff erent era was here to stay. The trend of changes in electricity price (as indicated by 3-year average changes) once again became a good predictor of changes in electricity use.
For the years
4 l
1977-1981, the data points in Figure 3 all lie close to a line with a j
j slope of -1.0, as did the data points preceding the oil crisis. The I
zero-price-change growth rate appears to have fallen from 1.5 percent f
per year to zero, implying that in the absence of future real price changes, per customer use would remain constant.
4 Based on the above analysis, I recommend that for purposes of estimating future growth in Residential electricity sales:
^
(1)
=6
- 1.0 pr (3), where r
r os = estimated annual change in Residential electricity sales, r
A = annual gr wth in number of Residential customers, r
p = prior 3-year average of annual change in real average price l
of Residential electricity.
Based on the PSNH forecast of the number of residential customers,*
the implication of Equation 1 is that total residential electricity use, i
in the absence of real price changes, will grow by 1.9 percent per year from 1981 to 1991. If real prices were to increase by 1.9 percent per year, the implied growth in total residential sales would be zero.
Conversely, a real decline in price of a' comparable amount would imply l
growth of 3.8 percent per year.
The future course of electricity prices, thus, is important to growth in electricity demand and the need for additional electrical gener-ating capacity.
Q.
Have you considered any factors that would influence future electri-a city prices?
4 A.
Yes, but I would prefer to discuss this subject af ter completing my i
historical analysis of the f actors that have influenced electricity I
consumption for the various customer categories.
General Service 1
Q.
Please present the details of your historical analysis of electri-l l
city use by General Service customers.
A.
My initial hypothesis wa s that changes in total sales of electri-city to General Service custome s could be largely explained 'by changes l
in real electricity prices to these customers and changes in demand for l
. Public Service Company of New Hampshire, Ten Year Electric Load
]
Forecast,.1982 Edition, Table 5-1 a
~.
. the services provided by the commercial businesses that constitute this customer class. As a pr' oxy for demand for business services, I used real New Hampshire Personal Income. To obtain real electricity prices, I divided average revenue per kWh by an index of the GNP deflator.
I applied multiple regression analysis to data on annual changes in total electricity sales and real Personal Income and prior 3-year average changes in real electricity prices, for the period 1963-1980.
The analysis indicated that annual changes in Personal Income were not i
reliable predictors of annual changes in General Service electricity sales, but that changes in real electricity prices were strongly associated (negatively) with such changes in electricity sales.
Income Effects: Common sense strongly suggests that growth in 4
General Service electricity use will be affected by the rate of growth in real Personal Income. The greater the amount of real income, the greater will be the demand for the services provided by General Service customers (commercial businesses). The negative finding of the regrension analysis does not mean that there is little or no relation between growth in General Service sales and growth in Personal Income, but rather that short-term changes in these are not closely related.
If Personal Income growth were to fall to 2 percent per year from its 1977-1981 average of 4.9 percent pev year, and were to stay this lower level for long enough for business to believe the change would be at relatively permanent, I believe that the long-run growth in General Service sales would decline by a roughly comparable amount (aside from eff ects of changes i
4
. )
i in electricity prices). Conversely, increased growth in income would lead, eventually, to higher growth in General Service sales, other things equal.
It is the growth in personal income that
'supperts the expansion of the businesses that account for General Service sales.
The idea that long-run changes in personal income and in General Service sales are closely related is supported by the data for the 1961-81 period taken as a whole. The average annual growth in General Service sales (adjusted for effects of changes in real electricity prices) was 5.3 percent per year, excluding 1974-76. The average growth in Personal Income for the same years was almost identical--4.9 percent per year.
Thus, I believe that for long-term forecasting it is reasonable to assume that General Service sales will be proportional to Personal Income.
It needs to be emphasized, however, that, as the statistical analysis revealed, this proportionality does not hold strongly for short-run changes in Personal Income, such as those that reflect business cycle ef f ects rather than changes in secular trends.
Price Effects: Figure 5 (S-2, p.2, Exhibits) shows the relation between annual changes in General Service electricity sales and the prior 3-year average of annual changes in real electricity prices for General Service customers.
The straight line in Figure 5 is the best fit regression line for the period, excluding the years 1974-76. The zero price-change intercept is 5.3 percent per year, and the slope is
.57.
This implies that a one percentage point increase in real price is associated with a.57 percentage point decrease in sales. This result is statistically highly signif icant.
A diff erent view of the same data is provided by Figure 6 (S-2, p.3, Exhibits), which connects the dots sequentially in time (excluding 1974-76). This tLae path shows that the rate of price decline accelerated throughout the 1960's, reaching its peak 3-year average in 1970. Sales experienced a parallel and even sharper acceleration in growth (the slope of the line through the points for 1963-71 is about -1. 5). The momentum i
' of this acceleration carried sales growth to its peak of almost 12 percent per year in 1972, two years af ter the year in which prices decreased the most.
The pattern of the 1960's reversed in the 1970's. Prices first declined more slowly and then began to increase. Sales growth moved down as prices moved upward. Notably, however, the downward movement did not retrace the path of the 1960's. The 1970's path lies well above the
. earlier one. What could explain this diff erence? Is it likely to persist?
At least three different factors could underly the higher 1970's path:
- 1) A higher rate of electricity-using innovation.
- 2) A higher rate of substitution of electricity for other fuels.
- 3) Faster expansion of the commercial business sector.
Consider each in turn.
Most innovation occurs in response to price changes; thus it seems reasonable that as prices declined in the 1960's, innovation occurred to allow greater use of electricity as a substitute for other, more expensive resources (such as labor and building materials). This is reflected in the rise in electricity growth as prices declined ever more rapidly in the 1960's.
It does not, however, explain the higher 1970's price-growth path, because at zero-price change, electricity growth was higher in the 1970's than in the 1960's. Thus, something aside from price-induced innovation was at work.
And, since the major focus of development af ter the oil crisis seems to have been on electricity-saving innovation, it seems unlikely that changes in the rate of innovation account for the higher path.
Because oil prices rose so much faster than electricity prices during the 1970's (even before 1973), it seems plausible that at least some of the upward shif t reflects a greater rate of substitution of electricity for oil heating.
But, although plausible, the data do not support this explanation.
Indeed, the shif t toward electric heat was high only in the early 1970's, before the oil embargo.
In 1971 and 1972, the increase in electric heating accounted for 2.7 and 3.2 percentage points of General I
Service sales growth (out of totals of 10.1 and 11.9 percent, respectively).
However, from 1972 through 1979, growth in electric heating use accounted l
s
for an average of only.2 percentage points per year (out of total growth of 4.6 percent per year).*
Thus, we are lef t with the third explanation for a higher sales-growth path in the 1970's: faster expansion of the commercial business sector. To explore this third explanation, I constructed an Index of Commercial Activity in New Hampshire (Table 1, S-2, p.4, Exhibits).**
Table 2 showc the growth rates of this Index (and also for real N.H.
Personal Income) for 5-year periods from 1955 to 1970 and for various periods from 1970 through 1981.
The pattern of growth in business activity revealed in Table 2 appears to largely explain the higher 1970's time path shown in Figure 6.
As argued previously, changes in the rate of expansion in business activity must eventually be reflected in the rate of growth of General Service sales. Starting from a low base in the last half of the 1950's, growth in business activity accelerated throughout the 1960's and remainea high until the 1974-75 recession created by the oil embargo. Rapid growth resumed in 1976 and continued through 1979, the last year for which data on business activity were available. This upward move in business growth provides reasonable explanation for the higher level of the electricity-sales: electricity-price growth relation in the 1970's as compared to the 1960's (Figure 6). Underlying the higher level is a higher rate of business expansion.
Figure 7 (S-2, p.4, Exhibits) shows the best-fit line through the points in Figure 6 for 1977-81. This line also provides a good fit for 1972 and 1973. Accepting that, over the long run, growth in Personal Income will, other things equal, cause an equal percentage increase in General Service sales, the line shown in Figure 7 can be used to derive a relation for estimating future sales growth. Using the 1981 value of Data are from Working Papers for the 1981 Ten-Year Load Forecast, PSNH, undated, pp.295-96.
The index equals the sum of Trade and Service components of New Hampshire GNP divided by the deflator for the Personal Consumption component of U.S. GNP. Because of approximations used in constructing the State GNP series and because only a national price deflator was available, the Index is only an approximate indicator of the level of commercial business activity in New Hampshire. It should, however provide a reasonably accurate indication of relative growth rates during various multi-year periods.
- the prior 5-year average of growth in Personal Income (.049 per year *)
to represent the long-run trend in income growth, the best fit line implies:
(2) i =.019 + y - 1.0 p (3)* " ***
c
$ = estimated annual change in General Service electricity
- sales, y = long-run annual growth in real Personal Income, p (3) = prior 3-year average of annual changes in real average price of General Service electricity.
Equation 2 implies that unless the economy soon returns to its previous rate of expansion, the sales: price growth relation is likely to shif t downward from that shown in Figure 7.
Personal Income grew by only 3.0 percent per year in 1980-81, and 1982 growth will also be low.
The longer the current economic slowdown continues, the greater will be the downward pressure on electricity sales growth. At some point, business mu st bring its rate of physical expansion in line with growth in demand.**
Adjustment Lags: The constant term (.019) in Equation 2 implies that electricity sales would continue to grow by 1.9% per year even. if there were no change in either income or price. This does not seem reasonable long-run behavior; thus we need to consider the source of this coefficient further.
The most plausible explanation is that the constant term reflects the continuing eff ects on electricity growth of the very long history of declining real electricity prices.
Q.
The period of declining electricity prices ended in 1974. How could the declines of that period still be contributing to electricity growth today?
A.
A significant portion of electricity use by commercial businesses
-is determined by building design. The electricity intensity (electricity use per unit of sales) of business increased steadily throughout the sixties and early seventies, as lighting levels were increased, air con-ditioning became more widespread, and electric heating became more common.
l This also is the 1961-81 average.
Although not too much weight should be given to results for any one year,1981 may have signalled a beginning of the downward adju stment. Sales growth in this year (which had unusually cold heating-season months) was well below the 1977-1980 sales : pric e relation.
(See Figure 7.)
V-
. l These trends were strongly encouraged (if not caused) by declining real electricity prices.
When prices began to rise, business took inmediate steps to conserve j
on electricity (as indicated by the significant, negative coefficient of p (3)), but conservation measures associated with changes in building design (such as improved insulation, greater use of natural light, and more efficient air conditioning) could be expected to aff ect electricity growth trends much more slowly.
First, building designs reflect long-run expectations about price, and for a time af ter the oil embargo most experts were predicting that rapid growth of nuclear power would soon put real electricity prices back on their old downward course. Only by the late seventies was it widely accepted that the era of declining prices was over.
Second, there are lags in bringing more efficient buildings into operation. Suppliers of equipment must redesign and retool. Architects must make new designs, businesses must be convinced of their practicality, and the buildings must be built.
Third, and most importantly, even af ter the electricity-intensity of new commercial buildings begins to fall, the average intensity of all plants will continue to rise for some time.
Figure 8 (Exhibits, S-2, p.7) is helpful for understanding why this is so.
This figure displays hypothetical but reasonable schedules of the electricity intensity of new buildings and the average intensity of all build ings. During the 1950's and 60's, when electricity prices were falling, the electricity intensity of new buildings was constantly rising.
The average intensity also rose, but because the average reflects th'e mix of buildings, it necessarily stayed below the schedule for new plants (as shown in Figure 8).
By the late 1970's, the schedule for new buildings begins to decline, but as long as new buildings have higher intensity than the average building, the average will continue to rise. As Figure 8 shows, it may well take some time before the crossover occurs and average intensity begins to decline.
The rise in average intensity that continued af ter electricity prices reversed their declining trend will show up as a contribution to the constant term in Equation 2.
l l
l
. Note, however, that as the rise in intensity begins to slow, as it must if prices continue to rise, Equation 2 will begin to overestimate growth in electricity sales. The degree of overestimation will continue to increase over time, becoming very large by the time that average intensity begins to decline.
4 Q.
Is there evidence that Equation 2 overestimates recent sales growth?
A.
No significant evidence.
Q.
What do you conclude from the above discussion and analysis?
The discussion provides a theoretical rationale for expecting the A.
constant term in Equation 2 to diminish over time (and even to become I
negative if electricity prices continue to increase), but there is not yet and will not be for come time reliable empirical evidence on the rate at which this shif t can be expected to take place. This is one of many elements of uncertainty in future electricity growth.
Q.
How do you propose to handle this uncertainty?
A.
I will treat it explicitly when I present my estimates of the range within which future electricity demand might reasonably be expected to fall.
Industrial -- Manuf acturing Q.
Please present your historical analysis of electricity use by Industrial customers.
l A.
I have separately analyzed two subcategories that together make up Industrial sales: 1) Manufacturing and 2) Commercial and Service.
Manufacturing sales constitute about two-thirds of Industrial sales. I will discuss them first.
I first used regression analysis to estimate how annual changes in Manuf acturing sales were aff ected by changes in the real prices of elec-tricity and residual fuel oil and changes in the level of manufacturing activity (as measured by GNP originating in manufacturing in New Hampshire).
Changes in oil and electricity prices are very highly _ correlated (a coeff t-1 cient of correlation of.92).
In this circumstance regression results that include both variables are not reliable.
I therefore concentrated on analyses that excluded the price of oil.
For reasons explained previously, price changes in 1974-76 were poor indicators of future price trends, and I dropped these years from the analysis. Also, Manufacturing sales growth for 1968 (19 percent) was so much higher than' in the surround-ing years (an average growth of 6.4 percent for 1965-70, excluding 1968)
.=
. that I hypothesize that some special, non-recurring factor (such as an acquisition of raw service territory or the start up of a large plant) was at work,* and so I omitted this year from the analysis also.
A regression was performed on data for the years 1963 through 1981, excluding the years mentioned above. This yielded the following:
(3) s
=.029
.62573 (3) +.52s, where IM 73
= estimated annual change in sales to Manufacturing (Industrial) customers.
g(3) = prior 3-year average of annual changes in the real price p
of electricity to Manufacturing customers (reported prices were deflated by the U.S. Producer Price Index).
i
= annual change in manufacturing activity, as measured by real g 7g Gross National Product (GNP) orig inating in Manufacturing.**
4 I have queried PSNH on this. The reply was negative.
(See response l
to request 6 in PSNH Responses to CLF Data Request Set No. 7, 1
July 30,1982). Company data show however, that 55 percent of the j
sales increase in Manufacturing occurred in Sales to Paper Industries (Working Papers, op. cit., pp.346 and 356), generally a relatively
]
slow-growing sector.
For years prior to 1969, real Mfg. GNP had to be escimated by using j
the U.S. Producer Price Index to deflate current aggregate dollar values of New Hampshire Manufacturing GNP. Real GNP data for 1969-80 were available that were calculated by applying sectoral U.S. prica deflators to current dollar figures for each of 22 manufacturing (2-digit SIC) categories. (Source: Gross State Product of New England 1977-79 and 1969-80 (two separate issues), Federal Reserve Bank of Boston.)
Because the mix of industry in New Hampshire diff ers significantly r
from the national average, the later-year figures are substantially j
more accurate indicators of real New Hampshire Manufacturing Activity.
2 Real Manuf acturing GNP of New Hampshire for 1981 was unavailable.
4 It was estimated using an analysis of the relation between U.S. and New Hampshire economic activity presented in Appendix A of my testimony.
j There are significant differences between the estimates of Manufacturing GNP derived by the alternative (aggregate and sectoral) deflation procedures. These differences, which can be seen in Table 3 (Exhibits, S-3, p.1), raise questions about the accuracy of the data prior to 1969. A regression on only the later-period growth data (1970-81), however, produced the same coefficient for gIM (0.5) as did the regression for the entire 1963-80 period. This suggests that, for the purposes of the present analysis, the deficiencies in the earlier-period data are not damaging.
.~..
. d The regression has high statistical significance, explaining over 70 percent of the variance. The t-values of both coefficients exceed 4.0.
The constant term in the regression equation (2.9 percent per year) measures growth in electricity sales that are not statistically " explained" i
by the variables considered. In the analysis here, the eff ects of long-run I
trends in price and manufacturing expansion are not fully captured by the short-run explanatory variables used.
4 The coefficient of manufacturing growth in Equation 3 is about 9
one-half, meaning that only one-half of the long-run growth in manufacturing is thereby predicted to be translated into electricity sales growth.
Over the long run, though, ele ~ctricity sales and manufacturing activity should change proportionally, other things being equal. Part of the i
constant term certainly reflects the one-half of long-term manufacturing growth not captured.by the coef ficient of gIM. Manufacturing output increased by an average of 4.6 per year during 1963-81 (and also during 1970-81).
Incorporating long-run manufacturing growth in the prediction equation reduces the constant term by one-half that amount. Equation 3 becomes (rounding the coefficients):
(3a)
=.006
.6 pyg (3) +.5 g73 +.5 g73, where 73 k
= 1 ng-run rate f manufacturing gr wth IM Only a small constant term remains. This is plausibly attributable to the lag phenomena that were discussed under General Service sales.
Although new manufacturing plants seem likely to be decreasing in electricity intensity under the force of rising electricity prices, the average intensity for all plants may still be increasing. Figure 8 is as applicable to the manufacturing as the commercial sector (although the rates of adjustment i
will almost certainly differ in the two sectors).
If this lagged adjustment is Laportant, Equation 3 will increasingly overestimate the growth of electricity sales to Manuf acturing as the proportion of more electricity-efficient plan'ts rises over time.
Figure j
9 shows actual and estimated sales growth. The estimating equation does a good job of tracking actual sales growth (even for most of the years omitted in the regression used to derive the equation). Actual growth-did f all somewhat below estimated growth in both 1980 and 1981, but the variance in year to year estimates is. too great to have confidence that.
f I
4
,,-,----c.
4
this constitutes a trend. For the present, the esistence and rate of this downward adjustment in sales growth must be considered uncertain.
Q.
Might not a portion of the constant term in the estimating equation reflect above-average expansion of electricity-intensive industries in New Hampshire?
f A.
It is true in theory that this could be the case. Rapid growth of very intensive users of electricity (per dollar of production) would raise the overall electricity intensity of manufacturing. This would be growth over and above that related to expansion in the size (as measured by Manuf acturing GNP) of the manufacturing sector. A detailed calculation for the 13 Manufacturing subcategories used by PSNH shows that this was not the case for the period 1972 through 1979, the only period for which data were available to permit the calculation to be made.
For this period as a whole, real Manufacturing GNP increased by 43%. If electricity-inten-sities (kWh's per dollar of GNP) for each subcategory had remained at the 1972 values, Manufacturing electricity use would have increased by
+
40 percent; thus diff erential growth within subsectors tended to lower the overall electricity-intensity of manufacturing.*
Q.
Does this complete your historical analysis of Manufacturing sales?
A.
Yes.
Industrial -- Commercial and Service Q.
Please present your findings for the Commercial and Service portion of Industrial sales.
A.
Commercial and Service (C&S) represents sales to large stores (the largest subcategory), hotels and motels, banks and insurance companies, private schools, hospitals, sales to utilities and miscellaneous other enterprises. This has been historically the fastest growing major category of sales, increasing at an average rate of 11 percent per year during 1966-81.
In analyzing C&S sales, it is helpful to treat sales to Utilities separately, as it is a special, atypical category. As can be seen in Table 4 (Exhibits, S-3, p.3), year to year sales fluctuate widely and show no growth trend. Prior to 1977, a large part of this category repre-sented sales to a gas pipeline, which then ceased business with PSNH.
Calculation based on constant-dollar GNP originating in 2-digit SIC manuf acturing industries (Gross State Product of New Hampshire, Federal Reserve Bank of Boston, various issues).
. Growth in more recent years represents construction power for Seabrook.
These sales will begin declining as Seabrook II nears completion.
For purposes of forecasting, I will project Utility sales separately.
The remainder of this section will consider C&S sales excluding the Utility category.
The most striking characteristic of historical growth in C&S sales is the distinct difference in growth before and af ter 1973.
For 1966-73, growth averaged 19 percent per year. For 1974-81, it averaged 3.4 percent 1
per year. Before 1974 there was only one year in which growth was below j
14 percent per year. Af ter 1973, there was no year in which growth exceeded 10 percent.
The full explanation for this sudden change in pattern is not apparent from the data considered here. The primary factor was undoubtedly the shif t from declining to rising real electricity prices, but this alone does not seem sufficient to cxplain the entire diff erence.
Figure 10 shows annual changes in C&S sales versus the prior 3-year average of changes in the real price of electricity to CSS customers. As usual, the points for 1974-76 lie above the sales: price relation for other years. Excluding these atypical years, there are two distinct clusters of data points: 1966-73 (with 1971 being atypical within the cluster) and 1977-81. Drawing a single best-fit line through both clusters (excluding 1971) would produce a coefficient of sales growth with respect to price change (a price elasticity of sales growth) of almost -2.0.
This is a far greater sensitivity to price than found for any other category of sales and does not seem reasonable. A more plausible interpretation is that there was a downward shift in the underlying growth rate
- between the two periods.
If we accept -1.0 as a reasonable approximation of the price elasticity of sales growth (the elasticity found for General Service), the underlying growth rates would be about 13 and 8.5 percent per year in the two periods. Figure 11 shows the implied price: sales growth relations for the two periods.
The one that would occur in the absence of changes in price.
' The sales: growth line based on an assumed price elasticity of -1.0 underestimates sales growth in 1980 and 1981. This is undesirable for projecting future sales. The best-fit line through the 1977-81 data points in Figure 10 has a slope of
.5 (a smaller price elasticity) and implies an underlying growth rate of 5.8 percent. This gives a much better estimation for 1980 and 1981 and is therefore pref erred, although the price elasticity seems low based on the results for General Service.
If so, the equation will overestimate sales growth when prices are rising rapidly.
The underlying growth rate for the recent period (5.8 percent) l implied by this analysis exceedo the growth rate of the New Hampshire economy as a whole. This does not seen surprising for this category of sales, f ir st, because it contains relatively rapidly growing sectors (banks and insurance, hospitals) and, second, because large businesses are undoubtedly growing faster than small ones (reflecting the national trend toward bigness).*
{
There is no empirical way to determine the sensitivity of C&S sales l
to changes in the size of the state economy. Sales are not sensitive to j
short-run changes in state income, and over the longer run income growth has been relatively stable compared to many other factors that influence sales. Thus, one can only appeal to reason. As argued above, it dees seem reasonable that C&S sales expand more rapidly than the average rate of business expansion. An estimating equation that allows for this eff ect and conforms to the 1977-81 best-fit price: sales growth relation (using the j
1977-81 average annual growth in personal income of 4.9 percent) is:
(4) i 1.29 - 0.5 pIC(3), where
=
IC h
estimated annual change in (Industrial) Commercial and
=
g Service sales, i
y = long-run annual growth in real Personal Income, l
pg(3) = prior 3-year average of annual changes in the price of Industrial electricity, deflated by the U.S. GNP implicit price deflator.
The more rapid growth of the early period may reflect in part the initial surge of larger businesses 'into New Hampshire. The slower present growth may, then, reflect a more mature economy in this respect.
. The coefficients in this equation are more uncertain chan in previous equations. It will be important to test the performance of the equation in the next few years.
Q.
Do you believe that the lag phenomena discussed previously is important in C&S sales.
A.
The major expansion in this sector occurred in the late sixties and early seventies, when electricity-intensities of buildings were already at high levels; thus the average intensity was probably close to the intensity for new construction in 1974, when the trend of electricity prices reversed. If this is correct, there should be less of a lag in this sector than in manufacturing and small commercial business. This may also account, in part, for the apparent downward shif t in the underlying rate of sales growth between the earlier and later periods, i
I have chosen coefficienes that make the const. ant term in Equation 4 equal to zero. This implies that average and new-construction electri-city intensities are equal. This is unlikely to be the case, but there are no good grounds for assuming one is higher than the other. Again, this is an area of uncertainty.
Q.
Does this complete your historical analysis of browth in electricity consump t ion ?
A.
Yes.
Q.
Could you please summarize the equations that will form the basis for your projections of future demand for electricity.
A.
Resid ential (1) ir " "r - 1.0 p (3) r General Service (2) 7 =.019 + 9 - 1.0 pc (3) e Indu strial-Manu f ac turing (3a) I
=.006
.6 573(3) +.5 gyg +.5 g73 7g Industrial-Commercial and Service I')
= -0. 5 pIC( ) + 1.2 9 IC
]
e
. l PROJECTING FUTURE DEMAND FOR ELECTRICITY Q.
Please describe how you will use the demand equations to project future sales of electricity.
A.
The electricity demand "model" described by the four equations makes future growth in electricity a function of future changes in electricity prices, economic activity, and population (households). Thus, given this model, the task of projecting electricity demand becomes one of specifying future changes 'in these underlying variables. There are significant uncertainties about the future course of the economy and future prices of electricity and these uncertainties will be reflected in the projections of electricity demand.* The result of the analysis will be a _ range of rates of electricity growth, with the values within this range explicitly linked to the values of the underlying variables.
Q.
What are the most important uncertainties aff ecting future electricity growt h?
A.
Future economic growth and fossil fuel prices are both uncertain, bu t by f ar the most important uncertainties concern Seabrook. The cost, timing, and PSNH ownership share of Seabrook are the dominant sources of uncertainty about future electricity growth.
4 Q.
How will you deal with these uncertainties?
A.
Because Seabrook is the most important factor in determining future demand growth, I will make separate projections of grouch under different assumptions about Seabrook cost and timing.
I will consider the eff ect of cancelling Seabrook II as a means of reducing the PSNH commitment to S eabroo k.
Underlying Assumptions The projections under diff erent assumptions about Seabrook cost are made using a fixed set of values for future economic and population growth, fuel prices, Seabrook performance, and parameters of the demand model. Values were used that I judged to be near the middle of the range of uncertainty.**
Growth in the number of households (residential customers) is also unc er ta in, bu t to a much lesser degree.
See Appendix A.
Uncertainties in the values specified are considered separately.
The Basic Economic Environment The basic economic environment assumed is favorable to expansion of electricity demand. Manuf acturing expands rapidly:
5.9 percent per year through 1985 and 4.7 percent per year through 1991. Real personal income continues to grow rapidly, averaging growth of 4.4 percent per year over the next 10 years. Fuel prices are constant in real terms through 1985 and rise only modestly thereaf ter (2 to 3 percent per year). If it were not for Seabrook, the basic environment would imply electricity growth of 2 to 3 percent per year throughout the next decade. Seabrook, however, will radically transform the cost structure and growth prospects of PSNH.
Projecting Future Electricity Costs Determining the eff ects of Seabrook on PSNH costs is an ectrencly complex problem. The eff ects depend not only on the cost and timing of construction expenditures, but also on the costs of borrowing, regulatory lags, and tax effects. PSNH and Rosen have made projections of future revenue requirements under various assumptions about Seabrook cost and timing using computerized financial models. Their projections of non-production costs (primarily capital-related costs, but including costs of administration and transmission and distribution O&M) are used as the basis for determining future electricity costs under diff erent assumptions about Seasrook. Fuel and O&M expenses were estimated separately, using the basic-environment assumptions and estimated levels of generation.
These expenses were added to the estimates of non-production costs to obtain estimated total costs.*
The cost projections were then combined with the demand equations to project future electricity sales. Append ix A describes and illustrates the methodology, i
l t
i Alternative Seabrook Cases Demand and electricity prices were projected for four diff erent The computer projections, which ended in 1989, were extrapolated to 1991 to permit 10-year projections of demand to be made.
-3 8 -
Seabrook cases.
- 1) Low Seabrook Cost is based on a financial forecast of PSNH.*
Seabrook I is assumed to begin commercial operation on 2/28/84 and Seabrook II on 5/31/88. The projection reflects the basic PSNH cost estimates of $2472 and $1549 million for the two units. No information is provided on how the two-year delay for Seabrook II aff ects the estimated cost, but any changes are implicit in the cost figures projected by PSNH.
- 2) Mid-Range Seabrook Cost is based on Rosen's Scenario A-4.
On-line dates for Seabrook I and II are July 1985 and 1988. Estimated costs are
$3129 and 2413 million.
- 3) High Seabrook Cost is based on Rosen's Scenario A-1.
On-line dates are the same as in A-4, but estimated costs are $4010 and $4561 for the two units. This is Rosen's "most probable" estimate.
- 4) Mid-Range / Cancel Seabrook Cost is based on Rosen's Scenario A-8.
Seabrook II is assumed cancelled on July 1,1982, and sunk costs as of that date are assumed written off over 10 years, with no rate of raturn on the unamortized balance. Seabrook I comes on-line in July 1985 and costs $3129.
In all of the cases, the current PSNH ownership share (35.6 percent) is assumed. Rosen uses the PSNH estimates for Administration and General Expenses and Transmission and Distribution O&M Expenses, so these are the same in all cases. Rosen's projected borrowing costs are generally higher than those assumed by the Company. The financial projections of PSNH and Rosen both assumed constant 8 percent inflation throughout.**
Rosen's estimated nuclear O&M costs were used in his scenarios. The PSNH estimates of O&M were used in the Low-Cost case.
Financial Forecast Computer Solution, Scenario Designation G&R i
82-3, PSNH, 6/7/82, NHPUC Docket No. DF 82-141, Attachment Panel-4.
In applying these cost data to the demand models, they were first converted to 1981 dollars so that results could be presented in constant dollars.
But, this conversion does not eliminate the 4
effect of the assumed inflation rate on the results.
Inflation af f ects both the estimated real cost of Seabrook and the rate at which its real contribution to the rate base declines af ter entering operation.
. Proiected Sales and Prices 4
Projected electricity sales and prices for each of the four Seabrook cases are presented in Tables 5-8 (Exhibits, S-4, pp.1-4). For the 10 year period considered, the average rate of electricity growth ranges from
-5.3 percent per year in the High Cost case to 3.5 percent per year in the Low Cost case. Electricity prices in 1991 range from 5.5c per kWh to 17.4c per kWh'(in 1981 dollars) for Industrial customers; prices for other i
classes exhibit sLnilar variation.
These results demonstrate the dependence of future electricity prices and sales growth on Seabrook costs.
Q.
Does this complete your testimony?
A.
Yes.
A J
i i
N
)
t j
l 4
n.
.,y --
^
l I
l EXHIBITS to Accompany the Testimony of Vince Taylor before the Public Utilities Commission of New Hampshire Docket No. DE 81-381 i
1 i
l October 8, 1982 Conservation Law Foundation l
t
c S-0, p.1 Table I Estimated and Actual PSNH Electricity Sales 8
l Growth for 1977-81 (Percent per Year) 1977 1978 1979 1980 1981 Residential Estimated 1.2 2.9 2.4 1.7
-0.7 Actual 2.6 3.2 2.2 0.8
.1 General Service Estimated 5.1 6.5 5.9 3.6 3.2 Actual 5.0 6.5 5.7 4.7 2.5 Industrial Manuf ac turing Estimated 6.9 7.5 5.5 0.1
-1.0 Actual 2.4 6.8 4.5 2.2 0.1 General & Service Estimat ed 4.5 5.6 4.1 1.5 0.4 Actual 2.4 6.8 4.5 2.2 0.1 1
Total" Estimated 3.9 5.2 3.8 1.5
-0.6 Actual 3.7 6.8 3.4 0.6
.9 i
a.
Estimated values are based on Equations 1, 2, 3a, and 4 of this j
testimony, using actual changes in real price and economic activity.
i i
b.
Excluding Sales to Utilities within this customer category.
I c.
Weighted by the fraction of sales in each customer class.
j l
Figure 1 Projected PSNil Prime Sales for Three Estimates of Seabrook Cost -- Present q
3 10 Ownership Share, Hid-Range GWH Economic Growth 8
Low-(_ ash 7
i s
Md-Ran3e v
Hof-Cui a
i ACTUAL PRO J ECTE D 7
I m
6 i
e i
i i
e y
1910 1915 1980 1985
/190 Y E l? R j
,up m i
d e
d.
g o
C n.k C
b 3
J.
w h
M o
L t
0 9
. 9 1
NE D
T C
E T
5 9
0 9
1 R
P r
o lff a
o es t)
R esss I
ceor l i t C a I
iN ra l
e SP mkl r
P i oo f
R u
yt oD g
d t sr i
eiEb1 0
F t c 8
A 9
ciee9 9 E ereS1 jt r
(
1 Y
och r eT Pl E
L R
U TC A
s 7
1'l 7
+
v4 0
2 9
n 1
0 5
0 S
s 3
1 2
I 4, Prh i
e u c
k 1
l l
1
.1 i
Figure III i
1.
l PSNH Planned Cencration 3
3O.
capability and acquirea ceneration -- tow seabrook (r.W H l._.r.!... i..
I i
..i cost I
g
- i..i.'..
i i
e a
a t.ld,. I i
i roJal G nkcalion e
i i
i C c-f' b l.
i a-i i.
i JI EQUIRED R
2.
i i.
i i
\\-G C N E R A Tl0N l.
1 rvel sul,s of i
-J C oc l c.4 N.,<.lec,e i.
iPo.sct14sES Oilt. M7E m)
Ipen.M, n i,.
' Re qunea p l
.i.'.
i i
i i
s l S;t... 4. _ \\ %\\> :
i I
i i
n I
?
i L.. 1 L
- 1..i i
. i..
. i.
i t
j l
1 i CDAL l
?
i
.i. (
i i
i t
i p.
l t
l I
['
6
. t' i
4
.l.....
b
_t _.
{
. J
"... k,. k
.5
. l!!
' '.i l. [
l..
I i
l.:
. i. ;. i.. L t.. l_
i I
i i
i i
i r
l I
... El jrtd r.NOCL 4 A l EWis 7L E.M N
u I
1 6
.i
.. i..
1
.i t.
I Hvono I
I
.j. i f
I
-*---A I
-6 f,1460!
liSE I
. Ii.70:
- a. Esc lv) $ o d.,5)..
C i
ra..., w.is.u I
,..j.
1.
Y e6 r-i
j i
e 1
I 3
Figure IV j
l t
PSNil Planned Generation f
3
/0 i
Capability and Required 1
i
[
i Ceneration - Mid-Range GW H*..
1 i
i.
i j-Seabrook Cost i
_.i i
j l
l i
[
l.l i
i j
1
.l0 i
i i
- i. ;
i C g5 l
- l. 1 i
l t
i 1
k l
l..
~
3 l
[
j 3
l p/ori.os t Generajien l
i a
g
.j g..
I, i
PURCHASE.S'$Eh2GD I
\\
5 (NST SALBS OF e
i i
QlL.($TEAln) I t
COAs. AND NUCLEAll t
j GENERATsoN l
i i _.
i i
I REQUIRED i
5:
i
?-
Plane.jNoc.lew i Ganen arion 7
i y //./
i i.
- l. i i
i L..
L...
i i
i i
.l m.
l l
. l..
l s
i
.g i
l i
1
.C o B L.-
l i
i I
j r
i I
i i
l l
i Ged edfi'ent i
i I
l l.
!e i
I l
1
- .i.
i i
i i
m D_Nh'CL2 4A f E#r#71.h.MdN
.__ _l
.j.
DJ AN b
6 a.
1 I
}
l HYDAO i
(
l j1 i
j
.o
.. !...L. 1
.1 e
e vi
._l..
t l
_j
..__{101 -
t t.
I1980! !.i.
i i118.5-l
_ _._ j_.. j..
19 A. E.cl.,j. 3. L,d,,.i C
._i.,
I j
.I...j..
I
!. ;: li.
T.A..q v.4sA i
i
..I. _ i.l..lY Bg r' I
i i
- i..
1 l
,r I
i i
.g l
., i..f.
!. !. 4,.
i i
l i
L.:
Figure V j.
l
[
.t.
1 3
PSNil Planned Generation
/0 Capability and Required j
.i i
C*""*'i
" -- nigh seabrook j
CrW H:.-
Cost e
s
.i....
e t
a g
g,,
l l'
t l
- i.... l i
i i
i 10 i
g j
.h lp T4kl Ge,. cal..n 4
i i
I i
cc. n t..l.1 a
.l.
3 I
i i
i i
'i' Mon. oil G enera tion e
[
t i[90RCHASES OIL (STEAM)'
'l
,(REQOtttED!
l e E T.f 4 L E.5 O F N
i
~ '
I (COAL AND L-N uc.t.EA ft t.
l I
i G EN E RRTio N
' 8t
.Planne.J Nad u c t
r I
s j
a
. RE QulRED CrEN E R STio N
... i i
t
- i.,
l.
F
.j..
i i.
1.
I i
i Conn.
i i
sea 41,!.,i i
i i
i i
I
- l. I l
l i
i i
i t..
=
AAlEnlun$ mint.
l s
i
.s.
- . s es it!
.t (L
I f
g lt i
1 8
Hvono i
I t
i j
..... L._ _.. I t __t. !. i i
i
.._l.___.. l
{ f 939o;.l
.j
!.l$80!
l 1.l
' 19'g fi-j
- 4. g,,tj, y,,,6
,l
_r i
I l
....i _. {
.i..._t.. i i
LL.,u d o,.1Al 1
i
!l.
i iYMr!
! ll..i..
i.,
j, e
.p i.
I i
l j
l' I
e I
_. ~ _ __. _
.~._ __..-. _..___.._ _
Figure VI Cancellation of Seabrook II:
t J
b Generation Capability And Requirements -- Mid-yy Range Seabrook Cost l
i i
i t..:
i i
i
- i..
I l
i e
i i.
l
..i 10 I
i j
i i
i
.. _l _
- i..
.i. I..I l..i i
2 l
8 i
i i.l i
1 i
i i
l t...
l j
I
.l l
t i
t t
i i
t l
' Tok l Genero liers i
..,. c< p.ut..l.1 "
i
.i.
i 3
'fuRC Ht15 Es O lli ($TE'AM)
REqul AED CENEtt.ATION
~
1 REQu t R TID
..a
/:/
N on-c.1 G eritmfien f
- i. l
, 5.
i
'#d sulu of coulunj P.L 4 NN E D..
V'C.L'E A R.
i i
. i, i
a l
t C. O A L:
i..- !
s t
I i
- l l
i i
l 1
I i
i i
i r
t 3
t i
kz80 sed, NQCL154A j t!# irs 7t.E.MyNds. _
U2
.. i.
g i
l HYPAo 1
i i
{
[
2.
- G 1
1 1.
4
[.
i y
1180 i.i.i19'B5i i
i. 4..
i I
'_..!..i roa,,,3v.o.kd,o, 3
E c.lo).,
C
,.lMO! ri a
i l
l i
.. t..
l
..&sA o
.'.i iYe4ri
'_-l__'
.'.1,
m6.*b o
/
9 e
0 w'
7 I. '
'i
. j e
n 7
f.
o 9
ec*
og u.
(
8 8
m k 8
- 8 o9 V
o1 y,
r g
7 t
- b. l
(,,
[8 o u g
eJ S
p
. h E
g e
S g
t d
e 5
4 8
e e
Y f
o 8
r
.t
.j P
3 t1
- 'O
.l
..' i 4{ *.'
ig
- . fi *.
a t~
4 2 'R N
f
[
n
.e,
. Il
. 'Y. '
1 t*
8 o
l*;i
.+
- lt yii
. ii S
e
,7
- 3 t
7 g
l n
t a
i s
e e
l so v
l eC s
enl l
')
coak e
n S o I
aI o
A I
CI er 7
V mb f ki a N
e oore r
OPS u
t r g
cbd e i
eaeg I
F f et n 7
fS ca A
E eR j-e od h
ri 5
T PM 7
Y 7
3 1
f f
o 0
0 0
o 0
I o
0 0
0 o
0 0
o e
0 0
0 o
0 r
0 9
0 s
o 0
5 W
r.
S 4
V 3
a o
5 2
1 t
k G
,Il
'j1li1
~
ug m.*
n i i.:
L i*
6
.(
i 6'
1
. i
.,Is. l
- 5. L.
o.
t -,
J
[
1
.,~
m_ [
9t i'
t.
L. {.
+41l
.. It_
l..L. - j.
{
~
I I
jy
~
. J.
iT1 a
._!. L s.
1 d
7 j,^
- k. 4.-
_. L_
& 1,? k.1'.
. : i e le.
t_ i i_
g ].
._. j.
~
[{
.i!
?
?
3
.r
. p.
_, h.
h
- ip.1:i,
~
e.
.. ( '}t.
.i1 n
),
~
.p u 8
. i.
Il' e
Sc
. L.* g.
.[:
~
. i.
hi s,
..* j.
.i.
8
[
r N
/
x j,
7 8
/
r_
f n
6 i
9 r/
T a
L
/
c
,u s
f g
j a
a r
g
/
o.
.rg
/
R V
k g
P G
m N
a a g 3
y 9
e S
2 t
i sy.
t
/
S
/
/
O
/
G F
'1 gd -
t ne t
i t ss d
u l ceo l ecC)
B eji h
7
- )
/
corkW c
I nrP ok I
aP o
li A
I C
yr r V
nt b e 7
~
f oiap 9
e o
ce r
I iS C u
t I r g
c t e1 iek c g8
(,
F f oen9 4.
7
\\
f ol a1 E rER(
b eal d h eai M
TS eM
/
R
/
tf
)
./
/
3
~
~
g h
b W 2
l 0
9 s
g g
y 1
1 1
f 1
t h
1I i
l i4
,l!
llI l
Figure 1 i
l Annual change in Residential Electricity Use'per Customer versu s 4w riva l 3-Year Average Ck in Change in Real Price of FessJe el Elec-Residential Electricity:
f ric l
../2f c us4.3 u e. pe,.
1963-1973
.,. c.
'69 *qx
. 10 e e 1g '61 e
e
- 69
?
.oif l
e y
'6& *
.of
,& Y e
'6r e'E 1
oaf 1
0
' ' 'O 017
.br
.. gv b.S -
ATE
. ; of. ;
l
',,jf,
+
C hah =, id je/gt el c/,,c) i 3
7f 8C f18b. ( 3 Q4M Apassg
-an
+sy,>..
.... a.
I
- h
.i
- l i,' ' g t
i e
.1
.,'8 l
?
l i
i i:
-. 950
, I i
m l,;!j...
i
?
t
'.o e.
4
4 S-1;..p.2
..i _.
. Anna:1_C.hangszi.in_Ris.identin(..m Electricity Use and Real
. Residential--Elect icity[ Prices -
.s.
i
.. Percaivif....
... J.__.... j Pec.
2
._...._._...1......
yyy,.
--.____..+..q...
d
{..Tvierea.se.iii..Res i en) :"
oe ev eo.s e in. \\t ea.l.
C ud o,,.e e.
u
..] o...
. ).0s e.
er i
._l
_..._... _ _..Pric...f. h si den h ot l Ele c.] rac ih 1
..,/.: -... _..%
I
.g
.. p.
'l p
. S..
g
_. /.
p.
l\\..
/
....f.
g_... __......_j_y.
1
/..
I-l-v km a
a f
-=
,i e
a a
q g
1966.
1970 ti 9..
I eo.
Yea.r.
l
.l.
g I
I i
i.
Ij r
s
-f.
l i
i.
1 I
I l
s l
I l
s 5
l
-to.
s I
l' i
i
-/f.
V I
e 1
i 1
Figure 3 Annual change in Residential Electricity Use per Customer
- anuAlCLA3e versus ik elecle.c'.h 3-Year Average us e y er-Change in Real Price of resJe.did
../2f Residential Electricity:
cus lo
,e,-
1974-1981
,.lO
.otf 196 3-1973 'l y R e.l d so n J
)
- 971-ei) j
.of ite.ldion j
.'76
.
- oaf
'7g H
.10
.oir or
.otr
- /
a.c
.sg by
.'ogg
,,f
.gf
~
c t.53e., ce a.1
- e. l e c.t r.c. 43 prie.s
@'O
,15
( 5 yaue avecq e of an%- Al cL < ges) 02f
'1 un "y
..aso 1
a I
i
S-2, p.1 There is no Figure 4
,.i
.l.._
i l I i i l i*I I i iiii l I i iit il l i
. l.. l.. - j. l.
l 1.
ll I i r !..
- .i.
j Figure 5 g f i
i 4
.g...
3._4 u_.
. ; _.g..
.. g..
$'.li 1 -. !- !j<
j
.I i
I; Annual Change in General Service Electricity Sales,,
[..
,, i versus I
i
. Abnuo'l ]chey...
3-year Average Change in Real Price
~
l i
t inl.C r.jerill.Sch
- c of General Service Electricity
'fldc.I riegl'4' Slp Average 1963-81 Relation, Excluding 1974-16 j
i
.i e i.
i.. i. \\.....
..u24.
. _ 9.. _..p...
q
_..g.,_q..g.
i j._... _.i.
4. g..p.
. _.L.
.L
.. l u
. 2t e
i, i
t g
l,
i l
. 1.
6
{.
gl jf)
..,o, ni,.6f 1
1 i e
. p.
-j.
- . ; {..
. i.
I I
t g ;-
p. -e.-l
. t..,
3 l
i.
- g7 j
' j.
'*}i
~!',}lg*
.1963-81 Re ld,o re y,
excloji.,3 39'7'/- 76 J p3 l,
,'sc
..0%
.- i j;
8 egg.
..j a
egf oe 78.
l1.
[j
,.t i.,-
c9 g e -
i i
4 g:
.j7f ii-g
- ll
...; l ;,
10 n..
i.,
. o,.c
. [ '\\..
... l. l i.
l og
-..~t~.
..s a.
4,.
....f.g
. 6..
.. [,
l.
N g: '{. -
.h.
i.:
l:
i s
pl l i ',
N e
l j g.
\\........
,75. '..
v l
i +
i i
l.g.,..
S4
- j. _'j
,r;
.i.. I
.1. j y
._j
. i t
I, ;i i.. i..
l.3 g g.
.. y..
i.j 1
..or
'.j... g.d2( ;.
',r.10 - I
. 07J~
_ l. i j. g l..l. d l.-
,fg l-
oa.c
..a c.
1 l lr l
.;;i C hc.n y
'n Fee i -.
. 7 Y., i :,.,
'.a l-L 1-
}
- I, ec'laceht Pel4ep,
% c Se nerdJ. Serv.sc<t. f i e'rq4 M j
.. pt!.r j..
i
( 3-yair. O i
3
.' i
.. l.. [ _.
. Il an ad C hag e5,)';
j i
un i
-l 4
t
+
l l
.N l
i ?
i i
3 3-i,.g l
~.
- I
- g
._ j...
,i e ;
[-
g i
..ff9
..i
- 1. -
.ll
.t
.i l.
i 1 l-I.
i
- l.
i j
i
.i
.j j;
i
. l e s l.
6 l
- i ~
t
.g I!'
.i l-I I
~
- l l l -l-l-j--l-j-f j - j-I lf j-l -l-l -l-j j i i l-li*,
8 i !
I i.l l Figure 6 i
i.q.---..,...
- ..g.
l l,
!.,i.i l 1fl
- Annual Change in General Service Electricity Sales
,t g
l, I.
versus g.
. Av1hi/A jCha p.
3-year Average Change in
.l..
- i ls
.I' IhdemrallSe
.! ' ! j.
ii.:
jes'1 15,le:
Real Price of General Service Electricity:
- a. !
- . it v
t i
f/cc fr 3
1963-81 Sales-Change: Price-Change Path, p;. j !..
M _.
Excluding 1974-76 i
I ~
L.
.I.
.t
. 1.
'n l.
. l'.
i
- j.,!.. l
.. l
,.g l :.
.1..
..... ;.. J. 3
....l j "
i.
'r 49
' 'g l 1
i.
g l '.. ]...,.,. j
, [ l --l !! f-i-l-).l sgo "It j'
l-
.l..i.l...
i.
- .! 4
.g.
lI;
's1 i
'73
- l.
li s
- g N
..oif.
i oss)
'.L
, ti
_s
- 7 y 3
- i
,f.,I 8
- eg;'ijii! ' i. +,.7d; g
.l..0I '
4..
I i
..i
.. i.
.t 36*/
I
_-i i i
.D.
llii i I
l g
1 l
i 3
6
)
l ; i.
l l j..
l.
i 74 j
'es a
i!li
{ j.i..
.li;4 i
.l,,.-
i 1
-.. j.
,.i.
5 ll; i
- . i I
.e
-ao oir
.or
' i.ozc : t!
l
..! ate
., s.l l. i.hg !
,,5 : < l.. ! g,! -
-l l
i llChqh. ir Ne gli l j,
i f lic c f r ee. % Andsi!;
i. ;.
!i '.i
. IW, I '/d f.coe l.-
%l. S@.i c '.e.. '
- torr i
- I (3
L etr. Sven M'e' d,;
.ll I.tj, An'no n.dhugt.5!)!
t.
r.
-,i us i
i i
i l,,
g
' 6 L J
l w
,s g
=
g l :
{-
a
..i.
t
..r
- e g-l g
m.
.,o, SD.
l !
i
.. i. l..
l l,. i Ld i
..a i
t l, g
e
(
i e
q l.,i.
...l i !
l !.. '.
l-
-l
. i i
i
,,i i
I j
p
- 1
.g -
-g j
e.
i
!i..
- g. j
_j..e.;
i j.! ;
. g j
- i.
. g..
i l
i i l
- . l...
_ _1._.l.
i. !..
t.!
. [.,1,
...l...i. :
i !
ilf Figure 7 i'
.i. '.
g l.
g..,..
.!l.j General Service Electricity Sales Crowth
- 1...
Illustrative long-Run Relation between i. _a i
i 1.
/)v3hdAliCita gA.
and Changes in Real Price of
. _l..:
eferkll5c
.l.ihq/r(cdg5.le:
- .vi a General Service Electricity, Assuming
... [. l. '
g i
iUcc
. 4.9 percent Annual Crowth in Personal Income. ; '
I L..
. {6_.J.
j..i.
- q. 7..
l.., j. _
7.._.
- g. g.
9.:
l
?
l i
.....i,
.S i
.1
'i
]".t
,n
.. l.
t.,.
i i
l
.. h.. [
.l*
h, ),1 l-
..i.
i 6
'6 f e.
.[
gg e
.go alt
.~
-l : ;
}
e l
- l.. : i. i.
j.
i 6
...l i
i
- g7 r
..i -
"n
'i i i !
'gg
..oif, j -
s s
- gs )
'L
,if s
syy
- r.yg
..*j
., 7 ;,$0
- i.. !,.
- i.. 63'
' 2
.i.
l l
.i
%; i. !... l. l..[ j. l i. '
'6 */
,.l
,,f Sc=;.p6 B j.f,.0 g(}) l i ;
I3 I
.} j.
- i i '.
i
..I l.g.
- t I
l' i !- i
- N I
i i
. lg,
. Ng,a l.
I *. } e.
i
. '.e N.
i i;
ji).'
!l i
.g. l l 4
+.I i
- i
- l. I i j i
a.
l
.i
.l:
[
j~...
'i i
.}o
.dir
.br i.o'1r : Il
..foir
.o s.l l. b io & !.l., A :' l.. Virl *_.
-i
' 'cksf.
k ir ke sl! !.
.i
'f.4 n p 4 i i Ilie c
,.c.
j E,1.jSc.<we.e.f.
. hY 7d.C. toe
.. !or:r ;
l.
i i
(3.x etr.. /bert.'c..o j
i.j u @rdu n.. Oiai3 di)i y
!!>.t j i
f
,lj.i* [,j l
i
.!-i.--
-l;.i
.o i
j
.1.
1
... oft) l I
i b
.l i !,...
I i
i.
- i
- l
- ! I
.'..'l i
i I
I i I
- t. i.
! l j g
- i i i
- i
.l l
.I i
i i
I i g
i ~
i '
I t-.
g-l e g.-
- i !
!i
S-2, p.5 Table 1 Index of Commercial Activity in the State of New Hampshire Year Value of Price Index of b
Commercial Deflator Real Commercial a
c Activities Activities 55 288 64 4 44.7 56 3 07 656 46.8 57 328 678 48.4 58 324 692 46.8 59 348 706 49.3 60 369 719 51.3 61 391 726 53.9 62 424 737 57.5 63 444 748 59.4 64 489 759 64.4 65 525 772 68.0 66 590 7 94 74.3 67 644 814 79.1 68 711 846 84.0 69 783 884 88.6 70 870 925 94.1 71 950 965 98.4 72 1023 1000 100.2 73 1243 1057 117.6 74 1384 1163 119.0 75 1463 1252 116.8 76 1683 1316 127.9 77 1898 1395 136.0 78 2169 1491 145.5 d
79 2427 1623 14 9.5 Notes to Table 1 a.
Equals the sum of Trade and Service components of New Hampshire GNP.
Source: Cross State Product of New England 1977-79, Federal Reserve Bank of Boston, undated.
b.
Deflator for Personal Consumption Expenditures component of U.S.
GNP; 1972 equals 1000. Source: Economic Report of the President, Government Printing Of fice, Washington, D.C., February 1982, p.236.
c.
Equals column 1 divided by column 2.
d.
Latest year available.
S-2, p.6 Table 2 i
Growth Rates of Commercial Activity and Real Personal Income in the State of New Hampshire: 1955-1981 4
(percent per year) 1 i
i Period Index of Real Personal b
Commercial Income a
l Activity j
1956-60
.028
. 03 1 l
1961-65
.056
.044 1
1966-70
.067
.051 1971-75
.043
.029 1971-73
.074
. 04 9 1974-75
.003
.002 4
1976-80 n.a.
. 0 58 1976-79
.062
.064 1977-81 n.a.
.049 2
1977-79
.052
.061 j
1980-81 n.a.
.03 0 Notes to Table 2 i
}
a.
From Table 1 i
b.
Sources: Prior to 1960, Historical Statistics of the United States, Part 1, U.S. Department of Commerce, Bureau of the Census, Washington 4
D.C.,1975, Series F297-348, p.244; 1960 to 1981, Survey of current Business, U.S. Department of Commerce, Bureau of Economic Analysis, i
Augu st 1979, July 1981, and April 1982. Deflation by the U.S. GNP j
implicit price deflator.
1 I
1 i
i E
4 4
=
9 l
BUB A
&4 04,64 36hl 09 b1 3 461 o t.st 3rst osst ss ss i
S*
o*F i
d*'fl!*G llU 0
7 v>3tigf f 5W2pt41 t o,s/ pod l 3 S 'l o's
- 1. c7s... n 3.. y,,
,e,n,n, -
S Ol*f64g$.y
- 'fI3 83uTPITng teiasaunnoa ut f*
Tl Kaisuaaul 1113T233aI3 Jo saInpayas a^11eaasnIII 8 aan3Td
~
S-3, p.1 Tablo 3 Alternative Estimates of Annual Changes in Constant 3
Dollar GNP Originating in Manuf acturing in New Hampshire Year l-33regate S ectoral Deflation
- Deflation 1970
.057
.064 71
.003
.001 72
. 03 1
.120 73
.060
.106 74
.113
. 04 0 75
.103
.095 76
.144
.142 i
77
.110
.109 l
78
.093
.099 l
79
.043
.08 1 80
.054"
.014d 81
~.028"
.050 Av erage
.015
.04 6 i
Notes and Sources 1
\\
a.
Total current dollar GNP originating in Manufacturing (Gross State j
Product of New Hampshire 1977-79, Federal Reserve Bank of Boston, undated) deflated by the U.S. Producer Price Index (Economic Report j
of the President-1982, Government Printing Office, Washington, D.C.,
January 1982).
1 b.
Constant dollar GNP originating in Manufacturing (Cross State Product of New England 1969-80, Federal Reserve Bank of Boston, undated),
j derived by applying U.S. sectoral deflators to current dollar values j
for each 2 digit SIC Manufacturing category.
Estimated by applying U.S. growth rates for the " Goods" component of c.
GNP to 1979 data for New Hampshire Manufacturing GNP.
d.
Estimated by applying U.S. real GNP growth in 1981.(.02) to the rela-
]
tion g
=.01 + 2.0 x (growth in U.S. real GNP); see Appendix A.
73 j
i l
i
~
S-3, p.2
\\
~-~-
g i
I t
AmunaI hjd 5
=
Np
,f 5
- i.
i i
l 4 8 I '
! i l i
.t(..
. i i
I f
e.g
- i i l
? ! !
I
<y
[i T Jlefual I.gn.
i-i
- Wi t 'J :
)
5:
! t
..____.._..g..
.t
'c,
- f. _
Penh,_f.[2 1.
N!
it
- 1
_9 2-r1 %
/ i it
- [ T.
- il.
I I.
\\n i-6 nn t
. ff
._....__ _.e05_.
8: I
,(:
5 4
L i!
l i'
IIi l
N
.I!
If
.E i
si t
i' A
y-i i
- _L k(
I
._14
_. -.1 9 f._ _... _. -. _.
t1a $~.
100-I 3!
3'! ' I e fe'
't
.,.'I:
. _. _,0 C..
A I-q I,.
- gg g.}.
4 g._
' I 't'
- -,-.-om%J. fee,-eeyws:m- -
N.._ _
_ P.aAJ._.4e.Jae v.s.7 ccJu.fici..
(psAf;e.%._ ___
. e.10..
E l
t
- g...
\\
l 1
Figure 9 Actual and Predicted Changes in Manufacturing Electricity Sales
S-3, p.3 l
t Table 4 Public Service of New Hampshire Sales to Utilities (within the Commercial and Service Category)
MWH Sal,es to l
Year Utilities 1965 62,300 1966 38,100 r
1967 41,600 l
'1968 52,332 i
1969 57,835 i
1970 70,986 l
1971 79,606 1972 79,268 1973 87,642 1974 79,938 1975 75,904 i
1976 67,368 l
1977 31,049 1978 34,705 1979 43,403 l
1980 52,682 l
1981 81,447 t
t i
(
l s
i Anrualc.bai.)c Figure 10 S-3 P 4 in C 45 Annual changes in Sales to g 4,,,,40
- "1"3 Commercial and Service Customers
[4 i
versu s 3-Year Average Changes in
,7#
.20 Real Price of Commercial e
'71
~
and Service Electricity e
- ' $7 73 I
t
..S$
'4r a
)
4 i
l
.. *I*
615 l
'7 s
'?t e
'71 i
~~~05
- ,79
) 8f b
i
'y y D
e l
'11
.=
_,', g
.,je
~. 05
.05
O f
changes m ree.1 t }a t.kvas e.sE3 presaa c2 ya.ae a.usea.s<.*
..as..
AwMuaIC,hanb43) i
.10 - -
i 1
.it..
.P
- l 1
I e
S-3. P.5 Figuro 11 Annual c.besp in Cj$
Relations between changes in dies.4r'es,43 Sales and Electricity Price for 5'
4 las for (Industrial) Commercial
& Service
'7#
'.2d Iq(,3-73 ftaldsay
[ ' '
3h
,73
?
'69 *
..I$
'Cl e
G Esclo)ed M4
- f om4s
.. 10 l97'7-fl RelAhio9
{71 (
- gg G
71
.. 01
'14
'77
'30 e
't1
^
l l
.If
- 10
~. of
.05 0
./5 c h nges in r ee. I eledrc.43 penass i 2 y s.u r u s, a.ss.*f
. 0f.
&WMua! Ch4Mg4J)
.10.
.It..
eP l
Table 5 Projected Electricity Sales and Prices - Low Seabrook Coat (Prices in 1981 dollars)
Resid ential General Service Industrial--------
Price Mfg.
C&S Retail Other Total Year Price Sales Price Sales Sales Sales Sales Sales Sales (C/kWh)
(GWH)
(C/kWh)
(GWH)
(C/kWh)
(GWH)
(GWH)
(GWH)
(GWH)
(GWH) 1981*
8.5 17 94 8.9 619 6.9 1258 5 68 4239 1382 5621 1982 8.4 1696 8.8 630 6.8 1231 569 4139 1326 5465 1983 8.6 1645 9.0 64 0 7.0 1254 584 4123 1322 5445 1984 9.7 1582 10.1 641 8.1 1279 595 4043 1218 5261 1985 8.6 1606 9.0 663 7.0 1342 620 4231 1247 5478 1986 8.6 1629 9.0 691 7.0 1397 639 4356 1273 5630 1987 8.3 1759 8.7 7 64 6.7 1500 68 2 4706 1375 6082 1988 7.9 1848 8.4 822 6.3 1556 7 03 4929 1385 6314 1989 7.5 1972 8.0 965 5.9 1683 742 5363 1467 6830 1990 7.2 2106 7.7 1057 5.6 1843 8 04 5810 1589 7399 1991 7.1 2223 7.6 1144 5.5 1982 865 6212 1702 7915 Annual Growth in Total Sales, 1982-91: 3.5%
a.
Actual values for 1981 un L
e 9.
Table 6 Projected Electricity Snleo cnd Prices - Mid-Ranga Seabrook Coct (Prices in 1981 dollars)
Residential General Service
-Industrial--------
Price Mfg.
G&S Retail Other Total Year Price Sales Price Sales Sales Sales Sales Sales Sales (c/kWh)
(GWH)
(C/kWh)
(GWH)
(c/kWh)
(GWH)
(GWH)
(GWH)
(GWR)
(GWH) 1981*
8.5 1794 8.9 619 6.9 1258 568 4239 1382 5621
~
1982 8.4 1696 8.8 630 6.8 1231 569 4139 1326 5465 1983 8.6 1645 9.0 64 0 7.0 1240 584 4123 1322 544.5 1984 8.4 1683 8.8 677 6.8 1349 613 4323 1381 5705 1985 10.7 1607 11.1 666 8.8 1361 620 4254 1312 5566 1986 11.3 1515 11.7 644 9.4 1344 622 4125 1221 5346 1987 11.7 1386 12.1 (05 9.8 1279 592
- 3861, 1137 4998 1988 14.2 1287 14.6 576 12.3 1213 55'8 3635' -
1013 4648 1989 14.2 1218 14.6 559 12.3 1210 5 64 3552 935 4486 1990 13.3 1190 13.7 561 11.4' 1241 58 0 3571 930 4502 1991 11.4 1306 11.8 631 9.5 1369 63 6 3942 1035 4977 Annual Growth in Total Sales, 1982-91: 1.2%
a.
Actual values for 1981 I
4 l-u 4
Table 7 Projected Electricity Snico end Prices -- High Seabrook Cort (Prices in 1981 dollars)
Residential General Service
Ind u s t r ia l-------
Price Mfg.
C&S Retail Other Total Year Price Sales Price Sales Sales Sales Sales Sales Sales (C/kWh)
(GWH)
(c/kWh)
(GWH)
(c/kWh)
(GWH)
(GWH)
(GWH)
(GWH)
(GWH) 1981*
8.5 1794 8.9 619 6.9 1258 568 4239 1382 5621 1982 8.4 1696 8.8 630 6.8 1231 569 4139 1326 5465 1983 8.6 1645 9.0 64 0 7.0 1254 584 4123 1322 5445 1984 8.4 1683 8.8 677 6.8 1394 613 4323 1381 5705 1985 12.2
~1522 12.6 632 10.6 1312 604 4070 1258-5329 1986 13.9 1328 14.3 567 12.3 1228 584 3707 1095 4802 1987 14.4 1135 14.8 499 12.8
. 1111 535 3280 996 4236 1988 20.4 97 6 20.8 422 18.8 1007 48$
2909 793 3702 1989 23.2 839 23.6 391 21.6 949 4 68 2646 662 3308 1990 22.3 763 22.7 3 54 20.7 913 457 2488 605 3093 1991 18.9 800 19.3 379 17.4 97 1 486 2632 64 2 3274 e
i Annual Growth in Total Sales, 1982-91
-5.3%
F Actual values for 1981 a.
.h w
Table 8 Projected Electricity Sales and Prices -- Mid-Rangs/Canect Sc: brook Cost (Prices in 1981 dollars)
Resid ential General Service Industrial-- -
Price Mfg.
G&S Retail Other Total Year Price Sales Price Sales Sales Sales Sales Sales Sales (c/kWh)
(GWH)
(C/kWh)
(GWH)
(c/kWh)
(GWH)
(GWH)
(GWH)
(GWH)
(GWH) 1981*
8.5 17 94 8.9 619 6.9 1258 568 4239 1382 5621 1982 8.4 1696 8.8 63 0 6.8 1231 569 4139 1326 5465 1983 8.6 1645 9.0 64 0 7.0 1240 584 4123 1322 5445 1984 8.4 1683 8.8 677 6.8 1349 613 4323 1381 5705 1985 10.7 1607 11.1 666 8.8 1361 620 4254 1312 5566 1986 12.2 1479 12.6 629 10.3 1319 558 3985 1180 5165 1987 12.3 1347 12.7 586 10.4 1249 550 3729 1105 4834 1988 11.8 1331 12.2 5 94 9.9 1238 569 3732 1105 4837 1989 11.1 1386 11.5 64 1 9.2 1331 610 3968 1177 5145 1990 10.5 14 91 10.9 7 08 8.6 1462 662 4324 1283 5608 1991 10.1 1601 10.5 779 8.2 1558 750 4718 14 00 6118 Annual Growth in Total Sales, 1982-91: 0.9%
a.
Actual values for 1981 i
US
.h.
l Y
U
,F
-s i
APPENDICES to the Testimony of Vince Taylor before the New Hampshire Public Utilities Commission Docket No. DE 81-381 October 8, 1982
A-1 Appendix A PROJECTING ELECTRIC LOAD GROWTH FOR PSNH:
ASSUMPTIONS AND METHODOLOGY Appendix A describes how the demand equations developed in the body of the testimony can be used to project growth in electrici,ty sales for PSNH. The underlying assumptions used ia 'he projections for the four Seabrook-cost cases are specified. The demand equations are then applied to an illustrative case. Quantitative assumptions and illustrative results are presented in tables at the end of the Appendix.
Assumptions Projected Basic Economic Environment The U.S. is assumed to recover gradually from the current slowdown, enjoy several years of prosperou, growth, and go through another normal e
l business cycle in the last half of the decade. Real GNP growth averages
}
2.7 percent per year for 1982-91, significantly better than the 2.3 percent average of the post-embargo period.
The superior performance of New Hampchire economy relative to the U.S. average during 1976-80 continues throughout the 1980's, but the margin of superiority declines somewhat over time.
To project New Hampshire economic performance from the assumed course of U.S. GNP, historical data for 1970-1980 were analyzed.
Using real (inflation-adjusted) growth rates, the following relations were found to hold:
(A-1) i(Mf g, US) =
. 04 + 2 i(GNP, US),
(A-2) s(Mf g, NH) = 1.2 g(Mf g, US), for 1970-75, 4
( A-2 b) g (Mf g, NH) =. 05 + 1. 0 i(Mf g, US), 1976-80 1
(A-3) g(PI, NH) =.027 +.35 5(Mfg, NH), where i stands for a real annual growth rate, and the abbreviations in paren-theses define whether it ref ers to GNP originating in Manufacturing (Mfg) or Personal Income (PI) for the United States (US) or New Hampshire (NH).*
Deflation of Personal Income was by application of the U.S. GNP deflator for both N.H. and the U.S.
4 A-2 Manufceturing grswth is baced cn Equnti:n 2b, which io much esro favorable to N.H. than 2a.
The constan?. growth term is assumed to decrease by.005 per year through 1986 and to remain at.025 thereafter. This implies continued strong manufacturing growth in New Hampshire, with the size of the sector increasing by 58 percent between 1981 and 1991 (average growth of 4.7 percent per year).
The average growth in real personal income is 4.4 percent per year, compared to an average of 4.2 percent for 1971-81, and 4.9 percent for 1977-81.
The year by year estiestes are presented in Table A-1.
To calculate the estimate of s(Mfg, NH) from the assumed course of $(GNP, US), Equation 1 was substituted in 2b to yield:
( A-2b ') d (Mf g, NH) =. 01 c& 2.0 s(GNP, US).
Prices of Fuel and Interchanged Power Real prices for fuel and net purchases of non-nuclear interchanged power are assumed to remain constant at the 1981 level (Table A-2) through 1985. Af ter 1985, real. oil prices are assumed to increase by 3 percent per year and real coal prices at 2 percent per year.*
The real price of existing nuclear-entitlement purchases is assumed to remain at the 1981 level (Table A-2). Real nuclear fuel prices are assumed constant at 0.7c per kWh.
Net sales of coal or nuclear power made surplus by Seabrook are assumed to be at 90 percent of the fuel price of oil-fired generation.**
Generating Capability Future generation capacity other than from planned nuclear additions was assumed equal to the prior 5-year averages (in 1981) of generation (by fuel type, including nuclear entitlements), modified to reflect the completion of Garvin and Eastman Falls Hydro projects and the conversion of the Schiller units to coal (assumed Jan.1,1984) (Table A-3).
The PSNH No purchases of interchanged power are required af ter 1985 in any case considered.
This is significantly more favorable to PSNH than would be the rate under present NEP00L policies, which would pay energy costs plus 35 percent of the difference between this cost and the cost of the fuel displaced.
A-3 present 7.5 MW interest in Millstone III was assumed to be retained (in-service date of 6/1/88).
Seabrook Performance Seabrook capacity factors are assumed to equal.45,
.50, and.55 in years 1 through 3 and.55 thereaf ter.*
This is between Rosen's esti-mated performance and the historical industry-wide average for all large units.
Lagged Response to Price Shif ts As discussed in the body of the testimony, there seems likely to exist lags not captured by the statistical analysis in adjustments of electricity intensity of new buildings and manufacturing. To account for these lags, the constant-growth terms of the demand equations for General Service (Equation 2) and Manufacturing (Equation 3a) are assumed to decrease linearly to zero over the next 5 years and to remain at zero thereafter. These adjustments lower the projected growth in total retail sales by 0.5 percent per year af ter 1985.
)
Methodology -- An Illustrative Example The use of the demand equations to project future electricity sales is described below. The illustrative case described diff ers somewhat in the underlying assumptions from any of the cases considered in the body of the testtsony, but the method of projection is identical.
I Trial Values of Required Generation To determine the production and purchase costs of power in a given year, the level of sales for the year must be known, but sales depend i
upon these costs (because they influence electricity prices). To get around this problem, a Trial Value of required generation is used to e
estimate production and purchase costs. This Trial Value is a "best guess," based on the previous year's value and anticipated cost trends.
PSNH has a 7.5 MW interest in Millstone III. For simplicity, the capacity factor for this small increment of generating capacity is assumed constant at.55.
i
A-4 The Trial Values of required generation are obtained sequentially, year by year. The entire estimation procedure is carried through for one year before choosing the Trial Value for the next year. Actual generation will generally dif f er from the Trial Value, introducing an error in the estimating procedure, but the errors tend to correct themselves in the next year. They do not create a systematic bias, but a random element that is relatively small compared to other uncertainties.*
Projected Production and Purchase Costs Using the Trial Value of required generation, production costs of electricity are calculated. The required generation is supplied from hydro, nuclear, coal, oil, and purchased generation, in that order. If i
more than sufficient generation is available, the highest-cost sources of generation are decreased until the required total is reached.
Fuel and purchase power costs are applied to the generation figures and O&M costs are added to obtain the total production costs.
Non-Production Costs Fon-production costs are estimated independently.** These comprise transmission and distribution O&M, administration, and capital-related costs (interest, dividends, taxes, depreciation, and insurance).
l Costs per KWH The cost per kWh of Prime Sales is calculated by dividing total cost (production plus non-production cost) by the Trial Value of Prime Sales.***
Table A-4 presents the data for this calculation for the illu-strative case.
If this model were placed on a computer, it would be desirable to i
eliminate the error by iterating until the Trial Value equalled the estimated value. The present calculations were done by hand, and the gain in accuracy did not justify the additional labor.
In the cases presented in the testimony, non-production costs were derived f rom the financial projections of PSNH and Rosen. Total revenue minus production cost equals non-production cost.
Equal to.938 times the Trial Value of required generation, allowing I
for losses and own use.
l l
l l
A-5 Price Changes by Customer Class The yearly change in cost (C per kWh) of. Prime Sales is calculated f rom Table A-4.
The change is assumed to apply equally to all customer classes -- that is, if Prime costs increase.2c per kWh, all customer rates are assumed to increase by.2c per kWh. Table A-5 shows the re-sulting projected average prices (in constant dollars) of electricity by customer class, together with the (logarithmic) year to year price changes.
Projected Retail Sales The price-change data (Table A-5) and economic growth data (Table A-1) are used with the sales estimating equations developed in the body of the testimony to project future sales. The equations *, data, and pro-jections are presented in Table A-6 through A-9.
Projected Prime Sales Sales for the retail classes of customer are added to yield i
Retail Prime Sales. The growth rate for Retail Sales is then used to estimate growth in Sales to Other Public Authorities and Firm Utility Sales.**
Adding these plus Street Lighting Sales, as estimated by PSNH, yields Projected total Prime Sales. The components and the totals are 1
presented in Table A-10.
Trial Values of Prime Sales are also shown.
Note the lack of any systematic error between Projected Sales and the Trial Values of Sales.
t The constant-growth term in the equation for General Service sales is assumed to decline at.004 (0.4 percent) per year and that for Manufacturing sales at.001 (0.1 percent) per year. This is similar to the assumption for the four Seabrook-cost cases, exc ept that in. the illustrative case the decline continues below =ero.
The PSNH projections for the three major customers within Other Public Authorities were accepted. The Retail Sales growth was applied to the remaining customers. Firm Utility Sales were adjusted downward to allow for the ef f ect of the ownership purchase of 50 PW of Seabrook by the New Hampshire Electric Co-op.
A-6 Table A-1 Projected Economic Environment Annual Real Growth Rates (percent)
Year 1982 83 84
- 85 86 87 88 89 90 91 Avg.
U.S. GNP" 1.0 3.0 5.0 4.0 3.0 1.0 0.0 3.0 4.0 3.0 2.7 U.S. P.I.
1.7 3.1 4.5 3.8 3.1 1.7 1.0 3.1 3.8 3.1 2.9 N.H. Mfg.
2.5 6.0 9.5 7.0 4.5 0.5
-1.5 5.5 6.5 4.5 4.7 N.H. P.I.
4.3 4.8 6.0 5.2 4.3 2.9 2.2 4.6 5.0 4.3 4.4 a.
Assumed b.
Calculated from the historical relation: 4 (PI,US) =.01 +.7 $ (GNP,US).
This series is presented for comparison with projected N.H. Personal Income. It is not used in the base case analysis.
i A-7 l
1 I
Table A-2 Cost of Fuel and Purchased Power - 1981*
Prime Sales
---Cost per kWh---
g enerated GWH or purchased sold Percent of d
~
Generated Power (C/kWh)
Prime Output Coal 1.74 1.81 1.93 29 Oil 2.19 5.64 6.02 37 t
Hydro
.357,
6 Subtotal - Generated 4.287 3.63 72 Subtotal - Sold 4.019 3.87 Purchased Power Nuclear Entitlements 627 2.55 2.72 10.5 Other Purchases 1,410 g
5.82 6.20 17.5 Purchases Exchange Sales (338)
Subtotal - Purchased 1,698 4.63 28 Subtotal - Sold 1,592 4.94 i
Total - Generated or Purchased 3.91 Total - Sold 4.17 100%
A-8 Table A-2 (continued)
Other Fuel Costs Used in Projections (1981 c/kWh)
Seabrook fuel cost
.7 Sales of Seabrook power 90% of the Fuel cost of oil-generated electricity Seabrook O&M Costs -- Mid and High Cost Cases" (Millions of 1981 dollars) 1985 86.
87 88 89 90 91 49.7 53.1 55.7 1 04.1 112.6 119.3 124.2 a.
Source: Statistical Supplement for the 1981 Annual Report, PSNH, pp.12 and 14.
b.
After losses estimated at 6.2% of generation.
Nuclear O&M costs are on a per year basis, including full year costs c.
for the years in which the Seabrook Units come on line. Source:
Richard Rosen testinony in this Docket.
A-9 l
Table A-3 Projected Generation Capability of PSNH, Excluding Planned Nuclear Additions (GWH)
Existing Nuclear Year Oil Coal Hydro Entitlements 1982 2320 1920 333 645 t
83 2193" 1920 348 645 84 1784 2760 348 645 85 1784 2760 348 645 l
86 1784 2760 348 645 l
87 1784 2760 348 645 l
4 88 1784 2760 348 645 I
89 1784 2760 348 645 90 1784 2760 348 645 i
91 1784 2760 348 645 i
a.
Assumes loss of 136 GWH due to shutdown of Schiller plants for conversion 7
to coal.
I b.
Gain of 840 GWH of coal generation (70 percent capacity factor) and loss of i
546 GWH of oil generation from conversion of Schiller plants (assumes shift to base lead).
i e
s l
I i
r
Table A-4 Projected Costs of Prime Sales - Illustrative Chse Fuel O&M Other Capital-Related Total Trial Value Cost per kWh l
Prime Sales
---(millions of 1981 dollars)-----------
(GWH) 1981 C Current C Year
=
1981" 234.4 19.7 67.4 110.9 432.4-5620 7.7 7.7 82 213.9 19.2 66.2 110.9 410.2 5500 7.5 8.1 83 217.1 18.5 66.5 110.9 413.0 5530 7.5 8.7 84 217.7 20.3 72.5 110.9 421.4 6046 7.0 8.8 b
85 99.4 25.6 68.6 399.0 592.6 5721 10.4 15.7 86 56.2 26.1 61.9 368.8 513.0 5159 9.9 14.6 d.
O 87 67.5 25.1 66.1 341.5 500.2 5515 9.1 14.4 88
-37.8 35.8 59.6 575.9F 633.5 4971 12.7 21.8 89
-38.0 36.9 61.3 527.5 587.7 5110 11.5 21.3 90
-45.6 37.9 61.3 483.6 5 36.6 5110 10.5 20.9 91
-18.3 37.9 67.4 444.9 531.5 5621 9.5 20.4 i
a.
Actual b.
Seabrook I enters commercial service on 1/1/85, at a cost to PSNH of $1308 million.
c.
Seabrook II enters commercial service on 1/1/88, at a cost to PSNH of $1476 million l
A-ll l
i Table A-5 Projoeted Electricity Prices (1981 C/kWh)
Residential General Service Industrial Year Price change Price change Price change 1981*
8.5 13.8 8.9 12.0 6.9 22.4 82 8.3
-2.4 8.7
-2.3 6.7
-2.9 83 8.3 0.0 8.7 0.0 6.7 0.0 84 7.8
-6.2 8.2
-5.9 6.7
-7.8 85 11.3 37.0 11.5 34.0 9.7 45.0 86 10.5
-7.0 10.7
-7.0 8.9
-9.0 87 10.0
-5.0 10.2
-5.0 8.4
-6.0 88 13.5 30.0 13.7 30.0 11.9 36.0 89 12.3
-9.0 12.5
-9.0 10.7
-11.0 90 11.3
-8.5 11.5
-8.0 9.7
-10.0 91 10.3
-9.0 10.5
-9.0 8.7
-11.0 a.
Actual (average) prices.
l l
A-12 i
Table A-6 Projected Residential Sales Sales Year p
p (3) s k
(GWH) r r
r r
1980
.019 81"
.138
.029
.032 1794 82
.0
.052
.007
.045 1715 83
.0
.046
.020
.026 1671 84
.06
.029
.023
.052 1760 85
.37
.103
.023
.080 1642 86
.07
.079
.022
.057 1551 87
.05
.083
.021
.062 1457 88
.30
.060
.021
.039 1401 89
.09
.053
.020
.033 1356 90
.085
.042
.019
.023 1325 91
.09
.088
.019
.107 1474 Estimating equation:
i =4
- 1.0 p (3) r r
r a.
Actual values for 1981.
b.
The correct value (Table A-5) in
.024.
The value listed was used inadvertently.
The error in growth for 1982 is
.8 percent.
c.
PSNH estimates of customer growth from 1982 Ten-Year Electric Load Forecast, January 1982.
A-13 4
4 Table A-7 J
Proiected General Service Sales i
Sales Year p
p (3) y c
c c
1980
.027 81*
.120
.049
.025 619 82
.0
.049
.049
.015 628 83
.0
.04-
.049
.019 640 84
.059
.027
.050
.084 696 I
85
.34
.34
.050
.041 676 86
.07
.07
.050
.021 661 87
.05
.05
.050
.028 643 88
.30
.30
.050
.019 631 1
89
.09
.09
.050
.016 621 90
.08
.08
.050
.023 607 91
.09
.09
.050
.116 682 Estimating equation:
$ =.019
.004 (Year - 1981) + y - 1.0 $c(3) c a.
Actual values for 1981 i
b.
Correct value is
.023.
See Note (b) to Table A-6.
]
A-14 i
l Table A-8 Projected Industrial - Manufacturing Sales
/\\
Year p
byg(3) 4 g
- IM gg 73 IM 1980
.102 81*
.224
.028 1258 82
.0
.109
.05
.048
.011 1244 83
.0
.075
.07
.052
.020 1269 84
.078
.036
.11
.C56
.107 1413 85
.45
.124
.09
.060
.003 1428 86
.09
.094
.07
.060
.010 1441 87
.06
.10
.03
.060
.015 1419 88
.36
.07
.01
.064
.006 1410 89
.11
.06
.07
.064
.029 1451 90
.10
.05
.09
.068
.046 1519 91
.11
.106
.07
.068
.129 1727 Estimating equations i
=.006
.001 (Year - 1981)
.5g
+.5g
.6p g(3)
+
g
,y g
a.
Actual values for 1981 b.
Correct value is
.029.
See Note (b) to Table A-6.
l l
A-15 Table A-9 i
Projected Industrial - Commercial and Service Sales IC IC IM til le Utilities Total 1980
.14 81"
.224
.001 486.7 81 568 82
.029
.117
.049
.034 470.4 83 554 83
.0
.065
.049
.018 478.9 90 569 84
.078
.036
.050
.120 540.0 90 630 85
.45
.124
.050
.039 519.3 89 624 86
.09
.094
.050
.009 514.6 87 602 87
.06
.10
.050
.015 506.9 62 568 88
.36
.07
.050
.015 514.5 30 545 89
.11
.06
.050
.025 527.5 30 557 90
.10
.05
.050
.035 546 30 576 91
.11
.106
.050
.191 661 30 693 Estimating Equation: $
= 1.7y - 1.0 p IC
- Estimating equation excludes C&S sales to Utilities, primarily to PSNH for construction of Seabrook. Pr)jected sales to Utilities reflect estimated levels of effort on Seabrook construction.
a.
Actual values for 1981.
b.
Correct value is
.029.
See Note b to Table A-6.
4
A-16 Table A-10 Projected Prime Sales (GWH)
~
Street Other Public Firm Total Trial Year Re tail Lighting Authorities Utility Projected Value 1981 4239 30 347 1004 5620 82 4141 30 317 987 5475 5500 83 4148 31 318 992 5490 5530 84 4500 32 344 1088 5961
,6046 85 4370 33 333 867" 5604 5721 86 4254 34 326 809 5424 5159 s
87 4087 35 318 707 5147 5515 88 3986 36 312 622 4957 4971 89 3986 37 312 622 4958 5110 90 4027 38 315 633 4967 5110 91 4578 38 349 765 5729 5621 a.
Major decline from 1985-88 reflects phased replacement of PSNH electricity by self-owned generation (Seabrook I) by N.H. Electric Electric Cooperative.
B-1 l
Appendix B l
Capital Cost of Seabrook II to PSNH Ratepayers for Rosen's Base-Case In his testimony in this docket, Richard Rosen presents a " base case" estimate of $1629 million for the PSNH share of Seabrook II.
Of this total, $559 million represents AFUDC accumulated at 12.5 percent per year.
Table B-1 presents an alternative estimate based on the same direct construction costs, interest rates on new debt, and return on common equity. In the alternative, costs of financing construction are calculated using projected rates for new borrowing and for the cost of equity (including taxes) rather than at 12.5 percent per year. This is the only diff erence.
The alternative estimate of $2413 million is the cost t hat would be borne by PSNE ratepayers in constructing Seabrook II under Rosen's cost a s sump tions. This is 48 percent greater than Rosen's estimate of $1629 million. The amount attributable to financing charges is $1348 million, 2.4 times the accounting value of AFL'DC in Rosen's estimate.
B-2 Table B-1 Capital Cost to Ratepayers of Seabrook II for Rosen's Base Casea Cumulative Capital Cost---- Cumulative Direct Debt &
Construction Cost Year Preferred Equity Total sE1981 91.3 66.6 158.5 123 1982 111.9 81.1 193.0 123 1983 165.2 119.7 284.5 166 1984 339.4 245.8 585.3 373 1985 569.5 412.4 981.9 597 1986 887.2 642.6 1529.8 865 1987 1225.0 887.2 2112.2 1039 July 1988 1399.5 1013.5 12412.91 1107d a.
Direct construction costs from testimony of Richard Rosen in this docket. Rosen's assumptions on financing cost were used: 18 percent on long-term debt and pref erred (58 percent of financing) and return on equity of 18.5 percent, Onplying a cost of 36.2 perc ent (including tax) on connon stock (42 percent of financing).
Interest, equity, and construction costs were calculated for each year. Financing to pay these costs was assumed to be divided between debt and equity at the 1981 ratio of 58:42.
Costs are in millions of mixed, current dollars and assume an inflation rate of 8 percent per year.
i I
j
\\
o 6
C-1 AppOndix C VINCENT D. TAYLOR PROFESSIONAL QUALIFICATIONS EDUCATION:
Bachelor of Science in Physics, California Institute of Technology,1958.
Doctor of Philosophy in Economics, Massachusetts Institute of Technology, 1964.
PROFESSIONAL EXPERIENCE:
Economics Department, Rand Corporation, Santa Monica, California, 1961-1969; consultant to Capital Researr 5, Inc., Los Angeles, on security selection, 1970-1973; senior staff member of Pan Heuristics, Los Angeles (a division of Sc-ience Application, Inc., La Jolla, California),
1974-1978; energy consultant to the Union of Concerned Scientists,1979; senior staff of the Union of Concerned Scientists, 1980-82; economic con sultant, 1982 to present.
PROFESSIONAL EXPERIENCE IN ENERGY-RELATED RESEARCH:
Beginning in 1974, my professional work has been exclusively on the economics of energy, with particular emphasis on the comparative economics of nuclear power and its alternatives. During the period 1974-1982, I performed research and wrote reports and articles on: the comparative economics of nuclear and coal generated electricity, forecasts of the future demand for energy, electricity, and nuclear electricity, the comparative economics of the use of uranium and plutonium on fuel for nuclear reactors, the economics of the uranium market, the economics of reprocessing of spent nuclear fuel, the comparative future energy potentials of nuclear power, synthetic fuels and improvements in the efficiency of energy use, the economic eff ects of the oil crisis, the contribution of electric utilities to oil consumption, and the economic implications of closing a nuclear power planc.
ENERCY-RELATED CONSULTING:
During the period 1974-1982, I have provided consulting services, research reports, or expert testi=ony for:
The United States Arms Control and Disarmament Ag enc y
- The Energy Research and Development Administration
- The California Energy Commission
- The Council on Environmental Quality
- The Nuclear Regulatory Commission
- The Pennsylvania Public Utility Commission
- The Vermont Public Service Board
- The Of fice of Technology Assessment 1
i I
J C-2 ENERGY-RELATED PUBLICATIONS:
The Uncertain Future of Nuclear Power, with Dennis Holliday, California Seminar on Arms Control and Foreign Policy, P.O. Box 925, Santa Monica, California, August 1975 i
Is Plutonium Really Necessary?
Pan Heuristics, Los Angeles,
Sept., 1976 (Revised)
The Myth of Uranium Scarcity, Pan Heuristics, Los Angeles, 4
April 25, 1977 How the U.S. Government Created the Uranium Crisis, Pan Heuristics, Los Angeles, June 1977 (Revised) l The Economics of Uran'1um and Plutonium, in " Moving Toward Life in a Nuclear Armed Crowd?", Minerva, Volume XV, Numbers 3 and 4 (combined issue), Autumn-Winter, 1977 Prepared Testimony of Dr. Vince Taylor in the Matter of GESMO, prepared for the California Energy Resources and Development Commission, Pan Heuristics, Los Angeles, March 4, 1977, Chapter A.
Energy:
The Easy Path, prepared for the U.S. Arms Control l
and Disarmament Agency, January, 1979 (published by the Union of Concerne'd Scientists, Cambridge, Mass.)
The Easy Path Energy Plan, Union of Concerned Scientists, Camoridge, Massachusetts, May, 1979
" Science and Subjectivity," Technology Review, Fe brua r y, 197 9.
"A Warning:
E. F. Schumacher on the Energy Crisis," MANAS September 3, 1980.
"The End of the Oil Age," The Ecologist, October-November, 1980.
Swords from Plowshares, with Albert Wohlstetter, et. al.,
University of Chicago Press, Chicago and London,1979.
Conservation, Equity, and Efficiency, testimony before the l
Vermont Public Service Board, Docket 4475, November 5, 1980 i
Electric Utilities:
The Transition from Oil, testimony before the Subcommittee on Oversight and Investigations of the Commit-tee on Interstate and Foreign Commerce Committee of the United States House of Representatives, December 9, 1980.
" Electric Utilities:
A Time of Transition," Environment, Volume 23, No. 4, May 1981.
Testimony on the Economic Costs of Closing Indian Point, testimony before tra Atomic Safety and Licensing Board of the Nuclear Regulatory Commission, Docket Nos. 50-247-SP and 50-286-SP (pending).
_~ _
l hf' f
0907 UNITED STATES T AMERICA NUCLEAR F.0'[ATORY CCM4ISSION 83 APR 13 M0:17 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD v
.,'U[c[
)
In the matter of
)
)
CONSOLIDATED EDISON CCMPANY T NEW YORK, INC.
)
Ibcket Nos.
(Indian Ebint, thit No. 2)
)
50-247 SP
)
50-286 SP PG4ER AtJTHORITY T THE STATE & NEW YORK
)
J-(Indian Ibint, thit No. 3)
)
12 April'1983
)
CERTIFICATE OF SERVICE I hereby certify that single" copies of TESTIMONY ON THE ECONOMIC COSTS OF CLOSING INDIAN POINT,. TESTIMONY OF VINCE TAYLOR was served upon the following by deposit in the U.S. mail, first class postage prepaid, this'12th day of April 1983, except where noted otherwise by asterisks.
"~
~
h
?
Steven C. Sholly Nunzio Palladino, Gairman Victor Gilinsky, Comissioner U.S. Nuclear Regulatory Comission U.S. Nuclear Regulatory'Comission -
Washington, D.C.
20555 m shington, D.C.
20555 John Ahearne, Comissioner lhomas Ibberts, Comissioner U.S. Nuclear Regulatory Comission U.S. Nuclear Regulatory Comission Washington, D.C.
20555 Washington, D.C.
20555 James Asselstine, Comissioner James P. Gleason, Esq., Gairman U.S. Nuclear Regulatory Commission Adminstrative Judge Washington, D.C.
20555 Atomic Safety and Licensing Board 513 Gilmoure Drive Silver Spring, MD 20901 x
y' 9
b "
Dr. Cbcar H. Paris James A. Laurenson Administrative Judge Administrative Judge Atomic Safety and Licensing Atomic Safety and Licensing Board Board U.S. Nuclear Regulatory U.S. Nuclear Regulatory Comissian Comission Washington, IC 20555 Washington, DC 20555 David Lewis, Esq.
Docketing and Service Section Atomic Safety and Licensing Board Office of the Secretary U.S. Nuclear Regklatory Commission U.S. Nuclear Regulatory Comission Washingto.
D.C.
20555 Washington, D.C.
20555
- Janice E. More, Esq.
Atomic Safety and Licensing Board Panel Donald F. Hassell, Esq.
U.S. Nuclear Regulatory Comission Henry J. McGurren, Esq.
Washington, D.C.
20555 Office of the Executive Iegal 1
Director Atomic Safety and Licensing Appeal U.S. Nuclear Regulatory Comission Board Panel Washington, D.C.
20555 U.S. Nuclear Regulatory Comission Washington, D.C.
20555
- Bernard Sanoff, Esq.
Assistant General Counsel Paul F. Colatulli, Esq.
Consolidated Edison Company of Joseph J. tevin, Jr., Esq.
New York, Inc.
Pamela S. Horowitz, Esq.
4 Irving Place Garles Nrgan, Jr., Esq.
New York, NY 10003 Morgan Associates, Gartered
)
1899 L Street, N.W.
- *Garles M. Pratt, Esq.
Washington, D.C.
20036 Stephen L. Baum, Esq.
Power Authority of the State myor F. Webster Pierce of New York Village of Buchanan 10 Colunbus Circle 236 Tate Avenue New York, NY =10019 Buchanan, NY 10511 Jonathon D. Feinberg Stanley B. Klimberg, Esq.
New York State Public Service General Counsel Commission New York State Ehergy Office
'lhree Dnpire State Plaza 2 Rockefeller State Plaza Albany, NY 12223 Albany, NY 12223 marles J. m ikish, E3q.
Marc L. Parris, Esq.
Litigation Division Eric 'Ihorsen, Esq.
'lhe Ebrt Authority of New County Attorney York and New Jersey County of Rockland Che Wrld Trade Center 11 New Hempstead Ibad New York, NY 10048 New City, NY 10956 lbnorable Ruth Messinger Alfred B. Del Bello Member of the Council of the Westchester County Executive City of New York Laurie Vetere, Esq.
District 34 148 Martine Avenue 3
City Hall Wite Plains, NY 10601 New York, NY 10007
W f' Ezra I. Bialik, Esq.
Andrew S. Roffe, Esq.
Steve [cipsiz, Esq.
thw York State Assembly Environmental Protection Bureau Albany, NY 12248 thw York State Attorney General's Office Ibnorable Richard L. Brodsky
'IW Nbeld Trade Center Member of the County Ingislature thw York, NY 10047 Westchester County County Office Building Ibnald Ibvidoff, Director Milte Plains, NY 10601 New York State Radiological Dnergency Preparedness Group Spence W. Perry, Esq.
Onpire State Plaza Office of General Counsel Towr Building, Room 1750 Federal Dnergency tunagement Agency Albany, NY 12237 500 C Street, S.W.
Washington, D.C.
20472 Ibvid !!. Pikus, Esq.
Richard P. Czaja, Esq.
Stemrt M. Glass, Esq.
Shea and Could Regional Counsel 330 Madison Avenue Federal Dnergency fbnagement Agency thw York, NY 10017 Doom 1349 26 Federal Plaza Ihyllis Rodriguez, Spokesperson New York, NY 10278 Parents Concerned About Indian Ebint P.O. Box 125 Garles A. Scheiner, Co-Gairperson Croton-on-Iludson, NY 10520 Westchester People's Action Coalition, Inc.
Richard M. Ibrtzman, Esq.
P.O. Box 488 Iorna Sal 2: nan White Plains, NY 10602 Friends of the Earth, Inc.
208 West 13th Street Alan Iatman, Esq.
thw York, NY 10011 44 Sunset Drive Croton-on-Hudson, NY 10520 Judith Kessler, Coordinator Rockland Citizens for Safe Ehergy Zipporah S. Fleisher 300 New Ilempstead Ibad W3st Branch Conservation Association New City, NY 10956 443 Buena Vista Road thw City, NY 10956 Renee Schwartz, Esq.
Paul Gessin, Esq.
Melvin Goldberg, Staff Attorney Iaurens R. Schwartz, Esq.
Joan Ibit, Project Director Margaret Oppel, Esq.
New York Public Interest Botein, Ibys, Sklar & lbrtzberg Research Group, Inc.
200 Park Avenue 9 Murray Street thw York, NY 10166 thw York, NY 10007 Ibvid B. Duboff Craig Kaplan, Esq.
Westchester Ibople's Action National Dnergency Civil Coalition, Inc.
Liberties Comittee 255 Grove Street 175 Fifth Avenue, Suite 712 Miite Plains, NY 10601 New York, NY 10010
l W
+
. Ms. Amanda Ibtterfield, Esq.
Jeffrey M. Bl'un, Esq.
Johnston & George, Attys-at-taw thw York University Law School 528 Iowa Avenue 423 vanderbilt Itall Iowa City, IA 52240 40 Washington Square South New York, NY 10012 Joan Miles Indian Ibint Coordinator Greater thw York Council on Ehergy New York City Audubon Society c/o D2an R. Corren, Director 71 West 23rd Street, Suite 1828 New York thiversity thw York, NY 10010 26 Stuyvesant Street New York, NY 10003 Ellyn R. Weiss, Esq.
William S. Jordan, III, Esq.
Steven C. Sholly Ibrmon and Weiss thion of Concerned Scientists 1725 I Street, N.W., Suite 506 1346 Connecticut Avenue, N.W., Suite 1101 Washington, D.C.
20006 Washirgton, D.C.
20036 5.
- Served by Messenger to Maryland Nat'l.
Bank Bldg., 7735 Old Georgetown Rd.
Rm. 10708, Bethesda, MD
- Served by Fed ' l. Express a
J