Loading...
HomeMy WebLinkAboutSusitna - Futile Quest for a Plan of Finance - 1986Lt/r i r`/-e Susitna Hydroelectric Project /,98"L FUTILE QUEST FOR A PLAN OF FINANCE N ALASKA two contemporary large- scale engineering projects are noteworthy. One is the development of North Slope oil resources, including construction of the trans -Alaska pipeline. This impressive achievement is the work of private com. panies and private capital. A second impor. tant contemporary megaproject in Alaska is one planned by a public agency of the state —the Susitna hydroelectric project. Planning activity on this project was re- cently suspended, and the current prospect for its revival is not good. Despite the un- happy fate of this project (or perhaps be- cause of it), the Susitna project is an interesting case study of public sector planning for a major infrastructure de- velopment project. This article considers only the finan- cial aspect of the Susitna project. As it happens, this is the critical dimension, because the failure to devise a workable and accegtable financing plan led to the project 's demise. In this paper I will review the history of financial planning for the project from 1982 (date of completion of the feasibility study) through March of 1986, when the project was put on the shelf. The objectives of this review are to explain why a workable plan of finance was so elusive and, more important, to seek insights from the history of the Susitna project that may benefit future planning for major energy projects, PROJECT DESCRIPTION The Susitna hydroelectric project was to include two dams along the Susitna River in the Talkeetna Mountains of sautf►eentral Alaska. When completed, the by Gordon S. Harrison project would have a combined installed capacity of 1620 megawatts and an aver- age annual energy yield estimated at 6200 gigawatt-hours. The Watana Dam, intended for oper- ation in 1996, was to be a rock structure 885 feet high and 4100 feet long, capable of generating 1020 megawatts. At this height, Watana would be the fifth highest embankment dam in the world, and the highest in North America, exceeding the Mica Creek embankment dam in British Columbia (794 feet) and the Oroville Dam in California (771 feet). The Watana reser- voir would extend upstream 48 miles; it would be 1 to 5 miles wide, and it would have a maximum depth of 680 feet. The Devil Canyon Dam, located 32 miles downstream from Watana, was scheduled to be operating by 2002. It was to be a double -curved concrete arch 645 feet high and 1500 feet long, capable of generating 600 megawatts. The dam's height would include it among the nine tallest arch dams in the world, including the Hoover Dam in Arizona (725 feet) and Inguiri in the Soviet Union (892 feet). The reservoir for Devil Canyon would be 26 miles long, 14 mile wide at its widest Point, and have a maximum depth of 550 feet. ALASKA POWER AUTHORITY In the United States, major public sector infrastructure projects are typically built, owned, and operated by quasi - independent public corporations. So it is in Alaska, where the Susitna project is under the jurisdiction of the Alaska Power Authority (APA). The APA is a public corporation governed by a board of direc. tors appointed by the governor of Alaska. It has its own professional staff but relies heavily on consulting firms to provide engineering and other technical expertise. APPROACHES TO STATE SUBSIDIZATION Large infrastructure projects that are developed by public corporations usually rely on the sale of revenue bonds for financing. Revenue bonds are debt issues (the interest on which is usually exempt from state and federal taxation) sold in the national capital markets that are secured by income generated by the proj. ect (road tolls, electricity sales, gate receipts, and other fees charged to users of the project). However, Susitna was such a large, expensive project that it could not be fi- nanced exclusively by conventional reve. nue bonds. Payment of interest and principal on revenue bonds sold to cover all project costs would result in an exorbi. tant price for electricity in the early years of the project. Therefore, it was always assumed that the State of Alaska would need to subsidize the project.' Two forms of state subsidy for Susitna were proposed during the course of project planning. One was referred to as state "equity" investment in the project. In this case, state appropriations would be used to pay some or all construction costs, and thereby reduce or eliminate altogether the requirement for borrowing. The second form of state subsidy was referred to as "rate stabilization." In this case, state ap- propriations would be used to help make Gordon S. Harrison was associate director of the Alaska Office of Manappernent and Budget, and a director of the Alaska Power Authority from 1983 to 1986. This paper was pnrwrlted to the conference on Global Infrastructure Projects, Alaska Pacific Univer- sity, Anchorage, July 8, 1986, sponsored by the International Federation of Institutes of Advanced Study. 22 The Northern Engines, Vol. 18, No. 2 and 3 payments of principal and interest on revenue bonds. Thus, state subsidy would be used to service debt rather than reduce the overall amount of debt. State subsidy to the project in the form of loans was among the financing mechanisms considered by project plan- ners, but loan alternatives were never fully developed and incorporated into financing plans for Susitna. The two main financing concepts of equity and rate stabilization can be illus- trated graphically. Line AE in Figure 1 represents the real wholesale price of electricity from a large, hypothetical hy- droelectric project that is financed entirely by debt. This line gradually slopes down- ward to point E because hydro projects are typically built to accommodate load growth (resulting in lower unit costs), and because of the effect of inflation on level nominal debt service. At point E, the initial debt is retired and the price of power thereafter is based on operation and main- tenance costs.' Line 8D in Figure 1 represents the pro- jection of wholesale electricity prices that would prevail without the hydro project. In the case of Susitna, this line represents the wholesale price of power in the Railbelt from gas- and coal-fired thermal plants. This projection assumes real price increases due to rising fossil• fuel prices and other costs of operations. Line BF in Figure t represents the whole- sale price of electricity from the hydro project with a combination of revenue bonds and state equity. In this case, the size of the state's equity investment reduces the amount of debt to that level which Produces an entry price of power from the hydro project equal to the price of power from the thermal alternative (point 8). In- creasingly larger equity investment in the project would further reduce the price of hydropower. if the project were entirely financed by cash grants from the state- 100 percent equity financing —the whole- sale cost of power would not have a debt service component, and it would represent EQUITY FINANCING Figure 1. Prov irh slats subsidy in the form of equity reduces the requirenhant for debt financing. In this figure the shaded use reprinsirm the amount of equity needed to make the whoiesele price of hydropower equal tO the wholess{e price of power from the thermal afternatiw RATE STABILIZATION Cost per kWh Thermd Alternative A p Rate stsbilisatioe C H o y subsidy _. E B 0 Years Figure 2. Providing hate subsidy through rate etshililation requires utilities to pay for hydropower along tM prejseted Curve of the thermal alternative until the crossover point C is reached. This fin&fwJne approeeh is more desirable then the equity approach from the .rate's point of vww. Me variable costs of operation and main- rate stabilization. At the crossover point Q tenance only (this scenario is not shown hydropower becomes cheaper than the in Figure 1). altemative, and no further subsidy is re - Figure 2 illustrates how rate stabiliza• quired (customers then pay along the line tion works. Here, state contributions to CE). An underlying assumption of this w the project do not reduce the amount of Mach is that customers will not be willing debt; rather, they reduce the price of to pay more than they would otherwise hydropower (line AC) to the level of the pay for electricity, notwithstanding future thermal alternative (line BC) until the two savings that the project will create. are the same (at point C). Customers will It is evident from the relative size of pay for electricity along the line BC, with the shaded area in these two figures that the state making up the difference through less state subsidy is involved with rate stabilization than with the equity ap- proach (on the basis of the general assump- tions underlying these curves). Also, it is no doubt evident that utility customers would prefer to pay along the line BF in Figure 1 than BCE in Figure 2. REAL AND NOMINAL DOLLARS Because of the long time involved in debt repayment, it is necessary to account for the effects of inflation when analyzing the cost of any major project. Thus, fi- The Northern Engineer, Vol. 18, No. 2 and 3 23 nance Planners and economists distinguish between real (or constant) dollars, which exclude inflation, and nominal dollars, which include the effects of inflation. In those terms, the cost of the Susitna project was estimated to be about S5 billion at prices prevailing in 1985 (real dollars), but more than S12 billion at the prices pre- vailing when the expenditures actually would be made (nominal dollars). REVIEW OF FINANCE PLANNING A review of finance planning for the Susitna project is best approached chrono- logically, beginning in 1982 when a major feasibility study was completed. 1982 In 1982 a feasibility study of the proj- ect was completed and three financing options proposed. During this time, how- ever, the long-term oil price outlook was deteriorating. Acres American report. In March 1982, the firm Acres American released a major feasibility study of the Susitna project. The firm had been under contract to the APA since late 1979. The Acres American report proposed the two -stage construc. tion schedule described above under "Project Description." This project con- figuration and the supporting analysis became the basis for APA's license appli- cation to the Federal Energy Regulatory Commission (FERC). With regard to financing, the Acres American report proposed three options: (1) 100 percent state appropriation of the total cost of construction, estimated to be S5.1 billion in 1982 dollars; (2) a state appropriation of $3 billion 0982 dollars), with the remaining project cost financed Table 1. Finance Plans for the Susitna Project with revenue bonds; or 13) a minimum state appropriation of S2.3 billion (1982 dollars) with the remaining project cost financed with revenue bonds. (The Acres American and other major financing proposals are summarized in Table 1.) It is noteworthy that one of the financing options was a cash grant from the state for the full cost of the project. At this time, it was widely presumed that Alaska's statewide hydroelectric development pro- gram would be funded entirely by state grants. The other two financing options are variations of the equity approach shown in Figure 1. An equity contribution of S3 bil. lion would represent an entry rate for the project somewhat below point B in the figure; an equity contribution of $2.3 bil- lion was calculated to represent an entry rate at point B (i.e., at a price equal to the Report' Constant S Construction Total costs (billionsl Nominal $ Construction Financing Total Revenue forecast Finance options Constant S (Same year as "Constant S" column) Acres American 5.1 15.3 0.0 16.3 Battelle 1. 100% state appropriation of total capital cost Feasibility Study (Jan. 1982 $I Reportz (S5.1 billion). Consistent with S825. (Mar. 1982) 15.3 1.6 16.9 2. State appropriation of S3 billion with residual bond financing. 15.3 1.7 17.0 3. Minimum state appropriation of S2.3 billion with residual bond financing. FERC License 5.1 15.3 2.0 17.3 Battelle State appropriation of $1.8 billion with residual Application (Jan. 1982 $) Report bond financing, ( Feb. 1983 ) Kentco Report 5.1 13.4 3.4 16.8 DOR mean State appropriation of $800 million in equity and for the Anchorage (1983 S) Sept. 1983 $778 million in rate stabilization. Remaining Chamber of Commerce financing requirements met by combination of REA (Jan. 1984) guaranteed loan and municipal bonds. APA Economic 5.4 11.8 5.2 17.0 DOR mean 1. State appropriation of $1.5 billion in equity and and Financial (Jan. 1983 $) Dec, 1983 $400 million in rate stabilization funds (RSFI, Up -date (Feb. 1984) 11 .8 4A 16.2 2. State appropriation of S1.7 billion In equity and $350 million in RSF, plus an REA-guaranteed loan of S1.5 billion, with residual bond financing, Draft FERC 5.4 12.7 7.8 20.5 DOR mean State appropriation of $220 million for rate stabil- License (Jan_ 1985 S) June 1985 ization, with revenue bonding of full project cost. Amendment (Nov. 1985) APA Draft Plan 5.4 12.7 7.8 20-5 Not slated State to provide $520 million for rate stabilization by of Finance (Jan. 1985 S) appropriation or pledging earnings from the Perma• (Jan. 20, 1986) nent Fund. $2 billion (nominal S) of project revenue bonds to be secured by Railbelt utilities. Residual bond financing issued by state and secured by Permanent Fund earnings. 'All reports arc available at the Alaska Power Authority, Anchorage. Alaska economic projections for estimating electricity, requirements for the Railbelt, Vol. 9, by S. Goldsmith and E. Porter, ISER, University of Alaska -Anchorage, Sept. 1982 report, Samiie Pacific Northwest Laboratories. 24 The Northern Engineer, Vol. 18, No. 2 and 3 price,of electricity from natural gas gener- ation at the time the project would begin operation). Char*49 nri'anue outlook. Worldwide crude oil prices had escalated dramatical. ly in the aftermath of the Iranian crisis of 1979. In February 1981, the contract Price for Alaska North Slope crude on the Gulf Coast had peaked at 536.90 per bar- rel, with experts predicting that prices would steadily increase into the distant future. Long-term revenue forecasts pre- pared in mid-1981, consequently, indicated that the State of Alaska would be phe- nomenally wealthy. The Acres American feasibility study referenced the long-term revenue forecast published by Battelle Pacific Northwest Laboratories as part of a major study of alternatives to the Su- sitna project. Table 2 and Figure 3 show this revenue forecast. Clearly, cash financ. ing of Susitna was a plausible option in 1981. In mid-1982, however, a dramatic de. cline occurred In the long-term revenue forecast, as indicated in Table 2 and Figure 3. Full cash financing for the Susitna proj. ect was no longer an obvious possibility, but some form of state subsidy remained dearly plausible. Because of the revised revenue outlook between 1981 and 1982, some disquieting commentary on the viability of the proj- ect began to appear. A report by Tussing and Erickson in September 1982, for example, argued that the oil prices of 1980 and 1981 were artificially high and could not continue to be tolerated in the marketplace; that lower oil prices nulli- fied most of the economic assumptions used to justify the Susitna project; and that, by implication, the stall would not be able to provide the cash grants neces- sary to finance the project? 1983 In 1983 an application for a federal license for the Susitna project was filed; it proposed two financing options. In spite of this, however, the APA initiated new financial and economic analyses for the project because of continuing declines in oil prices. FERC application. On 28 February 1983, an application was filed with FERC for a federal license to construct and operate the Susitna project. With regard to f!nan. Table 2. State of Alaska General Fund Revenue Forecasts, 1981 to 1986. (In $ millions, nominal.) Iftl 1982 1993 YEAR MARCH TUNE SEM DEC. MARO4 JUNE SEPT. DEC. 1915 1011 US7 3575 3367 2206 2911 3450 3514 3435 1916 9271 3979 4259 4261 3523 31" 35M 3192 3699 19: 10149 4569 4997 4217 3169 3393 3100 42AO 4172 19U 12179 4709 5147 Owl 4181 3540 3753 4106 4230 1919 13911 5242 5732 5379 4394 3534 M" 4442 4W 1990 15074 5141 5349 5096 4324 3651 4108 46% 5100 1991 16611 4717 4992 4549 4063 3374 3994 4290 4911 1992 17932 4696 4a66 4441 39U 3299 3913 4157 4863 1993 19395 4611 4679 4235 3971 3250 4103 4105 49% 1994 20326 4577 4652 4163 3990 3232 4173 4077 5DU 1995 20666 4261 4391 3892 3104 3092 3977 3927 4832 1996 20911 4033 4020 3609 3644 2930 3854 3612 4715 1997 20717 4246 4236 3786 3919 3001 4039 3741 4913 1996 20320 42% 4276 3937 3a19 3023 4129 3737 3110 1lt4 lfi3 1116 YEAR MARCH JUNE SEPT. DEC. MARCH JUNE SEPT. DEC MARCH 1985 3521 3340 345S 3343 32" 3253 3266 3290 3260 1926 37M 3475 35114 3402 3037 2964 2954 3215 2721 19t7 4042 Ml 395t U46 3001 2777 26M 2925 2D77 19111 4194 3957 4065 3241 2764 2470 2243 2474 1614 19a9 460 4149 43W 3290 2694 2403 2106 2397 1454 1990 49" 4175 4425 3245 2652 2324 2041 2310 1419 1991 4466 4237 4414 3204 2512 2259 1926 2299 1312 1992 4394 4431 4561 3212 2515 23U3 !950 2277 1232 1993 4521 45" 4639 3XI 2647 2337 1958 2327 1155 1994 45M 4592 44W 3162 2531 22U 1t62 2321 1096 19" 4510 4535 4479 3118 2445 2160 la34 2348 1045 19% 4516 4401 4465 3077 2313 2063 1753 =3 997 1997 4527 4248 4353 29" 22t2 2040 1750 2307 1093 19" 43M 4125 4327 2931 2257 1999 1750 2307 1061 Noe; The I"1 (aeecsc rea pegta ' by tlt Lud et of Said mad Eoosaatie Raaraek Uoivftsky of Alaeks for aut:e Natresat Labaaaarim dewed hat (uecau a(auoNr. aeveraaoe to and 11f'd9 iecaae by tit Atoka Depufesau of Revear, i• 1961. The lla2 -19" fixttass vwn ptepwd by Alataa 0(Ifce d M1UL1Vntwt Md lled9r; doll aea (Dacus of petmleate wvuva toe and tgeity utcoaa Lade by tha Alrka Dapaetaeu d R4v�te 7heaa (aeaess reprwst !a 50th pueatile Ptd4bdity est aeeeee (thee is a *4W Chao "as Baud va4e V a ba arcs er lea tba dtt toraeaild nhtL The Northern Engineer, Vol. 18, NM 2 and 3 25 Millions of Dollars 1 $12000 -r $10000 $8000 STATE OF ALASKA GENERAL FUND REVENUE FORECAST / 1981-1986 1981 (Battelle Report) $6000 March 1982 $4000 March 1984 " March 1983 $2000 March 1985 March 1986 _._....._...._.._........�. _. $0 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Fiscal Year sowce A4ika Daoanman• of Ra nu* aid OffK of kbn*9w"nt and eudgei Figure 3. Stan of Alaska General Fund revenue foneaar, 1981.19" cing, the license application stated that "costs for Watana through 1989 would be financed by $1.8 billion (1982 dollars) of state appropriations. Thereafter comple- tion of Watana is expected to be accom. plished by issuance of approximately S2.4 billion (1982 dollars) of revenue bonds." It also stated that the Devil Canyon phase would be financed entirely by revenue bonds. No doubt in response to revised revenue forecasts, the APA had dropped the full cash financing option, and recal- culated the minimum state cash contribu- tion to be $1.8 billion, or $500 million less than the minimum contribution of $2.3 billion identified earlier in the Acres American report. Concern about the future price of crude oil —the keystone of the project's economic and financial feasibility assessments —was thereupon expressed by FERC's staff. Noting several deficiencies in the state's application, FERC called for the APA to incorporate updated oil price forecasts in its economic and financial feasibility ittrdies. In response to this and to other critiques of the existing analysis, as well as to the changing oil price outlook gen- erally, the APA contracted with the firm Sherman H. Clark and Associates (SHCA) to provide updated forecasts. In the mean- time, a joint venture of two major engi- neering and construction firms, Harza Engineering and Ebasco Services (Harza- Ebasco), had been hired by the APA to provide engineering, design, and technical assistance in the FERC licensing process. Harza-Ebasco now initiated a review of the economic and financial studies for Susitna. 1984 During 1984 the financial dimension of the Susitna project began to receive serious attention from the APA, the legis- lature, and others. At mid -year, the long- term revenue outlook was robust enough to support an optimistic view that the project could be financed with the help of sizeable state grants. By the end of the year, however, it had become apparent to APA financial planners that a new approach was needed. 1964 Up -dare. In February 1984 the APA released the draft report Susfrne Hydro- electrk Pmwt Economic and Fkra 41 Up -date. Much of this report was the work of Harza-Ebasco; It incorporated the oil Price forecasts of Sherman H. Clark and Associates. The report validated the eco- nomic feasibility of the project, but con- tained a lengthy discussion of the major unresolved financing issues facingthe proj- ect. This report also introduced the subsidy mechanism of "rate stabilization." Several financing options were reviewed by the authors of the report, but two were advanced as the most feasible: (1) state appropriations of $1.5 billion for equity in the project, and S400 million for rate stabilization, with the remaining costs financed by revenue bonds; or, (2) state appropriations of $1.7 billion for equity and S350 million for rate stabilization, Plus a-$1.5 billion loan guaranteed by the U.S. Rural Electrification Administration (REA), with the remaining costs financed by tax-exempt revenue bonds (all figures in 1984 dollars). Thus, under these financing proposals, the state would not only pay a substantial portion of the project's construction costs; but would also create and finance a rate stabilization fund. This fund (as explained under "Approaches to State Subsidiza. tion") would then be used to offset enough debt service on the outstanding bonds to keep the project's wholes ila cost of power equal to the cost of the best thermal alternative until such time as 26 The Northern Engineer, Vol. 18, No. 2 and 3 the cost ofalternative power for the proj- ect surpassed the cost of hydropower (i.e., until the "crossover point" was reached). What characterizes the 1984 APA Up - dire is its somber assessment of the many conditions that would have to be met, and the public policy decisions and com- mitments that would have to be made, to finance the Susitna project successfully using multi -billion -dollar debt issues. Among these were the necessity for: (1) recognizing Susitna as one of the state's highest capital funding priorities; (2) pro- vidFttg adequate security for the very high volume of debt, which might require a con- stitutionally dedicated stream of revenue from the state's petroleum resources; (3) obtaining tax-exempt status for Susitna bonds; and (4) immediately providing for sizeable state appropriations to the Susit- na fund, as well as for the retention in the fund (by annual appropriation, if neces- sary) of the interest earned on that money. Kentco report. Also early in 1984, a re- port on the Susitna project was issued by the consulting firm of William Kent and Company (Kentco), which was working under a contract with the Anchorage Chamber of Commerce. This report, too, recommended a combination of state equity, a rate stabilization fund, and re- sidual revenue bond financing. The pro- posal, however, called for a larger rate stabilization fund ($778 million) and less equity (S800 million) than the 1984 Up- date. (These amounts are 1983 dollars.) The report further called for a majority of the debt to be guaranteed by RCA, with the remainder to be tax-exempt municipal debt. The Kentco report was optimistic in its treatment of the financing issue.' Ad- dressing the Anchorage Chamber of Com- merce, consultant William Kent stressed that his finance plan "allowed a minimum need for state investment, spread the need for state appropriation over a larger number of years, and did not present a tax exemption problem." The plan, he said, "suggests -a, need to start appropri- ating from 178 to 226 million dollars annually starting with this legislative session." Legislative action. During the 1984 legi. sladw session (January to June), two measures were enacted that dealt with Susitna financing: (1) the legislature ap- proved the Watana project at a cost of S3.75 billion in 1983 dollars; and (2) the legislature made a continuing appropria- Revenue outlook. Was it reasonable to ex- pect that $316 million a year (plus interest earnings of the fund) would be forthcom- ing from the legislature for six successive years to finance Susitna? In mid-1984, a plausible argument indeed could be made that the money was available, if the legi- slature had the will to see the project through. Note that the revenue projections shown in Table 2 and Figure 3 for 1984 are significantly higher than those made the previous year. If one were to project that the state's operating budget would grow at the rate of inflation (approximate- ly 5 percent) from a base of approximate- ly $2.2 billion in FY1984, then the 1984 revenue forecasts suggest that the State would have over $1 billion a year during FY 1986.1991 to allocate for the capital projects and loan programs. Under these fiscal circumstances, appropriations of S316 million per year to a fund retaining its own investment earnings certainly was not, on the face of it at least, impossible. By the end of 1984, however, revenue forecasts had fallen to their 1983 levels.' Also, additional oil price reductions seemed probable, due to a steady erosion of OPEC's influence over oil prices. There were other reasons, as well, to believe that the expectation of massive and continuing state appropriations for Susitna was unrealistic. Notably, the 1984 Legislature had appropriated only $100 million for Susitna for FY 1985, while total capital appropriations that session tion for "equity investment in and rate stabilization for the Susitna project" in the amounts of $100 million for fiscal year 1985 and $200 million for each of the six suoceeding fiscal years.` While the Watana construction cost figure of S3.75 billion was traceable to the Harza-Ebasco Up -date, the origin of the $1.3 billion (nominal dollars) total set aside by the continuing appropriation was a mystery. Many people assumed that it was based on the Kentco report and Wil• Iiam Kent's Chamber of Commerce speech. In any case, it bore no resemblance to the finance plans proposed in the Up -dare or those being discussed by the APA staff and board. Meanwhile, APA staff continued to maintain that some S2 billion (constant dollars) was needed from the state to help At its meeting of January 23, 1985, the finance the project. Thus, instead of the board of directors of the APA adopted a $200 million per year for FY1986.1991 staff recommendation for a Susitna plan appropriated by the legislature, $578 mil- of finance that called for state appropria. lion per year would be required —or at tions of $1.94 billion over the fiscal years least S316 million per year if interest 1985.1995 to a fund that would retain its could accumulate in the Susitna fund.$ interest earnings. This money would be used for both equity and rate stabilization. Minutes of the meeting show that the board considered this option the best pre- sented to date, and directed the staff to continue refining it. By this time, however, it was increas- ingly apparent to many people that if the project were to go forward, it would have to do so under a financing scheme that did not require such large state cash contributions. Among those recognizing this were high-level individuals in the par- ent companies of the Harza-Ebasco joint venture, who in January 1985 held infor- mal meetings with the Governor, APA executive staff, and board members to discuss a proposal for staging the con- struction of the Watana dam. Under this approach. Watana would be constructed in two phases (the first and the third Phase); the Devil Canyon dam would be the second, middle phase of the project. The virtue of developing the project in three phases instead of two was primarily financial. Three phases of construction would match more closely than two phases the growth of electricity demand in the Railbelt. As a consequence, there would be less unused capacity in the Wa- tans dam in the early years of project operation, and therefore a greater ability of utility customers to carry the burden of revenue bond financing. Thus, accord- ing to the staging proposal, all three Phases would be financed entirely by revenue bonds, with a comparatively modest state cash contribution remain - exceeded $1.2 billion, comprising the largest unrestricted general fund capital budget in the state's history. This was hardly a good indication of legislative will to sacrifice other capital projects in order to pay for Susitna. 1985 During 1985 the APA and its consul. tants redesigned the Susitna project in an effort to facilitate its financing. Toward the end of the year a team of financial experts initiated work on a definitive plan of finance based on the reconfigured proj- ect. Staging Proposal The Northern Engineer, Val. 1e, No. 2 and 3 27 Preparation of a financial plan, By late 1985 it was increasingly evident that the question of financing was critical for the Susitna project. in particular, financial advisers to the APA were concerned about the real -world problems of selling so much debt for a single project in the national market. These were the same individuals who had contributed the lengthy discus- sion of these problems to the 1984 Up- date. The task of marketing Susitna bonds was much more problematic now that ttm state equity contribution had been ew!UniLu altogether. Pressure also was coming from the legislature for the APA to produce a cred- ible plan of finance for Susitna. Finally, critics of the project, such as representa. tives of public interest advocacy groups ing necessary for rate stabilization only in the early years. In February the Board received a public presentation of the conceptual Proposal and authorized Harza-Ebasco to develop it further in an expeditious man- ner. At its meeting of May 3, 1985, the Board approved the staged approach, and directed the APA staff to begin preparing an amendment to the FERC license appli- cation that incorporated the reconfigured project. By October, APA staff and consultants lted prepared a comprehensive analysis of the economic and financial aspects of the new three-phase project. On the basis of assumptions about the cost of generating power from natural gas and coal (the next best alternatives to Susitna, the APA staff calculated that a rate stabilization fund adequate to keep the wholesale cost of Susitna power equal to its thermal competitor would require as little as $253 million (1985 dollars). During the 1985 legislative session the continuing appropriation to Susitna of 5200 million had been made, so there was already enough money in the bank to finance the project under this scheme (provided the interest on this money was allowed to accumulate in, or was annually appro- priated to, the fund). When the APA released its draft License Application Amendment in November, the estimate of state contributions to a rate stabilization fund had decreased further to $220 million (1985 dollars). The primary reason for these low estimates of rate sta- bilization was the assumption that without termination of the project two months Susitna large-scale coal plants would be later. required in the 1990s to meet Rail It energy demand, causing substantial rate Plan of finance. The draft finance plan increases. presented to the Board in January was built on the premise that very little state cash would be available, and that all proj. ect costs would therefore have to be covered through the sale of revenue bonds. Summed over time, these bonds would total more than S20 billion (nominal). The key question was whether the utilities and the state could successfully carry that much debt. To assess the debt capacity of the util- ities, the finance team calculated the max- imum annual revenue that the utilities could generate for debt service, using as a basis the assumption that the utilities' customers could tolerate a maximum rate increase of 3 percent (real) per year. Then, using a 25 percent estimate for the maxi- mum tolerable one-time rate hike that the Railbelt ratepayers could withstand in the event the project never operated, the and the environmental lobby, were openly asserting that the project was not finan. cially viable. They claimed that the bond market would not absorb so much debt for a single massive project intended for a comparatively small market area that was isolated from the power grid of the conti- nental United States. Late in 1985, APA's executive director assembled a team of financial advisers (in- cluding several bond underwriters, bond lawyers, and others) to begin preparing a definitive plan of finance for the project. 1986 The team of financial advisers charged with preparing a workable financing plan for the Susitna project presented a draft plan of finance to the board of directors on January 23, 1986. The revelations con- tained in this document led directly to the ... the f nance team concluded that only a commitment of the earnings of the Alaska P&manent Fund would suffice to secure the state's special revenue bonds." finance team estimated the maximum security that the Railbelt utilities could offer bondholders against the risk of the project's never being completed. the results of this analysis indicated that the upper limit of indebtedness for the utilities for the project was S2 billion (nominal). Thus, the State of Alaska would have to Issue special revenue bonds to cover the remaining project costs. The State of Alaska, however, could not ade. quately secure that amount of bonds, even with the pledge of its general obligation debt capacity. After reviewing all plausible alternative sources of security, conse. quently, the finance team concluded that only a commitment of the earnings of the Alaska Permanent Fund would suffice to secure the state's special revenue bonds. The financial team also concluded that, beyond issuing special revenue bonds and pledging the income from the Permanent Fund as security, the State of Alaska would also have to provide a rate stabiliza- tion fund of $520 million (1985 dollars; or $2.3 billion in nominal dollars) and an additional 5323 million 0985 dollars) pre- eollstructi0n licensing and development costs. The reason the rate stabilization requirement was higher than the APA estimate published in the draft FERC license amendment ($220 million, 1985 dollars) is that the draft finance plan sta. bilized rates to the level of a 3 percent (real) annual increase in retail electric rates, rather than to the somewhat higher level of electric rates estimated by the APA to result from the best thermal alternative. At its meeting of January 23, the APA board requested its executive director to submit the draft Susitna plan of finance to rigorous scrutiny by a major munici- pal bond underwriting firm, to test the validity of the finance team's findings. Under contract to APA, the firm of Pru- dential-Bache Securities then reviewed the analysis and conclusions of the plan, and concurred with them in a report dated March 21, 1986. Three days later, at its meeting of March 24, the APA board voted to withdraw the Susitna license applica- tion. REFLECTIONS ON THE QUEST FOR FINANCE: PROBLEMS WITH RATE STABILIZATION Even if a politically acceptable means of securing the state's Susitna revenue bonds had been found, it is doubtful that negotiations between the APA and the Rallbelt utilities would have been consum- 28 The Northern Enginaw, Vol, 18, No. 2 and 3 'mated Lander the finance plan advocated by the APA—that is, with rate stabiliza- tion providing the only vehicle of state subsidy. At the time the Susitna project col. lapsed, negotiations between the Railbelt utilities and the APA for conditional power sales contracts had been under way for some time, but they were still in very Preliminary stages. The underlying prob- lems of developing a contract that incor. porated a rate stabilization fund were therefore never fully identified nor con- fronted by the negotiators. Neither of the existing contracts be. tween APA and purchasers of power from its projects (the four -dam pool and Bradley Lake) incorporate rate stabilization. There- fore the following analysis of the rate sta. bilization approach is speculative insofar as the concept has yet to be implemented. Nonetheless, in the course of financial planning for the Susitna project, several seriously complicating features of rate stabilization emerged. Problems with Rate Stabilization There are two reasons for doubting that power sales contracts placing signifi- cant reliance on a rate stabilization fund could have been successfully negotiated between the APA and Railbelt utilities. The first has to do with the pervasive pub- lic opinion in the Railbelt region that the Susitna project was going to bring immed- iate rate relief, or at least stabilize electric energy prices at their then current level. The second is that probably neither the -T utilities nor the state would have been willing to expose themselves to the risks that rate stabilization entails. Public expectations. Financing for APA's other major hydroelectric projects, the four projects of the so-called "four -darn pool" and Bradley Lake, relies on state subsidy in the form of equity, In both cases, state cash appropriations to the projects cover approximately half of the cost of construction, with the remainder of project costs covered by borrowing! This financing assures customers of a wholesale cost of power that is comparable at the outset to the cost of power from thermal plants. 191ailbelt residents had come to expect the same of the Susitna project. The proj- ect, after all, had long been touted as the most economical source of Railbelt power available, and the best defense against sudden and dramatic rate increases likely to be caused by the expiration of existin "Thus, a definite risk existed that the Railbelt utilities might have to pay a substantial premium for Susitna power. Further, the potential magnitude of this premium was very great ...." have become very protracted and compli• cated indeed. Allocation of risk. Among the risks associ. ated with any major energy project, two are crucial: (1) the risk that the project will cost substantially more to build than assumed in feasibility studies; and (2) the risk that the price of competing energy sources will not perform as expected (i.e., will fail to increase, or not increase as rap- idly as thought).' Either eventuality will leave the project an overpriced producer in the market, at least in the near term, Somehow, then, these risks must be borne by the developer of an energy proj- ect or the purchasers of the power, or al. located between them. In the case of Susitna, contract negotia- tions between the APA and Railbelt utili- ties never progressed to the issue of the allocation of these risks. Nevertheless, the approach to project financing adopted by g The Northern Engineer, Vol. 18, No. 2 and 3 29 Figure 4. Risk of cart overrun Is the rams under either form of subsidy, but the risk of fossil fuel Prices being krww then -petted is less with equity then with rate rtabilization, FIV" 5. RISKS OF EQUITY FINANCING Coat per kwh — 7bermal sitk atire D / — Thermal aitarnstire Risk of cost overrun ' with price drop A— B Hydro cost with overrun Risk of price drop _ Hydro cost E 0 Years the magnitude of the total rate premium that Susitna customers might have to pay if those conservative but reasonable as- sumptions proved true. Exposure to the risk of declining al- ternative (fossil) fuel prices is significantly less under the equity financing approach than the rate stabilization approach. This is because the gap between the cost of Power from the hydro project and the cost of power from the thermal alternative will close sooner under equity financing. The differences between rate stabilization and equity in this respect are best explained graphically. Figures 4 and 5 illustrate that the risk of cost overruns are identical un- der both financing approaches, but that the risk of declining fossil fuel prices is greater under rate stabilization. The cross- over point C in Figure 4 occurs much sooner than the crossover point C in Fig- ure 5, thereby reducing the length of time consumers would have to pay a premium for hydropower in the early years, if an unexpected decline in fossil fuels should occur. F, asf the consumers' perspective, a sizeable equity contribution is the pre- ferred .wethod of providing state subsidy to an energy project such as Susitna, be- cause it minimizes risk and offers the pros- pect of rates lower than those that would otherwise prevail. From the state's per- spective, on the other hand, rate stabiliza- tion is the preferred approach because it minimizes the state subsidy. The experi- ence of the APA with the Susitna project suggests that to the extent it is relied upon exclusively, rate stabilization may simply not be viable, particularly when used for a sizeable project and particularly in a period of unstable fossil fuel prices. When state subsidy in the form of cash grants is made to a project, the money should be used to reduce the overall level of debt for the project, rather than reduce the debt service burden in the early years of operation with the aim of keeping wholesale electricity costr comparable to a long-term projection of the avoided costs from alternative generation sources. SUMMARY In the course of planning for the Susit- na project, three sources offinancingwere proposed: (1) state appropriations to cover some portion of construction costs (equity); (2) state appropriations to cover some portion of the debt service on reve- nue bonds during the early years of proj- ect operation (rate stabilization); and (3) revenue bonds. Planning for the SUsI" project began with the assumption that cash appropria- tions from the state's general fund would cover all project costs—i.e., a 100 percent equity approach. Later, it was proposed that a mix of state equity and revenue bonds be used to finance the project. Fol- lowing that, the concept of a rate stabili- zation fund was added to the combination of equity and revenue bonds (because rate stabilization tended to reduce the amount of required state equity). Finally, the equity component was eliminated alto- gether, and it was decided that financing for the Susitna project would be accomp- lished entirely by revenue bonds and a rate stabilization fund. This evolution of Susitna's financial planning was driven by the eroding out- look for state revenues and by uncertain evidence of legislative resolve regarding financial commitment to the project. From the beginning, it was recognized that Susitna would require a substantial subsidy from the state. Ultimately, an acceptable plan of finance for the project eluded the APA because the state did not have enough money to provide the subsidy the project needed. The end came because of the prob- lem of providing adequate security for the large volume of revenue bonds called for by the finance plan, and this problem stemmed from the state's inability to pro- vide equity investment in the project sufficient to reduce borrowing require- ments to levels that could be secured by the utilities through conventional power sales contracts. 30 The Northern Enel mm, Vol. 18, No. 2 end 3 RISKS OF RATE STABILIZATION FINANCING Coat — Thermal aiternative per kWh DThermal altemattwe Risk of cost overrun / with price drop rf Rate stabillsatioo C ' r Hydro cost with overrun B A —Hydro cost Rink of price drop E 0 Yean Even if a politically acceptable means had been devised to secure the Susitna project revenue bonds, it is unlikely that a workable contract could have been suc- cessfully negotiated between the Railbelt utilities and the APA that relied heavily on a rate stabilization fund of a fixed amount. There are two reasons for this evaluation of the situation: (1) utility customers in the Railbelt expected the Susitna project to protect them from re- tail electricity rate hikes, when in fact the rate stabilization approach assured them of rate hikes and would not result in sav- ings to customers for many years, and (2) the concept of rate stabilization entailed risks that neither the state nor the utilities would be willing to assume. From the point of view of public policy considera- tions, rate stabilization might be the pre- ferred approach to providing state subsidy to large energy projects because it mini. mizes state contribution. However, the experience of the Susitna project suggests that it is not practical. State subsidy for future hydroelectric projects (to the ex- tent it is necessary and available) should take ft form Of equity, i.e., it Mould re- duce the need for borrowing, NOTES 'State subsidy was regarded by many people as desirable from a public policy perspective. because it provides a means of distributing the stste's oil wealth to citizens. Other as - Pam of the issue of subsidy for the project are discussed in Gordon S. Harrison, "Science, Susitna and political decision making." The AfOrMenr Engit , 1984, Vol. 16, No. 3. 2 In the real world, a project would never be without debt, because major renewals and replacements of the turbine, generator, and nvimhyard equipment would have to be fi. nanced through the issue,of new debt. Arlon R. Tussing and Gregg K. Erickson, "Alaska Energy Planning Studies," Policy Analysis Paper No. 82.13, a review of three consultant studies submitted to Alaska state agencies In fiscal year 1982, November 18, 1982. See also, Richard Emerman, 'The Probable Effect of Lower State Revenue Fonutb on the Projection of Electricity Demand in the Railbelt," policy and analy- sis paper 82-10, Division of Policy and Development and Planning, Office of the Governor, September 21, 1982. See testi. many of Gregg Erickson on SB 25, SB 26, and SS 2" before House Finance Commit- tee on kuy 18, 1981 (minutes, p. 1325). The Kamm plan of finance was not realirtic, howevw, because of cutbacks in federal funds for REA. In any case, the Susitne Project would not have received favorable eonsidan, ion by that agency because most of tan power from the project would be sold to "urban coopertrtiv il, " which are amorded a low pdorlty in the distribution of REA funds. Sea 'Tranacript of Questions and An. r war Session" following address by U.S. Senator Td Stevens totter Thlrteanth Alaska State Legislature. February 1984. sThe continuing appropriation was declared unconstitutional on August 30, 1985, by the Alaska Superior Court. Figure 5. Risk of fossil fuel pricy being lower than expected is greeter under the rate robllnation subsidy It taken longer to reach the oeoasower point than with the is"try approach, Figure a. a See minutes of APA board meeting of Novem. ber 9, 1994. 7It should be noted that the 30th percentile, risk -adjusted forecasts developed by the Do. partment of Revenue wars even lower, sig- nificantly, then the mean probability fore. cast shown in Table 2. The 30th percentile forecast reflects a 70 percent probability that the estimate will be exceeded, and is used by the executive and legislative bran. ches for budgeting purposes. a In the use of the four -dam pool, the debt component is a state -funded long-term sub. sidized loan. In the case of Bradley Lake, which has just begun construction, the debt component will be project revenue bonds is- sued by the APA and secured by contracts with the utilities purchasing power from the project. aA third major risk is that the forecast demand for the output of the project will not mate- rialize. This was a major risk of the Susitna project, but one that was not taken seriously by Railbelt utility managers, who constantly chided the APA for its corxervativs estimates of load growth. Thus, it moms unlikely that allocation of this risk would have impeded contract negotiations with APA. 'oThe rswits of the analysis, in table format, were included in a package of material pre - Pared by APA staff end consultants and dis- tributed to the Board of Directors at the meeting of October 2, 1985. The table is ti- tled "SanshMty Anah/sh; " but has r*hher table number nor page number. Further, the table is not reproduced In the APA's draft F ERC Ikense amendment, although the gen- eral outcome of the sensitivity analysis is al- luded to In Exhibit D, p. D4-6, of the draft amendment. The Northern Engineer, Vol. 18, No. 2 and 3 31