RESTRICTED Report No. EC-53-a FILE COpy This ieport was prepared for use within the Bank. In making it available to others, the Bank assumes no responsibility to them for the accuracy or completeness of the information contained herein. INcTERNATIONAL BANK FOR RECONSTRUCTION AND DEVELOPMENT COST OF CAPITAL IN THE CHOICE BETWEEN HYDRO ;J.\ND THERMAL POWER January 17, 1957 Economic Staff Prepared by: Robert Sadove COST OF CAPITAL IN THE CHOICE BETl·U:;IN HYDRO AND THERMAL PO~'T?R . INTRODUCTION - 1. HYdro power installations are regarded in many quarters as one of the primary symbols of economic development. There are, of course, real economic advantages in hydro, the most general of which are distinctly lower operating costs including depreciation. Uydro power may also eliminate fuel imports and thus ease the pressure on the. balance of payments. In some cases, there are extensive multipurpose benefits 'tdth hydro. In some special cases, there may be significant advantage in the use of hydro for pealdng purposes or for stand-by reserve in an integrated electric system. 2. On the other hand, the advantages of thermal power are not negligible. The most important are the feasibility of expanding power output in relatively small stages, and in locations whi6h mini- mize transmission requirements, greater ability to maintain reason- able average total unit costs if the level of demand falls below exp_f;l~tations, shorter construction period, and lower and more easj.ly det~tmined capital reqpirements. These factors are particularly favorable to a developing economy where power requirements frequently do not justify the immediate installation of large scale generating facilities, where load factors may be relatively low, where demand is difficult to forecast, and where capital is scarce. I. STATEMENT OF '!HE ISSUE 3. The scarcity of capital is a particularly important factor in properly making the comparison of hydro and thermal projects because ,of the dominant role that the cost of capital usually plays in the cost of producing power, particularly in a hydro station. The cost of capital is the cost of obtaining money for investment purposes. This includes interest, and profits before income taxes. The cost of capital normally averages about 80 to 85% of the unit cost of power in a hydro station while in a steam plant only about 40% of the cost of generating power. The other factor of comparable importance in a thermal plant is fuel which is typically more than 40% of the total cost of power. 4. In considering whether a hydro plant is preferable from , the economic point of view to an alternative thermal plant in most Situations, the question to be answered is, in essence: Is .it. WQrth . spending the additional capital needed to build a hydro plant in • order to save the cost of fuel? The answer to this question depends to a large extent on how short capital is or how abundant other in- vestment 9Pportunities are in the country concerned. - 2 - 5. 'The savings in fuel costs, or more precisely in operating costs, can be stated as a percentage of the diffeJ."ence in total in- vestment betweep'hydro and thermal with the resultant percentage considered as a return 011 the additional capital invested in hydro. '!he charges made for pOlver, it shou14 be noted, do not enter this calculation. In £i ve of the Bank projects for vl:nch it was possible to make this calculation on the basis of readily available information the yields calculated in this manner range bettveen 6.5;~ and 13.9% To judge the adeauacy of the return requires a general assessm~nt of al ternati ve investment opportunities Hhich a.'llounts to the same as determining the "real" return or "real!l cost of capital, admittedl~r an elusive concept but, nonetheless, a relevant one~ 1'1hi1e it is not possible in most cases to determine the real cost of capital very precisely, it should not be impossible to establish in a given country which rates of return are lot·r, lvhich are adequate, and within what range reasonable doubt exists. As experience is accumulated in looking at this problem in these terms the basis for judgment can undoubtedly be strengthened. It is clear that this kind of compu- tation leads to conclusions 1..rhich cannot be reached by using the nominal cost of capital, hal-rever adequate nominal costs may be in judging profitability of alternative investments to the particular enterprise. 6. A survey of the nominal cost of capital invested in several government owned hydro projects which have been considered fOr Bank financing shows the average to be around 4t% and the total financial return to be only slightly higher - around 6'f~ (see Table 1). Such a nominal cost of capital is defined as the actual cost of 'bQrro~dng and, in some cases, either an imputed or a required rate of return on the e~ity capital. The rates actually charged for electricity in the projects included in the survey ordinarily yield a net finan- cial return as ind:i.cated in Table 1 which in most cases is slightly in excess of the nominal cost of capital as defined above. In the five cases where suf'ficient information is readily available to permit a detailed analysis of the costs of the alternative thermal plant, the nominal cost of capital used for both thermal and hydro in this comparison varied as shown in Table 3 fror.l 3.8% to 6.2;~. A similar survey of the nominal cost of capital invested in several privately owned hydro projectlD which have been considered for Bank financing sho'tvs the average to be around 9% and the total financial return to be only slightly higher - around IO~'~ (see Table 2). 7• The method usually applied consists of comparing thf:!. unit " cost of producing a. ktvh of pOl'ler in hydro and thermal plants all the basis of the nominal cost of capital to the enterprise. This is ecpivalent in its effect to calculating the return to the additional - 3 .. capital invested in hydro and selecting hydro when the return exceeds the nominal cost of capital. In the Bank I sexp eri enoe hydro has been rejected in favor of thermal when cost calculations,,,,lllade on the basis of nominal money costs, show a lower cost per ~1h for hydro only ~.,hen it 't-las virtually impossible for the public utility enter- prise to obtain the addi tiona! funds required for hydro. This was the case in the Verbundgesellschaft loan (Austria) where the Bank urged a substitution of thermal power for part of the proposed hydro development. The object of this memorandum is to suggest that the principle applied in substituting thermal for hydro should be exten- ded to cover cases where the additional capital for hydro can actually be obtained but only at the sacrifice of higher yielding investment opportuni ti es els et\fhere. 8. There is strong evidence that the nominal costs of capital usually employed in the atlalysis of governmentally oHned hydro projects do not represent the real cost of capital in the countries concerned. 'Ihis is sometimes the case also for privately owned enterprises. A. Differences in the nominal cost of capital to power enterprises within countries and among countries are based on such factors as the corporate finan- cial structure, incidence of taxes, g;overnment pricing policies luth respect to power and other fae tors unrelated to the real return on capital in the countries concerned. B. The nominal cost of capital, particularly to govern- ment owned public utili ty enterprises, is frequently belOl-l the co s t of borrotrling money generally in the countries concerned, an~ government interest rates are frequently not meaning'ful in view of the fact that the market for government bonds both internally and externally is exceedingly ·thin. In countries Hhere capi tal is scarce government funds are often obtained at subsidized rates. c. liiost important, in a number of cases tbere is evidence that the real return 011 a1 terna tive public investments is ,a.t levels substantially higher than the nominal cost of capi tal to the governm.ent (vlhether subsidized or not). 'll1e evidence is e'V'en more compelling if comparisons are made 'vi th the , private sector. - 4- A. Factors Influencing Coat of Cap! tal to PO'WGr EntE'lrprises 9. As s tated abov:a the nominal cos t of capital fior several government owned hydro projects considered for Dank fin.ancing has averaged around 4~%. In several private projects con~lidered for Bank financing, the average nominal cost is around 9%. Broadly speaking, this difference in nominal costs reflects th!9 terms on Nhich different types of enterprises can finance thems.elves accord- ing to 'tvhether they have access to goverrnnent funds or' not. These levels of nominal money cost may be compared t-d. th '~hepractices of the utili ty indus try in the TJ. S ., where the cos t of rl~ising c api tal is about 12% for private companies and where the cost of capital used in the analysis of !?ublic power projects is ~% 'fl./; in Brazil where private power companies are currently paying more than 11%; in Italy lvhere even some of the recent projects financed by the Bank pay from lO/~ tc;> 12%; in the U.K •. where government pe)'t-ler projects are analyzed on the basis of 3. .5~~ cost of capital. It seems clear that the difference between Bank financed projects al;ld these other cases, as 1vell as the difference auong these examplE-is themselves, bears very li ttJ.e relationship to the real cost of capi tal in the countries concerned. 10. These differences simply reflect for the most part the insti tutional arrangements under which capital is m~~de available to different public util~ty enterprises. In many (~ountries there is a great difference beuveen the financial structu3~e of private and government enterprise. In the U.S., for examp:le, t.he .govern- ment regulatory agencies lock lvi th disfavor tlp017\ a ;fina.ncial structure for priva te utili ties l'11hich includes much over 50;6 debt, H'hereas, j. t is common practic e among public po'tver projects to rely almost ex... elusively on debt. Oi ~t'1n the different rates of r,ieturn which are necessary in order to attract equity ca.pi tal and bo,rrowed funds, this produces a 5ubstan.tial difference between the nominal capi tal cost at which public ~1 tili ty enterprises can obtain funds as compared to private. Another factor contributing to the diff(~rence in the TIlis 61s% is calCUlated on the basis of the cl.lrrtent borrot-ring costs and us~s an allowance as recommended by the FPC for "taxes ;foregone". The differences between the nominal. cost of cap:;.tal in proJ,t1J.cts considered for Bank fine.ncing an(1 these nom:inal costs shown for U.S. power projects is largely compensated in its effects on total costs by the more conservative method of calculating uepreciation allow- ances employed by the Bank compared 'l'1i ththe method used for the U itS. data surveyed. In all instances being C01'rlpare~ the same def'ini tion of coat of capital is used,that is one which inclu,des both interest charges on debt and the imputed or required rate of return on the equity capital. - 5- nominal cost of capital available to private and governmel1t Ol'1ned utili ties is the incidence of taxes on business profits. In the U.S. for example, the corporate profits tax adds about 5 percentage points to the capital costs of private utility enterprises. Finally, the fact that governmentacan frequently borrow on more favorable terms than private corporations, tends to reduce the cost of money to government ot~ed utility enterprises. 11. '!he resul ting differences in the nominal cost of capital to public and priva.te utility enterprises is frequently justified on the grounds that goverrnnent olIDed utility enterpriseshave as one of their major objectives the provision of electric power at a low price. In fact, even private utility enterprises are frequently given special access to low cost capi tal in pur sui t of this same objective. 12. It is becoming increasingly recognized that the proVJ,.s~on of capital on special terms, vJhile consistent 't.n. th the pricing poligies adopted by many goverrunents in the electric power field, distorts the allocation of investment resources. This is particu- larly evident in the choice between hydro and therm':U pO'tver. In order to correct, at least partially for such distortions in the choice bett-Jeen hydro and thermal pOl-Jer in the U .8., the FPC recommends that in calculating the costs of capital to a government hydro project an allowance be made for tt taxes foregone". ~ren after this correct- ion is applied, the cost of capital to government projects in the U.8.i5 roughly half the cost of capital to private projects, as noted above. 1fuile it might be argued that no great harm is done ill the U .8. by the resulting tendency to undertake perhaps too many capital intensive projects in the public sector, such arguments are hardly applicable to those countries Hhere capi tal is scarce. B. R~:-l(. vance of Actual Borrowing Rates 13. Further evidence for the thesis that the nominal cost of capi tal to utility enterprises usually understates the real cost of capital, is found in the fact that in some cases it is clearly belov¥ the normal cost of borrowing money in the particular countries congerned. The follo~dng examples illustrate the fact that the scarcity of capital is not adequately reflected in the nominal costs of capital used in some of' the Bank's hydro projects. In two Austrian hydro power projects surveyed the average cost of capital paid by the government owned public utility enterprises is 4.1% and 5.5% whereas interest in that country' normally seems to be 9% to ~l%. - 6- In one of the Italian hydro projects celpi tal was obtained at a cost of 7.7% whereas the cost of obtaining funds in most utilities in that country, as was noted above, is oVl3r 10%. In the two Finnish projects that are government ovmed, capital costs only 6% whereas the official. borrowing rates in that c10untry have been betl"1een 7% and 9%, and even at such rates only li'mited funds have been available. For several Colombian hydro projects a, nominal cost of capital well below 6% was used despite the fact thclt normal y:i.e1ds on govermnent bonds in recent years had ranged abov(~ 10%, and very seldom below 6%. In the Baygorria (Uruguay) project, pomparative costs of hydro and thermal power have been made on the b:FlSis of a cost of capital of 6.2%. This rate is based on the present price of UTE and goverrJllent bonds in the Hontevideo "market" which consists largely of government pen- sion funds. In the l.:ords of the most recent economic report on Uruguay, "the capacity of the 'market,' to absorb government bonds is limited, ••• the resul thas been that funds available for public works have been small. n In most instances surveyed, it is clear that the cost of money to the govermnent and in turn to the public utili ty enterprise is hardly an accurate measure of the real return on investment in the particular countries involved. The government bond markets in many of these countries are very tbin and very Ii ttle in fact) is bor~owed at the in~erest rates which prevail. Cb Return on Alternative Investments 14. In several instances it is possible to estimate the "real" cost of capital by reference to the expected return on other public investment projects. Governments have heaV1J responsibilities for providing for investment in m.any different fields, and the real return must be the highest return among alternative uses. ~ncile yields are more difficult to calculate for much of the basic over- head investment for wr~ch governments assume responsibility, it is clear that the real return to the economy in such projects must be substantial. This has certainly been the case for a ~roup of non·- power projects financed by the Bank. It is estimated that the average return for several of them is at a minimum around 12% and more typically around 16%. Included are a port in Nicaragua which reports an estimated ret~rn of around 17%, another port in Syria with a return of 40%, and a Honduras highway project with a return of 12%. 15. Even where the return to the economy from undertaking additional public investment in a particular count~ is not high, one should still investigate the possibility of channelling into - 1 - the private sectors' the resources t-lhich t-lould be released if a thermal project were accepted. Representative returns in the private sector are not usually below' 15%. In the manufacturing projects financed by the Bank, investment yields are e~ected to reach well beyond 20% and may average (;'ver l5/~. Recent Italian industrial projects indi- ca te possibl e returns of 1.5% to 30;& qn investment, and actual earnings in recent years above 20%. Recent J'apanese industrial projects indicate a return on total investment of 10~ to 20%. Average indus- trial yields of around 14% have been reported in the past in Colombia. The average rate of return to total capital in various industries in the U. S., during 1954 was about 17;~. Included were such rates as 23% for electrical 1ilachinery in.dustries, 20% for chemicals and 31% for motor vehicles and parts. hThile the yields cited above refer to the financial return in the individual enterprises and not to the economic return to the economy at large it is unlikely that they would understate or overstate by a wide margin the average economic return on investment. In considering the minimum economic return on investment in the i~dustrial sector, the Staff Economic Committee concluded in II Economic Considerations Affecting Industrial Projects" that lI'tn th the scarcity of capital that exists in most underd6,reloped countries, any project that does not have an economic yield of at least 10% should be regarded with suspicion. Ttventy percent is probably a more reasonable miniroun;1." These figures are clearly on the high side for the purpose of this memorandum, since the risk factor in industrial projects is in general greater than in the utility field. They are, hOvJ'ever , given here as an indication of the order of magni- tude of the yields that should be obtainable if ne't'J' resources could be directed to, and absorbed by, industrial enterprises. III. APPLICATION OF PROPOSED ME1rIDD --------------------~---- -- 16. TaKing into account the scarcity of capi'/jal that exists in many countri-es borrowing from the Ba.nk, and the fact that the libmina1 cost (if capital available to public utility enterprise::; does not necessari~·w reflect the real cost of capital to the economy, a need clearly ebp.sts for a method of comparing hydro and thermal power costs which dOE:!~~ not rely exclusively on the nominal cost of capital to the en-t,erpri~ye. The method proposed here calls for an explicit judgment as to t,hc:! adequacy of the return on the adc1i tional capitlal to be invested in~~dro. 17. As a starting point for the appraisal of the adequacy of the yield on the additional in,vestment, the f{')llowing guide posts might be '. - 8- consj,dered. In appral.s~ng a hydro power proj ect, a return of below 6% on the additional investment in hydro should be considered inade- quate in most cases. ]I.. return between 6% and 9Y; ma.y be adequate in some cases; but this is still doubtful because in most coup.tries higher returns seem feasible in certain alternative uses where the risks are not substantially different. In such cases, in order to justify the additional investment in hydro a serious consideration of the thermal alternative is necess~J. Such consideration should include an examination of both the alternative investment opportuni- ties for the funds released by the less capital intensi11'e thermal power project as well as the preparation of more exact cost estimates for the thermal project itself. Where the return on the additional investm~nt is found in the preliminary analysis to be 10% to 12% a less car.'.eful consideration of the thermal alternative is necessary and only a very cursory examination of the thermal alternative need be made when the j~elds are above 12%. 18. These guide posts to a judgment of the adequacy of the return on additional investment are admittedly arbitrary. The appraisal of alternative investment opportunities in terms which will yield a meaningful answer for the choice bett-reen hydro and thermal power in specific countries must be approached on a case to case basis. The proposed method is easiest to apply in the case where there are specific alternative projects in direct competition for the funds involved. The economic yield on the investment in these alternative projects can frequently be calculated and directly compared with the yield on the acldi t~onal investment in hydro. 19. ~fhen the specific alternatives in competition for capital resources are not known in sufficient detail it is still possible to form a judgment of the adequacy of the yield on the additional capital invested in hydro even though such a judgment can hardly be precise II In many of these countries the Bank has a judgment on the burden which high priori ty publ~~ investment is imposing on available investment resources. These judgments, in each case, could be translated into a range of acceptable yields. In some of the countries the interest rates at which go'vernments can borrow, plus an additional premium to cover the risks associated with specific projects, might be consider- ed as indicative of alternative yields. This method is of limited usefulness, however, whenever the market for government bonds is very thin and is not at all applicable where the government is strictly ra tioning its limited investment funds. 20. Where the availability of capital to the public sector leaves a trrl.de margin for exploiting investment opportunities i.t may be appropriate to evaluate the return on the additional investment in - 9 - hydro in terms of the yields which can be earned in the private sector §ince the possibility of channelling additional investment resources into the private sector should in such a case normally be considered. Yields in the private sector would, however, have to -be discounted for the higher risks that may be associated 'tri. th such investment and in countries with a history of inflation such yields may also reflect upward price movements. 21. Appraising the adequacy of the yield on the additional capital invested in hydro in tenns of the yield on alternative invest- ments poses some real problems when the yields on specific alternatives are not known. Unfortunately, it is exactly when the application of the propos ed method encounters the grea test difficulty that the nominal cost of capital to the public utility enterprise is likely to be least representative of the real cost of capital. Thus, an explicit judg- ment as to the adequacy of the return on the addi tiona! investment, wrdle admittedly more arbitrary than the use of nominal costs of capi tal, should be a sounder guide to the choice between hydro and thermal power. IV. ILLUSTRATIVE CALCULATIONS 22. Usir..g the information in Bank's reports there are five cases for which it is readily possible to calculate the return to the additional investment in hydro. They are listed in Table 3 and include Asvran, Li tani, Yanhee, Baygorria and Kariba. 23. In each case, the computation made on the basis of the nominal cost of capital resulted in a cost per ~lh which was lower than it would have been with a thermal alternative. ?or example, in the Li tam with a nominal cost of capital of .5.3%, hydro p01;oJ'er costs 9.7 mills per kwh compared to 11.5 mills for thermal. Similarly, in the Baygorria project at a nominal cost of capital of 6.2% hydro power costs 8.3 mills per kwh compared to a cost of 11.9 mills per kwh for thermal pO"VJ'er. On the other hand, the additional capital requir- ed amounted to substantial investments: $22 million in the case of Litani, $30 million in the case of Baygorria, and $98 million in the case of Kariba. In all cases except Kariba where the additional investment yields 13.9%, the yields on the additional ~nvestment is in o -10 - the neighborhood of 6.5~ to 8.8% 1/, that is within the range called ndoubtful" in paragraph 17 above.- It is suggested here that where similar calculations made at an early stage in the consideration of a hydro project reveal a return on the additional investment falling in the .. doubtful" category, it is appropriate to investigate whether the amount of additional capital could not be invested elsewhere at a higher rate of return. 24. The implications of the proposed method ca.n pe '(illustrated by a series of hypothetical calculations applied to some representative cost situation as shol-m in Table No.4. Assuming a fuel price of ,50 cents per million BTU which is the typical cost to the economy of fuel used in pot-rer plants in most countries, (see Tables' 5 and 6) a load factor of 70% and an investment of ~P.15 per kw of installed capacity in the thermal alternative: (for comparative cost data see Tables 8 and J.9). Hydro projects with capital inve'stment requirements above ~~c22 per kw of installed capacity including transmission costs, are out of the question in most cases as they yield less than 6% on the additional investment. HYdro projects ~vith capital requirements beUleen $475 and $622 are doubtful investments as they yield between ~ and 97b on the additional capital. Hydro projects In th investment re- quirements of ~~425 to ~466 yield 10% to 12% and are thus less doubtful but require careful consideration of the thennal alternative. Hydro projects with os-pi tal requirements of less than ~~425 yield more than 12% on the additional capital and are thus most likely sound invest- men~s and require only a cursory examination of t?e thermalalternative~ The \\effect of different assumptions regarding fuel cos ts are brought out 'Jin Table No.4. The effect of different load factors are indi- cat~d in Table 13 which shoH's that relatively high load factors are advantageous to hydro, whereas the low load factors which are common :in The calculation of the return on the investment in pO't-1er in the case of the High ASt-ran Dam was based on what was inevitably an arbitrary division qf costs bett<1cen pON'er and the other parts of the. project; wh~~ti--;::Oi:,her benefits are taken into aOocp~nt, the return on the total investment i;n the project is of c9urse higher. Similarly, the calculation,s given in the text in the case of Litani and Yanhee. do not take into consideration the possible bene!i ts from mw. tipurpose aspects of those projects, and, in the case of Kariba" tbe additional investment require- ments for the production of dbmestic fuel which ~·rould be used if a thermal alternative were chosen. - 11 - underdeveloped countries are advantageous to thermal. riOl-1er. v. OTHER ECONor1IC FACTORS IN THE CONPARISONOF HYDRO AND THERI'1AL 25. The return on the additional ca,pi tal invested in hydro has been formulated above in terms of the savings in production costs. This assumes that the amount of power to be produced is equivalent and that the hydro project does not produce any benefits other than "power. In fact, even though the return to the addi ti0nal capital invested in hydro is low does not mean that projects cannot be justi- fied on the basis of addi tiona]. benefits which accrue from mul tipUl~ose programs. Where non-power belaefits are significant, it is iIl1;?;1ortant to determine the return based on the full benefits resulting from the hydro pro j ect. This was done:, for example, in the case of ~he High Aswan p r o j e c t . ' 26$ The method of calcul~tting the return on the additional capi tal is not substantially a), tered by the necessity of taking into consideration mul tipurpose ben~~fi ts such as those which accrue from , irrigation works. The calcul~!/t.ion in this case l-Tould indicate the yield, or rather the sum of thEI savings in operating costs plus the net benefits from irrigation aSI a percent of the difference between the total investment in all phc:~ses of the multipurpose project and the investment required for a ~\hermal alternative. 21.. In eval.uating the saVings in production costs secured by the hydro al ternative, it is ir.1portant to evalucl.te fuel costs at t~eir cost to the economy ,of tln.e country and not their cost to the particular enterprise if these}' t~vo are different.· In some cases there are special taxes, tar:t.f:fs and monopoly prices on fuel, the effects of which must be elim:i~nated. Furthermore, in sorne instances fuel costs expected to prevai;1. in the long run"may differ' substan- tially from present prices. ~'Jhen this is the case some assumption must be attempted as to what the long ruri price is likely to be. If fuel is to be supplied from local sources, as in the' case of Kariba, which incidentally yields a relatively high return on the additional capital petore any adjustments are made, and if the capacity for fuel product.~i:On and distribution would have to be substantially increased if a thdrmal alternative were adopted, it may be very important to ,appraise the real cost of future fuel supplies. 28. The importa~ce of imported fuel to at least half of the thermal projects financed by the Bank, suggests that the foreign exchange sa.vings due to hydro may be an important balance of payment - 12 - . consideration. This can be taken into account in computing the investment ;}rieJ.d by placing an assumed premium on foreign exchange savings. Assuming all fuel to be imported, the result of a series of calculations for a particular set of examples based on the same assumptions as in Table 4, indica ted that vIi th an average price for fuel of abou t 50 cents per million BTU which is typical for imported fuel oil, hydro investment of ~p375 per kw of installed capacity would yield the same return on the additional capital required, as an in- vestment of ~~!.a.O if a 10% premium is assumed on foreign currency savings. Similarly, if a 30~& premium is assumed hydro could cost ~~425 and yield the same return and wi th a 60% premium, could cost ~~4.5o (see Table 1.5). In other words, a proper eValuation of the foreign exchange savings which result from hydro may 1'1el1 al ter an original determination that thermal was more economical than hydro. This is of some importance since many of the important hydro projects considered by the Bank involve capital costs above $400 per 1m. 29. This adjustment for foreign exchange savings does not, however, apply with equal force to all unrterdeveloped countries for the follow~ng reasons: The above calculations 'trJ'ere based on the assumption that the foreign currency share of hydro capital investment is 40% and of thermal 50i~, which is similar to the Bank's average experience (see Tables 20 and 21). If the assumptions are char..ged slightly to be more nearly representative of the situation in many underdeveloped areas where the foreign currency share of power investment is higher, that is,for example, 6.5% for hydro and 10% for thermal then the effect of the foreign exchange premium is almost nil (soe Table 16). This results from the fact that with a higher imported component in invest- ment, the net effect of the foreign exchange savings on fuel is sub- stantially offset. 1/ Even a 100% premium on foreign exchange savings would normally result in less than a one percentage point change in the return to the additional capital invested in hydro. In other words, a project wi th hydro costs of ~~50o per kw that was doubtful, would still be doubtful no matter what premium is attached to the foreign exchange pavings. The foreign exchange benefits of This assumes that the foreign exchange premium to be applied is no greater during the operating stage of the project than in the investment stage. -13 - hydro may p,asily be exaggerated for many of the underdeveloped countries. y 30. All of the above illustrative calculations assume that power output is expanded at the same rate whether the choice be in favor of hydro or of thermal, This is no t necessarily "bhe most economically beneficial approach. The construction of thermal stations may be phased to more nearly coincide wi th the expected increase in power demand. If such is the case then not only is the investment more fully utilized at all times but a significant part of total investment is postponed to a time when the demands upon scarce capital resources may be less pressing in relation to their availability•• VI. RH,A1iON m ELECTRIC P01nJER. RA1'ES 3.1. As noted above the determination of the return to the add- i tional inves tment in hydro is unrela ted to the pOlrTer rates charged to consumers. The savings in oper~ting costs in a hydro project and the additional capital costs are the same, whatever the charges made for power. The :i.ssue of higher or lO~'1er rates, though obviously important, is not directly relevant to this particular argument, namely, whether it is better to build a hydro or a thGrmal plant. However, an incidental efiect of the policy proposed in this paper might be some- l'1hat higher rates for e1 ectricity unless offset by other factors such as a change in regulatory practices. This results from the fact that rates are usually based on nominal costs whereas it is recommend- ed here that the choice between hydro and thertrJ.l be ba.sed on com- paring real economic costs. As has been indicated above, thermal costs per kwh may be higher on a nominal cost basis but lower on a real cost basis. In other words, prices charged for electric power, to the extent they are influenced by the nominal cost calculation for a particular project, would be higher in such a case if the thermal project were accepted. !I The recent Suez cr~sJ.s hus once more brought home the strategic or defense value of domestic sources of energy. This is often adduced as an argUment in favor of hydro rather than thermal power based on imported fuel. Quite legitimate in many cases, this argument is not an economic but a political one and should be recognized as such. '-14 - 32. Another aspect of the problem of electric power rates may be noted. TIle Bank bas become increasingly concerned ~dth the problem of electric power rates and is attempting, in order to build up funds needed for reinvestment and to attract new funds, to secure in many instances rate adjustments which, if applied, would raise the expected financial return on the total investment considered for Bank financing.. If as a matter of policy goverrnnents accept higher financial returns on investment in the power field as normal, capital may no longer be made available on such favorable terms to utility enterprises. In such a si tua tion, the nominal cost of capital may be more nearly representative of the real cost of capital~ lhis is, however, rarely the case at present. Table No. 1. The re.turn on total capital in various government owned hydroelectric projects. 2. The return on total capital in various privately owned hydroelectric projects. 3. The return to the additional capital invested in hydro in various projects considered for IBRD financing. 4. Levels of capital investment for hydt:o installations yielding various returns. 5. Price of imported fuel oil for various ports of the world. 6. Cost of coal in various representative countries. 7. Construction costs and operating data for various private hydroelectric plants in the U.S. 8. Construction costs and operating data for ,rarious private steam plants in the U.S. 9. Investment and operating data for various government con- structed hYdroelectric projects in the U.S. 10. The return to capital in various non-power projects financed by !BRD. 11. Representative rates of return in variouS United States manufacturing industries. 12. Illustration of the calculation of the return to the additional capital method. 13. Illustration of the calculation of the return to the additional capital method, on the basis of various load assumptions. 14. Levels of capital investmsnt for hydro installations yielding various returns on additional investment. 15. Levels of capital investment for hydro installations yielding various returns on additional investment after premiums for foreign exchange savings. 16. Levels of capital investment for hydro installations yielding various return on additional investnent after premiums for foreign exchange savings adjusted to investment conditions typical in the more undeveloped countries. 17. Levelt~ of fuel costs to thermal plants adjusted for premiums on fOl~eign exchange savings yielding various returns on additional investment in hydro installations. IS. Investment in hydroelectric projects considered for !BRD financing. 19. Investment in thermal projects considered for IBRD financing. 20. Investment in foreign currency in hydroelectric projects considered for IBRD financing. 21. Investment in foreign currency in thermal prQ j ects for IBRD financing. Table 1 The return on total capital (profits before taxes plus interest) used in the financial analysis and the COf3t of ca.pital 11 used to calculate the uni t cost of power in various goverr..ment owned y hydroelectric projects considered for mRD financing. Estimated II . .; Cost of capital. financial return as a % of total as a %of total Project -iDlvestment investMent Damodar (Malthon, Panchet Hill) 4.~ 4.0 Lebrija, Colombia 2.3 2.3 Chidra1, Colombia 1.0 8.3 La Insula, Colombia 6.0 6.4 Anchicaya, Colombia 1.3 1.3 Cunucyacu, Quito, Ecuador 3.0 3.0 Litani, Lebanon 5.3 7.8 Seyhan, Turkey 4.8 4.8 High Aswan, Egypt, 1st Stage 3.8 3.8 Tokke, ~rorway 2.6 4.4 Lunersee, Austria 4.1 4.1 Chainat, Thailand 5.0 5.0 Kariba, Rhodesia, 1st Stage 4.8 7.2 Pirttikoski, Finland 6.0 6.0 Junkoski, Finland 6.0 6.0 Ypps-Persenberg, Austria 5.5 6.8 Yanhee, Thailand, 1st stage 5.0 8.9 Baygorria, Uruguay -2d ..1L.Q Unweighted mean }ledian 1I This represents the cost of capital used in the project reports in cal- culating the cost per kwh. In some cases calculated for an enterprise's total operations rather than for a specific project because of the lack of information in reports consulted. gj Government ownership is d~fined as government equity interest of at least 51%. J/ Financial analysis involved analysis of company data in some cases rather than project as such. Table 2 The return on total capital (profits before taxes plus interest) used in the financial analysis and the cost of capital 11 used to calculate the unit cost of power in various privately owned a/ hydroelectric projects considered for I8RD financing • . .. Estimated II Cost of ca.pital. financial return as a % of total as a %of total Project investment investment Kaltimo, Finland 10.0 17.1 Coscile, Italy 9.6 9.8 Luzzi, Italy 7.1 7.7 Capodi'Pollte, Italy 8.7 9.3 Cast~lla:!lO Asco1i, Italy ...§.a1 ..2.:J Unueighted mean ..Y 10.6 Median ~ ...2.:.l 1/ This represents the cost of capital used in the project reports in cal- culating the cost per kwh (in some cases) calculated for an enterprise's total operations rather than for a specific project because of the lack of information in reports consulted. 2/ Private ownership is defined as private equity interest of at least 51%. j/ Financial analysis involved analysis of company data in some Cases rather than projel,ct as such. 11 i Table 3 The return to the additional capital invested in hydro as compared to thermal alternatives and the comparative cost of power per kwh in various projects considered for IBRD financing. High Aswan Litani Y Kariba J/ Yanhee kI Bqgorria 21 - • 1st stage 11 1st stage 1st stae Uruguay Installe4 capacity ot ;. hydro in thousand kv 720 84 soo 140 10) Net generation in million kwh 1,62$ .336 1,167 409 4$1 Cost of hydro in $ per installed kw 310 418 442 550 455 Cost of. thermal in $ per installed kw 154 220 245 290 187 Additional capital required tor hydro in million $ Annual production expenses 112.0 .llil 98.0 - 47.9 29.6 for thermal including de- preciation in million $ 16ID8 2.9 16.1 4.1.&. 3.8 Annual production expenses for hydro incl .. depreci- ation in million $ Savings in annual production 8.1 ...l!! -,- 7.• 5 ...hl ...!!l. costs in million $ 8.7 1.8 13.6 .3.1 2.6 Return on additional capital invested in hydro 1. 8% -8.2% - 13.CJ!, 6.5% -8.8% Cost per kwh of hydro power in mills per kwh 10.2 9.7 4•.3 12., 8•.3 Cost per kwh of thermal power in mills per kwh Cost of capital used in pro- 13.2 ll., 6.9 16.8 11.9 ject report 3.8% S.3% 4.8% ,.0% 6.2% = = 11 Calculations based on 1965 estimates from Tables 18-19, Project report TO 94b and a~location of investment to power as indicated in that report. Y ':a.Jtal cost of dam allocated to power. The cost of other works to be used ex- clusively for irrigation have been excluded as indicated in Annex " TO 87a. ~ Calculations based on 1964-1965 estimates from Annex 5, Project report TO 116. EI Calculations based on USBR report, pp. 176, 162 - completion of initial develop- .. ment, 1961. Thermal alternative calculated on basis of footnote 1, Table 17, page ~ 176, with 100,000 kw of installed capacity operating at 40% load factor. 21 Calculated on basis of data presented in project report TO 120, with fuel costs adjusted to price or 60 cents per million BTU, rather than the Urugu~ fuel monopo~ price of about 90 cents per million BTU. Table 4 Levels of capital investmt11t for, hydro installations yielding various returns on the additional invest- ment over that required for thermal alternatives. (in ::~ per JaT of in'stalled capaci ty)~~ L01'1 price IlrIedium price High price ., for fuel for fuel for fuel at 25 cents at 50 cents at 70 cents per 'million per million per million Br.ru Return on additional capital -------------...;;;...-----_._------ BW BTU Below 6::1,":) abo've ('·1.Jl7 (~) (u'622 above ,,) above ~~786 6(!f ,/0 to 9(1 ,0 (i34a - ",lit7 t.P :)475 - ~i;622 "611 .) - ~:~i7 86 10% to 12% :)310 ... f::i333 ~)425 - ~~466 ~;)5l5 - ~~573 Above 12% below ~~310 1belO't'1 \t ~?425 below (~5l5 * Assumptions: Both thennal and hydro stations have installed capacity of 100,000 KW and operate at '"(0% load factor (annual net ~eneration 614.0 million K'ttJH). Fuel consumption for the thermal plant :'s 10, 000 nTU per KHH, with capital inves1;ment at 0175 per installed KU, and production expense excluding fuel and depreciation at 0.50 mills per KUH,. Annual c1.eprecj atiOl1 allowance based on a sinl~ing fund method is 2. 75;~ of total capital inves'fjed. The depreciation charge for the hydro station is 1.5~(, 1-1i th produc.l0ion expenses excluding depreciation at 0.40 mills per I0JH. The above assumptions wi th respect to fuel consumption imply a thermal efficiency typical of ne't-J' plants installed in the U.8. Such efficiency may involve operational pressures which are too high fol" some under- developed countries. On the other hand, ~lch countries rarely achieve a 70% loac1• factor. This means that on balance the assumptions adopted here generally are somel-J'hat 'unfavorable to thermal. TABLE , DELIVERED PRICE OF n1PCRTED BOILER FUEL OIL, CIF PeRT OF IMPCRTATION, FCR VARIOUS POR'lS OF THE WCRLD AS OF AUGUST 1955. Price per barr,el Estimated cost per in U.S. $ . million BTU in U.S~cents* New York 2.50 42 Los Angeles 1.80 30 Vancouver 2.97 50 Punta Arenas 3.03 51 Rotterdam 2.79 47 Rijeka 2.93 49 Piraeus 2.87 48 R~stanura 1.64 27 Cepetm'1n 2.82 47 Bombay 2 •.39 40 Yokohama ).08 52 vJelllngton 3.)8 57 * Based on assumption 18, 000 BTU per lb. and 6.04 bt\lrrels per ahort ton. Source: National Planning Association CQnfidential ,IStudy. -.,. o TABLE 6 COOT OF COAL 11 USh1> FeR palER PRODUCTION IN VARIOUS REPRESENTATIVE COUNTRIES DUR.ING lQ5' Cost in Cents Country Per 11111ion BTU United Kingdom 40 Austria 6; France 57 vJ est Germany 5D Italy 55 Turkey 80 India 25 Japan 90 Union of South Afr~ca 15 &-azil 100 Australia 40 Denmark 45 Greece )0 Netherlands 50 Yugoslavia. )0 1/ Cost delivered to steam plants. Source: Econgmio Aspects of Electric Power Produgtion in Selected Countries, Harry I. Nil1er, F. Douglas Campbell, July 1955. (lCA) Appendix D. ·6 TABLE.7 , Construction costs and operating data for various privately Qonstructed hydroe1ectl"ic plants in the U.S. (Annual data for the year 1225). Pacific Connect- Wiscon- Gas & icut sin Calit. E1ec.Co. South Light & River Oregon (N.Fork Carolipa Minnesota PO'tr1er Pm-Jer FOlier Feather Elec" & Pouer &. Co. Co. Co. River) Gas Cae Li~ht CO e . : ~!J : I: II ~~, ~3' Ui' : l~r.- lo' \ Installed cap. in thou. roof 37,200 15,000 18,000 U3,400 130,000 67,400 Date of construction of plant 1955 1951 1951 1950 1930 1930 Uet generation mill. KWH 204.0* 80.3 104.6 355.2 103.1 343.9 Plant factor in % 6'§r 61 66 35 9 58 Peak demand in thou. Kt'1 48,000 18,800 18,000 110,000 108,000 69,700 Original Cost of Plant «(~ per installed I{\~) Land & land rights Structures & improvements 60 21 183 83 21 - 18 20 43 8 10 12 ReservOirse dar.1S & l-latert"Jays 285 ~\46 157 218 101 58 Equipment ~osts 111 94 45 33 14 25 Roads, ~ai~oads & bridges 0 1 9 3 1 0 Total cost per Elrl of in- stall€~ capacity 477 607 232 352 167 105 Total c.bst (in thou. $) 17,726 9,110 4.,162 39,247 21,693 7,06.3 ~oduction Expenses in mills per K1·JH Operation, labor, super- viSion, engineering 0.30 1 , 0.55 0.08 0.14 0.74 0 • .38 Supplies ruld expenses 0.03 0.31 0.03 0.01 0.12 0.03 lwfaintenance other expenses, e.g. rent 0.06 0.30 0.31 0.07 0.03 o~05 0.27 0.09 ',', .. Total Production expense 0.39 1.16 0.49 0.23 1.13 0.50 Estimated overhead using Federal Power Commission Standards 11.06 13.80 4.86 13.50 25.60 2.51 Total costs in mills per Kt'IH 11.45 14.96 - 5.35 13.7.3 26.7.3 - 3.01 ~~ Annual rate as plant only in operation several months during 1955. -- W3IEa Construction costs and operati -data for various privately constructed steaITl ot-iTer pla..'1ts in the U.5. Annual date for the year ,19,4)". n Public Long Minn. Amer- Ser- South Conn. Is. Pm-1er ican vice Gulf vlest- Salt Idght Light- & Gas & Co'. of States ern River & ing Light Elec. New Utilit~ Public P.ro- Pot-ler Co. Co. Co. Hamp- ies Co. Ser- ject, TVA Co. "Far "Aur- "Kan- shire "St- vice IISag- "King- tlDevon" Rock";;;: ora" awha" "Schi1- ation Co.' uaro'! stontl conn,. away" HiM. River le~" No.2" tlEast.. Ariz~ Tenn. (1) (211 (3) (4) lS} • (6) (7) (.8) (9),' Installed caj,). in ~hou.KVl ,272,000 90,000 88,000 3,1,500 80,000 140,000 6S,000 100,000 405,000 Date of construc- tion of plant 1953 1953 1953 1953 1950 1951 194f 1953-4 1954 Net generation , ; mill. KIrJH 1,)88.'4 704.6 353.9 3,027.7 479.2 8~i.1 1~.30~4 575.'0* 3,38'u~ Plant factor in %58% 89% 46% 98% 68% 7($ 76% 66%{H~ 95%~~ Peak demand in thou. Kti 2711~200 115,000 84,700 460,000 97,100 161,,000 80,,500 102,000 422,090 Original cost of plant (in $ per installed KW) land & rand rii#S I 6 0 2 1 o 1 1 3 Structures & im- provements 33 62 23 29 65 11 17 11 ,"4 Equipment 92 156 123 120 182 104 100 112 1(22 Total Cost }:eX' KtnJ 126 224 146 151 248 . 115 . 118 124 159 Total'Cost in =======================}~c'=====--================= thou. $ 34,323 20,167 12,852 52,923 19,822 16,115 7,657 1~,381 64,363 Production Expenses (in mills. per ro'm) Operation, labor, " supervision et~0.37 0.28 0.38 0.09 0.46 0.23 0.38 0.32 0.44 Supplies & Exp. 0.03 0.06 0.03 0.03 O.Oj~ 0.05 0.2, 0.17 0.12 ~~intenance 0.53 0.15 0.23 0.16 O.95~ 0.16 O.,~ 0.11 0.27 Other expenses . 0.01 0.03 0.01 Fuel - 4.27 3.57 1.75 3.59 1.12 ;t.09 2.37 1.94 Tpta1 production expenses 5.30 4.06 2.04 2.77 Estimated over- head using FPC standards 3.24 .3.75 4.75 2.57 ,,$.46 2.4! 2.32 1.83 1.57 :rotal costs in ~lls per~m =8.=5~4==~7=~=81====\9=.5=3===4=.6=1==~lQ~.~5=3==#~.=Ol====4=.3=O==;=4~.8=1:;_~,.,_~4=.3=4== 11'ajor Fuel Used: Av, t3TU per l~M 11,953 9,894 11,,511 9,113 9,,666 12,200 15,314 10,854 9,877 Type Coal Coal Coal Coal Oil Gas Gas Gas Coal Cost delivered plant $ per short ton $ per barrel .. or ~ per MCF 9.40 9.49 8.37 4•.30 2.17 10.00 6.39 24.00 4.8S Cost per mill. BTU,rj 35.76 36.04 7.10 21.83 19.59 *C\ Completely moderniz~d ** Annual rate since first operation June 1954. TABLE 9 "INVESTMENT AND, OPERATING COST DATA FGR VARIOUS GOVERNMENT CONSTRUCTED HYDROELECTRIC PROJECTS IN THE U.S. (ANNUAL DATA "FOR THE YEAR 1955) Bureau of Bureau or Reclamation R~c1amation City or Anderson D~vis Dam Los Angeles D8l,I1, Boise Colorado Bishop.Dam Ri ver. Idaho' River. Arizona Ca1:t.f~rniA. (1) (2) (3) Installed capacity in thousand KW 27,000 225,000 37,500 Date of construction of plant 1951 1951 1952 Net Generation, million lU-lH 129.8 ).,185.1 136.9 Plant::, factor in % 55 60 :~2 Peak demand in thousand KW 38,000 272,000 39,000 Jnyestment in p1~ J$ per install eel K\v) " Land and ~and rights Structures and improvements 59 47 53 Reservoirs, dams and waterways 58 186 232 Equipment· costs 64 45 50 Roads, raJlroads and bridges 6 Total co~ per KW of installed capacity 181 278 341 Total investment in thousand i 4,878 62,681 12,784 JEToduetion bosts in milIR' pAr KW~ Operation, labour, supervision, engineering 0.28 0.09 0.21 Supplies and expenses 0.01 0.06 0.02 Maintenance 0.12 o.m 0.15 .., Total production expenses 0.41 0.22 0.38 .-- Estimated overhead us~ng Federal Power Commission standards. 2.68 3.78 6.65 Tatial costs in mills per KWH 3.C9 4.00 7.0) - - - - TABLE 10 THE RF,TURN TO CAPITAL IN VARIOUS NON-PQIER PROJECTSFINANC~j) BY IERD (AS A % OF TarAL INVESTlvlENT) Annual increase Rate of in net profit return to Capital invested a.ttributable to capital in project (in investment (in invested Projeet thouAand dollArs) thousand dollArs) (in %) Port of Corinto, Nicaragua 4,74 0 800** 16.,9 Port of latakia, Syria 14,75 0 1,400 39.2 Honduras Highway Maintenance 8,200 1,000 12.2 Kawasaki Steel, Japan 52,3 00 5,560 10.6 Karnaphuli Paper, Pakistan 20,000 1,980 9.9~B~* Pacasmayo Cement, Peru 5,000 630 12.6 Mitsubishi Shipbuilding, Japan 2,000 220 11.1* Toyota Motor Co., Japan 9,64 0 1,390 14.4 Yawata Steel, Japan 13,900 2,920 21.0 Nippon Steel Tube, Japan 11,3 00 1,260 1l~1 Japan Seamless Tube 2,200 470 21.5 Total $l44,03 0 $17,630 12.3% - --- - Unweighted mean rate of return *** Foreign exchange savings 22% ** Annual savings estimated for the year 1955. * Annual savings on engine production alone. TABLE 11 REPRESENTATIVE RATES OF RETURN* AS A %OF TarAL CAPITAL INVESTMENT IN VARIOUS UNITED STATES MANUFACTURING INDUSTRIES DURING lQS4. Rate of~ r·eturn Rate of return Industrv before t,axp.s after tmces Apparel & Finished Textiles 9.0% Paper & Allied Products 18.0% Chemicals & Allied Prcducts 19.8% RUbber Products 17.0% Leather Products 12.4% Primary non-ferrous metals 15.3% Primary Iron & Steel 14.4% Products of Petroleum & Coal 13.9% Nachinery 16.3% Electrical Machinery 22.8% Motor Vehicles & Parts 30.8% Unweighted Mean 16.7% -- - * Net of interest payments. However, debt averages only 8.8% of total capital in above industries. .TA~~ Illustration of the calculation of the return to the additional capital required for hyc1roelectric stations as compared to thermal power alternatives for various cost assumptions for hypothetical projects. Project Projec~ Project A B C 1. Installed capacity in IGo] 100,000 100,000 100,000 2. Net generation in million IDA~ 614.0 614.0 614.0 3. Plant load factor in %=(2) ~ (1) x 8760 brs 70 70 70 4. Cost of hydro plant in ~~ per installed I