Estimation of water supply costs in Cali, Colombia

DRAFT THE WORLD BANK DEVELOPMENT ECONOMICS DEPARTMENT URBAN AND REGIONAL ECONOMICS DIVISION URBAN AND REGIONAL REPORT No. 76-14 ESTIMATION OF WATER SUPPLY COSTS IN CALI, COLOMBIA JOHANNES F. LINN September 1976 These materials are for internal use only and are circulated to stimulate discussion and critical comment. Views are those of the author and should not be interpreted as reflecting the views of the World Bank. References in publications to Reports should be cleared with the author to protect the tentative character of these papers. I................. ...... ...... DRAFT THE WORLD BANK ESTIMATION OF WATER SUPPLY COSTS IN CALI, COLOMBIA September 1976 Prepared by: Johannes F. Linn for Research Project 671-18 (Pricing and Financing Urban Services) ..............L..... TABLE OF CONTENTS Page No. I. INTRODUCTION .................................................. 1 II. WATER SUPPLY SYSTEM IN CALI ................................... 2 III. PROJECTED WATER CONSUMPTION AND COSTS ......................... 7 A. Water Consumption ........................................ B. Capital Costs ............................................ 9 C. System Costs ............................................ 10 D. Zone-Specific Capacity Costs ............................. 1 E. Other Capital Costs ...................................... 20 F. Distribution and Density ................................. 22 G. Short-Run Marginal Costs ................................. 2 IV. AVERAGE INCREMENTAL COSTS OF WATER SUPPLY IN CALI ............. 31 A. Systems Costs ............................................ 3 B. High Level and Low Level Costs ........................... 39 C. Yumbo, Siloe-Lleras and Terron Colorado .................. 2 D. TotalCosts... . .. . . .. . .. . . .. . . .. . . .. .4 I.. INTRODUCTION This preliminary note reflects an interim statement on an ongoing exercise of estimating water (and eventually, sewerage) supply costs in Cali, Colombia. The objectives of this exercise are to: (a) provide alternative estimates of average incremental cost (AIC) depending on changing assumptions regarding (i) the base period chosen (ii) the time horizon (iii) the discount rate (iv) the level of aggregation of costs across space (v) the shadow price of foreign exchange in the particular context of the expansion program for water supply in Cali, as outlined in the Consultant Report by Planes-Indecon in October 1975; (b) compare these AIC cost estimates with other possible cost estimates, such as average cost or short-run marginal costs; (c) provide the basis for eventual combination of these cost estimates with demand and consumption pattern estimates in ascertaining the efficiency, distributional and financial effects of alternative pricing systems. The present paper completes task (a) and provides the data base for task (b) above. Further work is scheduled to complete (b) and (c) and to repeat a similar analysis for sewerage services. 1/ Planes-Indecon, "Programa de Obras Para. la Etapa II del Plan Maestro de Acueducto y Alcantarillado," Cali, October 1975. Prabhas Sharma and Nelson Valverde have provided computational assistance in the preparation of this paper. II. THE WATER SUPPLY SY'STM IN CALI Figure 1 reflects schematically the major features of the Cali water supply system, which will briefly be explained in this section. Presently the system consists of two major supply zones: the High Level Zone and the Low Level Zone ("high" and "low" referring here to altitude, not income levels). The High Level system is supplied by gravity from the San Antonio Plant, which draws high quality water from the Cali River. This River is relatively small and subject to substantial seasonal variations in flow in turbidity. Some treatment of this water is required. The Low Level system is presently supplied by the Rio Cauca Plant, which draws low quality water from the Cauca River. After extensive treatment the water is pumped from the Plant to the Low Level system. Construction is presently underway, on a further water intake, treatment, and pumping facility, the Puerto Mallarino Plant, which is to ba completed by 1976. Plans for future expansion envisage an increase in the capacity of the Rio Cauca Plant in 1984, and an expansion of the Puerto Mallarino Plant in 1987. Furthermore, a repair and minor expansion of the San Antonio Plant in 1976 and 1977 will increase the wet season capacity of this plant, without however changing its dry-season capacity. The High and Low Level Zones are interconnected such as to permit at present transfer of water from the former to the latter, but not vice versa. In order to facilitate the transfer of water from the Low to the High Level network a special pumping station (La Normal) is planned to enter into operation in 1978. Its function is to permit supplementing the high water supply during dry seasons and to provide a margin of ,_.r _ . , .а� � . _ _�. , . ,. _ _ ,. _ _ . _ � , - , . , r � � oU.N�'йt N . �h��S R-p`�J('1� $ it-o� t�,��c�o'" � 1 � �� � � � � � •� �� � .�� t � GLJ��a r � ,j . ` ` � . f r �� - � � �' , - � � ' . .` �� � � .�� ! � ` - ' � '�� /� , � J � � � �. J /f � г ,� `� .� ! �� �� �' , . f� � � !�` � J` ! % ? ` � - /f �( К ' � f / l! � �. ! l' � � � / �� , . �� f r` , � � • ' 1 � � l � � �' r � ' ,� ° � �о` � г'� � �� fJ� f г �' 1 ' г � А` �+ i � � � � . i / / � � �• f г,, * / �L1 / / � �'l l �1 � � Е V � �.. �' �.. о•лi Ч и м в о 9.._. Е. v � t.._ � �о � �,°'LL.c�` �„h„�т GAC.r 'Ри.�'� r�вд-►Ав' ио �1 J�.[ Т;„АА7С .-- �- .�- G�uca ,�2� и�R- � �..`�.`r ; � � � � i�1 � � Fzgure 1: '�НЕ WATER SUPPLY SYSTEM IN CAI,I s �: � f w _� ' 1 й � 4 security against the risk of plant failure in the San Antonio Plant. Table 1 and Figure 2 summarize the projected demand and capacity increments. L In addition Fo the two main supply zones there are presently a number of high lying areas requiring secondary pumping, provided by local pumpaxig stations. The two main areas where this is necessary are the Barrios (Neighborhoods) SJI oe-I-I eras and Terron Colorado. Furthermore, EMCALI (the CaliPublic Utility Cbmpany) is planning to extend its service to the neighboring municipality of Yumbo by building a water main from the existing low level network to Yumbo. The overall expansion of the water supply system further requires additions to the storage capacity for the High Level, Low Level, and Yumbo distribution networks. Table 1: CALI WATER DEMAND AND PLANT CAPACITY, 1975-1995 (m3/sec) Maximum Plant Capacity Year Daily Demand A B C D E F Total 1975 4.16 1.2 2.5 3.7 1976 4.61 1.2 2.5 4.0 7.7 1977 4.89 1.2 2.5 4.0 7.7 1978 5.16 1.2 2.5 4.0 7.7 1979 5.48 1.2 2.5 4.0 7.7 1980 6.09 1.2 2.5 4.0 7.? 1981 6.56 1.2 2.5 4.0 7.7 1982 7.08 1.2 2.5 4.0 7.7 1983 7.60 1,2 2.5 4.0 7.7 1984 8.13 1.2 2.6 1.0 4.0 8.7 1985 8.69 1.2 2.6 1.0 4.0 8.7 1986 9.36 1.2 2.6 1.0 4.0 8.7 1987 10.05 1.2 2.5 1.0 4.0 4.0 12.7 1988 10.76, 1.2 2.5 1.0 4.0 4.0 12.7 1989 11.53 1.2 2.5 1.0 4.0 4.0 12.7 1990 12.28 1.2 2.6 1.0 4.0 4.0 12.7 1995 16.9 1.2 2.6 1.0 4.0 4.0 4.0 16.7 A - San Antonio B - Rio Cauca C - Extension of Rio Cauca Plant D Puerto Mallarino - Stage I E - Puerto Mallarino - Stage II F - Puerto Mallarino - Stage III Source: Planes - Indecon, op. cit. ―… 7 III. PROJECTED WATER CONMPTION AND COSTS Given the objectives of this study as laid out on page 1 it is necessary to analyze carefully the projected water con6umption patterns and cost components from the available engineering data. The aim is to break down consumption and costs by major zones within the supply area, for each year of the planning period, and according to whether or not the input into water prod-action and distribution is domestically produced or imported. Tables 2 to 9 lay out the consumption and cost data in all the necessary detail. This section will describe each component and its derivation. A. Water Consumption Total projected system water consumption is based on consultantsJ estimates. These in turn are derived from per capita consumption and population projections, and from estimates of industrial water use. The population projection-s extrapolate the rate of population growth of Cali from the intercensal period 1964 to 1973, and are thought to be quite reliable. In particular, they reflect the slow down in population growth in Cali experienced during the last ten years, as compared with earlier years. In contrast, the per capita demand projections probably overestimate actual future demand. The consultants project a rising per capita demand from a level of about 285 l.p.c.d. in 1975 to a level of 393 l.p.c.d. in 1995.' despite the fact that between 1967 and 1975 per capita demand hardly increased at all (from 270 l.p.c.d. in 1967 to 286 l.p.c.d. in 1975). The reason given for this increase in per capita demand is the low system pressures in some areas in recent years and the expected increase in per ............ -8- capita income over the next twenty years. But given that future new connections will continue to involve mainly low income families and given that the relative price of water will probably continue to increase in future years, one would not expect major increments in water per capita consumption. This assumption is supported by the independent judgement of EMCALI personnel. Despite these shortcomings no attempt was made here to revise the consultants' estimates donwards, since this would probably also have required a revision of the investment program. As our primary objective in this study is to test the effects of alternative pricing strategies in terms of their efficiency, distributive, and financial effects, rather than to develop an optimal investment strategy for EMCALI, we will take as given the investment plans as of June 1975 and the demand projections on which they are based. / Overall demand may be broken down by supply zones as shown in Tables 2 and 3. The consumption figures for Yumbo are based on surveys of prospective industrial. users in that municipality and estimates of the relatively minor residatial demand component. Consumption projections for the High Level network are based on interpolations from past to future per capita consumption and population estimates for that area. Low level consumption is then derived as a residual from total system consumption projections. Consumption projections are available for the period 197,5 through 1995. For the period 1969 through 1974 actual consumption figures were used. 1/ Ideally one would of course want to link the pricing rules back to the investment decision, first in their capacity to serve as investment guides, and secondly, in their effects on demand projections. -9- B. Capital Costs As with the consumption informationcapacity cost data are organized by supply zone. Depending on the nature of the investment made costs are attributable either to the entire system, or to particular zones within the system. The decision determining cost attribution is made on the basis of whether or not the investment (and the resulting maintenance and operations expenditures) is required by increments in the consumption in any particular zone only, or by consumption increments taking place anywhere in the system. This will then allow the computation of zone- specific incremental costs as opposed to costs incremental for the system as a whole. Cost estimates were assembled for the years 1969 through 1974 on the basis of actual system expansions, recorded in the four-monthly investment reports of EMCALI. For the years 1975 through 1985 cost projections are derived from the previously cited consultants' report and from the invest- ment program of the Puerto Mallarino Plant (Stage I) and associated works that remain to be carried out in 1976 and early 1977. For the period beyond 1985 no cost estimates are presently available. All historical data are inflated to reflect constant June 1975 prices, using Ministry of Public Works price indeces for construction materials most suitable to water extension programs, according to EMCALI staff. For all major capital works maintenance and operations costs are calculated. Operations costs exclude all short run marginal costs such as energy for pumping and chemicals for treatment. These costs are further analyzed below. Since maintenance and operations costs were not projected in the Consultants' report for future system expansions, they had to be - 10 - interpolated from past experience with existing plants. Given the rough nature of these estimates, a substantial margin of error must be allowed for. In order to bracket the likely range, two alternative assumptions are testedi For Alternative 1 maintenance and operations costs are computed as a constant sum per installation, independent of capacity use. For Alternative 2 maintenance and operations costs are computed on a per unit basis, and are assumed to increase in proportion with capacity use. C. System Costs (Table 2) Consider first these capacity costs which are. attributable to the system as a whole. Since all zones - High Level, Low Level, Yumbo, etc. - are interconnected an increase in production capacity, whether it occurs in the High or Low Level plants, is attributable to the entire system. Thus, the construction costs of the Puerto Mallarino Plant, of the Rio Cauca Plant extension, and of the San Antonio -extension, as well as the associated maintenance and operations costs are attributed to the entire system. Linked to production capacity are also expenditures on the reforestation up the Cali River, and on maintenance equipment. Investment in storage capacity is also treated as a system cost, since an increment of water use anywhere in the system requires increased storage which may be provided anywhere in the interconnected system. Somewhat more questionable is the inclusion of extension of water mains among system costs. It is justified on the grounds that these represent costs required virtually throughout the system and cannot be attributed to specific areas, or consumers. Finally, for the case of Alternative 2 (assuming maintenance, operations, and other indirect costs increase in proportion with capacity use) the incremental costs of maintenance, Table 2: SYSTEM CONSUMPTION AM CAPACIY (COSTS IN COL* M11IONS, CONSTANT JUNF 1975 PRICES) 1/ WaterP R 0 D U C T 1 0 14 Water P O U T O Consu=ation Poerto Nallarino Construction . Rio Cauca Extension San Antonio Expansion Aver.ge Annual aintenance and Operation an Y'ar =-sec. (r3,.00) Plant AlternativP 1 Alternative 2 - Plant Alturnacive 1 Alternative 2 Plant Alternative 1 Alternati%e - 1963 2.46 77,380 0P 2.59 81,760 1971 2.65 83,585 31 1572 3.10 97,820 3.2 (52.0) 1973 3.13 98,915 8.8 (52.0) il-?l 3.18 100,010 119.2 (52.0) 1575 3.34 105,120 124.6 (52.0) 6/ ! IPA 3.69 116,435 100.0 (52.0) 6.6 (37.5) 19-7 3.91 123,310 5.9 (520) 4/ 11.0 (.0.0) 7/ 19 4.13 130,305 14.2 14.2 2.8 2.8 9 4.38 137,970 14.2 15.0 2.8 0 100 4.87 153,580 14.2 16.7 2.8 3. 1951 5.25 165,564 14.2 18.0 22.0 (50.1) 2.8 3. 1'32 5.66 178,494 14.2 19.5 22.0 (50.1) 2.8 1'sl 6.08 191,739 14.2 20.9 22.0 (50.1) 5/ 2.8 4.1 1r; 6.50 204,984 14.2 22.3 3.6 - 3.6 2.8 4. 1"S5 6.95 219,175 14.2 23.9 3.6 3.8 2.8 47 1 7.'5 235,205 14.2 25.7 3.6 4.1 2 5.1 19-7 8.04 253,549 14.2 27.6 3.6 4.4 2.8 5.* 1935 E.61 271,525 14.2 29.6 3.6 4.7 2.S 5 13 9.2 230,762 14.2 31.7 3.6 5.1 8 o 1-43 9.62 309,684 14.2 33.7 3.6 5.4 2.8 6.? I51 1C.40 327,974 14.2 35.7 3.6 5.7 2.8 7.0 195r 11.04 348,157 14.2 37.9 3.6 6.1 2.8 7. 1'93 1:.72 309,602 14.2 40.3 3.6 6.4 2.8 7.9 19j" i1.40 391,046 14.2 42.7 3.6 6.8 2.8 '.4 1931 13.11 413,437 14.2 45.1 3.6 7.2 2.8 8.9 1 ?t;Urt': in -a'eti .how thc proportions of foreign excbanpa in total costs in percntage cecas. All data for 1969 through 1974 are actual; data from 1975 onwards are projezed. Ai t'ttenace coats co=-ated as a lumi-aln per installation not increasing with capacity use. -." n 1. 7 . 2 ,A:in a-r.ance costs coputad on a per unit basis, as;uming maintenace costs increase proportionately with water use (note actual increase between 1965 and 1974). .: raa :cpoct; attal figures w'ere not nvailble. -es uriL ca1;i'Z .ozt is the sane for Rio Cauca and Puerto mallarino Plants, which means that the cost of the Rio Cauca Plant has to be scaled up by a factor of 1.6, cqLal to the ratio of the ca-.acity of the Puerto Nallarino Plant (4.0 m3/sec.) to the capacity of the Rio Cauca Plant (2.5 m3/sec.). 5/ Asse-es SSflu-:it capacity cost for Rio cauca before and after expansion, wIhich ineans that the actual 1975 maintenance coat is multiplied by 0.4. Assu-sz that all costs are actually spent for expansion; replacem2nt costs strictly speaking should have been treated separately. Ass aLssiie unit capacity cost for txpadcd, as for actual San Antonio Plant !Furca: s :CALI, Water and sewerage Laster Plan, 4-Monthly Reports, 1971-1975; Planes-Indecon, op. cit. *H I Table 2: SYSTEM CONSUMPTION AND CAPACITY (COSTS IN COL$ MILLIONS, CONSTANT JUNE 1975 PRICES) (Continued - Page 3 of Table 2) Administration T 0 T A L C 0 S T S and Other Total Maintenance and Indirect Expenses Total Capital Operations Costs Overheads Year _ Alterr.-itive 2 Costs Alternative 1 Alternative 2 Alternative 2 Year 1969 27.4 42.9 1969 1970 28.3 3.4 ( 0.0) 44.6 . 1970 1971 29.6 7.1 (14.-7) 46.3 1971 1972 34.6 3.6 (50.1) 0.1 0.1 54.1 1972 1973 35.0 21.1 (44.4) 0.1 0.1 54.8 1973 1974 35.4 120.7 (51.5) 0.1 0.1 55.4 1974 1975 37.2 129.7 (50.3) 0.1 0.1 58.2 1975 1976 41.2 140.7 (47.4) 0.1 0.1 64.5 1976 1977 43.7 . 57.5 (18.6) 0.6 0.6 68.3 1977 1978 46.1 17.9 ( 9.4) 18.0 18.0 72.1 1978 1979 48.8 18.0 (30.4) 18.3 19.3 76.4 1979 1980 54.3 6.6 ( 0.0) 18.3 21.5 85.0 1980 1981 58.6 24.6 (55.4) 18.3 23.2 91.7 1981 1982 63.2 24.6 (55.4) 18.3 25.0 98.9 1982 1983 67.9 24.6 (55.4) 18.3 26.8 106.2 1983 1984 72.5 9.4 (35.6) 21.9 32.3 113.5 1984 1985 77.6 9.4 (35.6) 21.9 34.5 121.4 1985 1986 83.6 22.4 37.7 130.8 1986 1987 89.7 22.4 40.4 140.4 1987 1988 96.1 22.4 43.3 150.3 1938 1989 102.9 22.4 46.4 161.0 1989 1990 109.6 22.4 49.4 171.5 1990 1991 116.1 22.4 52.2 181.6 1991 1992- 123.2 22.4 55.5 192.8 1992 1993 130.8 22.4 59.0 204.6 1993 1994 138.4 22.4 62.5 216.5 1994 1995 146.3 22.4 66.0 228.9 1995 1/ Without replacement costs Source: EMGALI, Water and Sewerage Master Plan, 4-Monthly Reports, 1971-1975; Planes-Indecon, op. cit. e 2: SYSTE1 CONSi:-PTION AND CAPACITY (COSTs IN COLS MIILIONS, CONSTANT JUNE 1975 PRICES) (Continued Page 2 of Table 2) P RODUCTION, C cd. - ST O R A G E, EXTENSION OF NAINS xiscing Plant aintcCnance a::d " ncration Haintnace Tanks II Naintenance Maintenance and 0ý,cretio?ns Ai trnativeŽ 2 Reforestation Eauipmant Tanks I Land I (incl. Mains) Alternative I Alternative 2 Capital Costs Alterintiva 1 Alt-rnative 2. i5.5 16.3 3.4 .u 171 16.7 4.1 3.0 (34.9)- 79.5 ./ 0.4 (34.9) 0.1 0.1 173 19.8 10.8 ( 86.8) 1.1 (10.9) 0.4 0.1 0.1 3 37- 20.0 0.5 (10.9) 0.8 0.2. (34.9) 0.1 0.1 1. 975 21.0 3.7 (10.9) 1.4 0.1 0.1 976 23.3 6.6 11.4 (100.0) 2.0 (10.9) 14.1 ( 4.5) 2/ 2/ 0.1 0.1 577 24.6 6.6 4.4 (100.0) 8.1 (10.9) 21.5 (10.9) 0.5 0.5 0.1 0.1 '78 26.G 6.6 11.3 (14.9) 0.9 0.9 0.1 0.1 27.6 6.6 11.4 (48.0) 1.2 1.2 0.1 0.1 9 30.7 6.6 1.2 1.4 . 0.1 0.1 33.1 2.6 (100.0) 1.2 1.5 0.1 0.1 35.7 2.6 (100.0) 1.2 1.6 0.1 0.1 523 38.3 2.6 (100.0) 1.2 1.7 0.1 0.1 9,4. 41.0 2.6 (100.0) 6.8 (11.0) 1.2 1.8 0.1 0.2 43.8 2.6 (100.0) 6.8 (11.0) 1.2 1.9 0.1 0.2 !5 47.2 1.7 2.6 0.1 0.2 50.7 1.7 2.8 0.1 0.2 iJ 54.2 1.7 3.0 0.1 0.2 58.1 1.7 3.2 0.1 0.2 92 61.9 1.7 3.4 0.1 0.2 19 5 1.7 3.6 0.1 0.2 L "92 69.6 1.7 3.8 0.1 0.2 73.6 1.7 4.1 0.1 0.3 78.1 . 1.7 4.3 0.1 0.3 62.6 1.7 4.5 0.1 0.3 1995 Ar:. Apraisai Report; actual figures were nat avaý4îble. 3 sad on per unit scor'ýe capaciry cost- for m,aintenance, etc. in 1975, asauming the same unit costs apply to each addition in storage capacity (24,000 m in 1976, 15,000 in 1977, 15,G00 in 1984 and 1985 each). c, 2UCALI, Water and Sewerage master Plan, 4-Monthly Reports, 1971-1975; Planes-Indecon, op. cit. operations and other indirect expenditures associated with the existing system must be allowed for. For 1975 these are taken from the actual accounting data of EMCALI water operations and are extrapolated back-and forward in line with the proportionality assumption. D. Zone-Specific Capacity Costs (Tables 3 and L) For the High-Level Zone the costs of the pumping station "La Normalt? must be ccmputed. For the Low Level Zone the cost of constructing new mains is here included. For Yumbo, the cost of laying the water main from Cali to Yumbo is counted. For each of these capital costs, alternative maintenance and operations costs are computed. Since no independent estimates for these costs was available they were assumed to equal 2 percent of capital costs on an annual basis. For the 'High and Low Level zones maintenance and operaticns costs of the existing distribution system could also be estimated for Alternative 2, since the distribution system in the High Level Zone consists mainly of cast iron piping and that in the Low Level Zone of asbestos cement piping, and since separate maintenance cost figures are kept by EKCALI on each type of piping. Finally, for the Barrios Siloe Lleras and Terron Colorado, the capital and maintenace costs of recently installed secondary pui.ng stations were estimated together with the consumption increments made possible by these pumping stations (Table 4). The total cost of secondary pumping stations is shown in Table 5. From the available cost data it was not possible to allocate this cost total to individual pumping stations beyond the data presented in Table 4. Table 5 also shows the total of actual and projected capacity costs for the period 1969 through 1986, summing all cost streams:attributed Table 3: CONSUMPTION AND CAPACITY COSTS FOR HIGH AND LOW LEVEL SYSTEMS AND FOR YUMBO (Costs in Col$ millions, constant June 1975 prices) H I G H L E V E L Distribution Consumptiog La Normal Pumping Station Maint. Cost Year (m3,000)1/ Plant Alternative 1 Alternative 2 Alternative 2 1969 28,067 1.6 3/ 1970 29,013 1.7 1971 29,328 1.7 1972 33,74 1.9 1973 33,428 1.9 1974 33,74 1.9 1975 34,690 2.0 1976 35,951 11.0 (95.7) 2.1 1977 37,212 10.2 ( 0.0) 2/ 2.1 1978 38,474 0.4 ~ o.4 2.2 1979 ho,o51 0.4 0.h 2.3 1980 41,628 0.4 0.5 2.4 1981 43,204 0.4 0.5 2.5 1982 4h,781 0.4 0.5 2.6 1983 46,358 o.h 0.6 2.7 1983 48,250 0.4 0.6 2.8 1985 50,142 0.4 0.6 2.9 1986 52,034 0.4 0.7 3.0 1987 53,927 0.h 0.7 3.1 1988 55,819 0.4 0.7 3.2 1989 58,026 0.4 0.8 3.3 1990 60,234 0.4 0.8 3.5 1991 62,441 0.4 0.9 3.6 1992 64,649 0.4 0.9 3.7 1993 67,172 0.4 1.0 3.9 1994 69,695 0.4 1.1 4.0 1995 72,217 0.4 1.1 4.2 1/ Estimated for 1969 through 1974 by assuming that the proportion of High Level consumption in the total decreased by 0. percentage points annualy from 36 percent in 1969 to 33 percent in 1975. 2/ Assumed to equal 2 percent of capital costs. Maintenance cost of cast iron pipe, distribution net, which is mainly High Level Source: EMCALI, Water and Sewerage Master Plan, 4-Monthly Reports, 1971-1975; Planes-Indecon, op. cit. Table 3: CONSUMPTION AND CAPACITY COSTS FOR HIMR AND LOW LEVEL SYSTEMS AND FOR YUMBO (Continued - Page 2 of Table 3) L 0 W L E V E L New Mains I and II Distribution Consumption Capital Costs Maint. Costs 2Year '(m ,000) I and II Maintenance 1 Maintenance 2 Alternative 2 1969 49,313 4.0 3/ 1970 52,747 4.3 1971 54,257 4.4 1972 64,076 5.2 1973 65,487 5.3 1974 66,266 44.9 5.4 1975 70,430 104.6 5.7 1976 80,484 79.5 2/ 2/ 6.5 1977 86,158 5.0 5.0 - 5.0 - 7.0 1978 91,831 5.0 5.3 7.4 1979 97,919 5.0 5.7 7.9 1980 104,383 5.0 6.1 8.4 1981 11.1,007 5.0 6.4 9.0 1982 118,576 5.0 6.9 9.6 c 1983 126,459 5.0 7.3 10.2 1984 134,028 5.0 7.8 10.8 1985 142,543 5.0 8.3 11.5 1986 151,689 5.0 8.8 12.3 1987 161,148 5.0 9.4 13.0 1988 171,240 5.0 9.9 13.9 1989 182,278 5.0 10.6 14.8 1990 193,001 5.0 11.2 15.6 1991 204,038 5.0 11.8 16.5 1992 216,967 5.0 12.6 17.6 1993 232,843 5.0 13.4 18.7 1994 244,719 5.0 14.2 19.8 1995 259,857 5.0 15.1 21.0 1/ Estimated for 1969 through 1974 by assuming that the proportion of High Level consumption in the total,decreased by 0.5 percentage points annually from 36 percent in 1969 to 33 percent in 1975. 2/ Assumed to equal 2 percent of capital costs. 3/ Maintenance cost of asbestos cement pipe distribution net, which is mainly Low4 Level. Source: EMCALI, Water and Sewerage Master Plan, 4-Monthly Reports, 1971-1975; Planes-Indecon, op. cit. Table 3: CONSUMPTION AND CAPACITY COSTS FOR HIGH AND LOW LEVEL SYSTEMS AND FOR YU1vBO (Continued - Page 3..of Table 3) Y U M B 0 Consumption Mains Year (m3,ooo) Capital Costs Maintenance 1 Maintenance 2 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 48.1 (66.0) 1979 52.1 (31.4) 1/ 1980 7,569 2.0 2.0 1981 11,353 2.0 3.0 1982 15,137 2.0 4.0 1983 18,922 2.0 5.0 1984 22,706 2.0 6.0 1985 26,490 2.0 7.0 1986 32,482 2.0 8.6 1987 38,474 2.0 10.2 1988 hh,b.66 2.0 11.7 1989 5o,458 2.0 13.3 1990 56,h9 2.0 14.9 1991 61,495 2.0 16.2 1992 66,541 2.0 17.6 1993 71,587 2.0 18.9 1994 76,632 2.0 20.2 1995 81,363 2.0 21.5 1/ Assumed to equal 2 percent of capital costs. Source: EMCALI, Water and Sewerage Master Plan; 4-Monthly Reports, 1971-1975 Planes-Indecon, op. cit. Table 4: CAPACITY COSTS OF SECONDARY UPING STATIONS IN SILOE-LLERAS AND TER10N COLORADO (Cost6 in Col$ million, constant June 1975 prices) S I L 0 E -L L E R A S P U M P I N G S T A T IO N T E R R 0 N -C 0 L 0 R A D 0 PU P I N S T A T I O N Consu::prio 2ConsumptiÜ 2/ 2' Year (.3,co) Plant Maintenance 1 - Maintenance 2 -(.3,000) Plant Maintenance 1 - Maintenance - Yea U;0i Dj71 - 17^ 197219 1973 1- 1974 2.8 (0.0) 2.8 (0.0) 5 1975 3.4 (4.4) .3.4 (4.4) . i976 1,150 0.06 0.06 1,054 0.06 0.06 197 1977 1,218 0.06 0.06 1,117 0.06 0.06 ig 1973 1,287 0.06 0.07 1,180 0.06 0.07 1 79 1,362 0.06 0.07 1,249 0.06 0.07 1930 1,517 0.06 0.08 1,391 0.06 0.08 1931 1,635 . 0.06 0.09 1,499 0.06 0.09 1932 1,763 0.06 0.09 1,616 0.06 0.09 1962 1983 1,893 0.06 0.10 1,735 0.06 0.10 19½4 2,024 0.06 0.11 1,855 0.06 0.11 195 2,164 0.06 0.11 1,984 0.06 0.11 196b 1E6 2,332 0.06 0.12 2,138 0.06 0.12 0 197 2,504 0.06 0.13 2,295 0.06 0.13 14-1 1i98 2,681 0.06 0.14 2,458 0.06 0.14 1959 2,871 0.06 0.15 2,632 0.06 0.15 1590 19"5 3,058 0.06 0.16 2,803 0.06 0.16 199, 1991 3,239 0.06 0.17 2,969 0.06 0.17 1992 1992 3,438 0.06 0.18 3,152 0.06 0.18 1993 1993 3,650 0.06 0.19 3,346 0.05 0.19 1994 1994 3,861 0.06 0.20 3,539 0.06 0.20 1995 1995 4,083 0.06 0.21 3,743 0.06 0.21 1/ ässL:es that the proportiorn in total consumprion of city in 1975 renains unchaaged, taking 250 lped as the basis for estimating 1975 consumption. 1 .stcd to equal 2 percent of capital cost. Szcrce: Office of the Water and Sewerage Masterplan. Table 5: TOTAL CAPACITY COSTS FOR WATER SUPPLY (IN COL$ NILLION, CONSTANT JUNE 1975 PRICES) 1/ System, High Level, low Level, and Yumbo Secondary Pumping Stations Grand Total Naintenance and Maintenance and 1Zaintenance and COerations Costs Overbends Operations Cost! Operations Costs Cerheeds Ca.pita1 Costs Alternative I Alternative 2 Alternative 2 Capital Costs Alternative 1 Alternative 2 Capital Costs Alternative 1 Aternative 2 Alternative 2 Ye 3 48.5 48.5 1969 0 3.4 ( 0.0) 50.6 3.4 '( 0.0) 50.6 19,0 7.1 (14.7) 52.4 7.1 (14.7) 52.4 11 3.6 (50.1) 0.1 0..L 61.2 2/ 3.6 (50.1) 0.1 0.1 61.2 l972 21.1 (44.4) 0.1 0.1 62.0 1.1 (55.4) 22.2 (44.9) 0.1 0.1 62.0 197 165.6 (37.5) 0.1 0.1 62.7 0.5 (55.4) 166.1 (37.6) 0.1 0.1 62.7 975 234.3 (27.8) 0.1 0.1 65.9 3.7 (55.4) 238.0 (28:2) 0.1 0.1 65.9 76 231.2 (33.4) 0.1 0.1 73.1 2.0 (55.4) 233.2 (33.6) 0.1 0.1 73.1 72.7 (14.7) 5.6 5.6 77.4 8.1 (55.4) 3/ 3/ 80.8 (18.8) 5.6 5.6 77.4 ,<8 66.0 (50.6) 23.4 23.7 81.7 0.3 0.3 ~ 66.0 (50.6) 23.7 24.0 81.7 9 70.1 (31.1) 23.7 25.4 86.6 0.3 0.3 70.1 (31.1) 24.0 25.7 86.6 11. 6.6 ( 0.0) 25.7 30.1 95.8 0.3 0.4 6.6 - ( 0.0) 26.0 30.5 95.8 3 24.6 (55.4) 25.7 33.1 103.2 0.3 0.4 24.6 (55.4) 26.0 33.5 103.2 1951 24.6 (55.4) 25.7 36.4 111.1 0.3 0.4 24.6 (55.4) 26.0 36.8 111.1 19'- 24.6 (55.4) 25.7 39.7 119.1 0.3 0.4 24.6 (55.4) 26.0 40.1 119.1 73 9.4 (35.6) 29.3 46.7 127.1 0.3 0.5 9.4 (35.6) 29.6 47.2 127.1 : 9.4 (35.6) 29.3 50.4 135.8 0.3 0.5 9.4 (35.6) 29.6 50.9 135.8 I7S 29.8 55.8 146.1 0.3 0.5 30.1 56.3 146.1 29.8 80.7 156.5 0.3 0.6 30.1 61.3 156.5 19 29.8 65.6 167.4 0.3 0.6 30.1 65.2 167.4 29.8 71.1 179.1 0.3 0.7 30.1 71..8 179.1 -9 29.8 76.3 190.6 0.3 0.7 30.1 77.0 193.6 19c, 29.8 81.1 201.7 0.3 0.8 30.1 81.9 201.7 29.8 86.6 214.1 0.3 0.8 30.1 87.4 214.1 29.8 92.3 227.2 0.3 0.9 30.1 93.2 227.2 29.3 98.0 240.3 0.3 0.9 30.1 98.9 243.3 29.8 103.7 . 254.1 0.3 1.0 30.1 1C4.7 254.1 3 erld by aiddirg all costs in Tables 2 and 3. r:.'ed from appraisal report. ssu_ed to equal 2 percent ofCcapital cost. i:CALI, I,nter and Sewerage 'aster Plan, 4-Nonthly Reports, 1971-1975; Planes-Indecon, op. cit. - 20 - to the system.as a whole, to the High Level, Low Level, and Yumbo Zones, and to secondary pumping stations. E. Other Capital Costs Other capital costs (see Table 6) include the extension of the secondary water distribution network; the extension of the distribution system to a number of highlying areas presently not served (Los Cristales, Menga, Bellavista, Cristo Rey, etc.); and house connections. The cost of extending service to selected highlying areas was also treated separately since this progr,m consists mainly of the secondary 1/ distribution network. - From the consultants' report it is not entirely clear whether it will also require secondary pumping stations. If so, these should have been included with the cost stream of secondary pumping stations discussed in the previous section. In any case however, cost data for such secondary pumping stations are not included in the donsultants' report. House connections are also treated separately here, as with the secondary distribution network. The Consultants' report estimated these costs on the basis of a projection of new connections, multiplied by the average cost per connection of Col$1,676-..21 as of June 1975. 1/ This is important for the AIC method, where the secondary distribution system is excluded from the costing framework. - 21 - Table 6: OHER CAPITAL COST (IN COL$ MILLION, CONSTANT JUNE 1975 PRICES) High Zones Distribution House Connections Distribution Maintenance Maintenance Network: Capital and Operation Capital and Operation Year Capital Costs Costs Alternative 1 Costs Alternative 1 1969 1970 10.5 1971 7.4 8.5 0.4 2/ 1972 7.8 46.8 0.7 1973 33.8 9.4 2.6 1974 23.1 3.5 3.0 1975 13.6 13.9 3.1 1976 1.7 22.2 3.7 1977 4.2 12.7 4.6 1978 (8.0) 15.0 13.5 5.". 1979 (8.0) 24.1 0.3 1/ 14.6 5.6 1980 (8.0) 24.1 0.8 26.6 6.2 1981 (8.0) 13.3 1.3 19.7. 7.3 1982 (8.0) 13.1 1.6 19.7 8.1 1983 (8.0) 13.1 1.9 19.7 8.9 1984 (8.0) 13.1 2.2 19.7 9.7 1985 (8.0) 13.1 2.5 19.7 10.5 1986 2.8 11.3 1987 2.8 11.3 1988 2.8 11.3 1989 2.8 11.3 1990 2.8 11.3 1991 2.8 11.3 1992 2.8 11.3 1993 2.8 11.3 1994 2.8 11.3 1995 2.8 11.3 1/ Assumed to equal 2 percent of capital cost starting with one year lag after construction. 2/ Assumed to equal 4 percent of capital cost staring with one year lag after installation; the 4 percent figure was obtained by computing the per connection cost of connection and meter maintenance for 1975 from EMCALI accounts (approx. Col$70 per annum), and applying this to the capital cost of one new connection in 1975 (col$ 1,676.21). Source: EMCALI, Water and Sewerage Master Plan, 4 Monthly Roports, 1971-1975; Planes-Indecon, a.it - 22 - F. Distribution and Density From a sample of water distribution projects executed by EMCALI (predominantly in low income neighborhoods) and by a private developer (predominantly in middle-income developments) it is possible to establish a quantitative estimate of the relationship between water distribution costs per subscriber and per hectare on the one hand and density of settlement on the other. Table 7 summarizes the dath specifying the neighborhood, the developer, the year of the project, an estimate of the average household incane, density, cost per hectare, and cost per subscriber. All cost data are in June 1975 prices, and they include house connections, as well as the pricing and secondary distribution pipes in the project area. Taking first the relationship between density and cost per area it was found that there is a positive and significant linear relationship as reflected in the following ordinary least squares regression equation: (C/A) = 13.4198 + 2.7538(S/A) R2 = 0.5220 where: C is the cost of the project, S the number of subscribers, and A the area convered by the project. This regression coefficient is significant at the 0.1 percent level. This indicates that in Cali the water distribution costs per hectare increase with increasing density. 1/ Cuellar, Serrano, Gomez & Salazar de Cali Ltda. / A logarithmic relationship was also estimated butliit showed a very poor fit and no significant relationship between the two variables even at the 10 percent confidence level. In (C/A) = 3.8659 + 0.1987 ln (S/A) R2= 0.1015 二 24 Turning then to the relationship between cost per subscriber and density one finds a significant negative correlation which is especially highfor the log-linear regression equation: (C/S) = 14-4278 - 0.2545(s/A) R 2 = 0.2627 2 = o.6618 ln(C/S) = 3.8810 - Mug in(s/A) R The regression coefficient is significant for the linear relationship at the 2.5 percent confidence level, and for the log-linear equation at the 0.1 percent lEvel. In other words water distribution cost per subscriber decreases as density increases; however, costs decrease by less than in proportion to the increase in density. This in turn is explained by the positive relationship between density and cost per area,, as can be inferred from the definitional identity (CIS) =_ (C/A) A.- (S/A) If (C/1) does not vary with (S/A) then (C/S) varies inversely with (SIA), or the elasticity of (CIS) with respect toQ(S/A) is eq-qal to minus one.!/ But when (C/A) increases with (S/A), as was shown to be the case for the sample of projects in Cali, then (CIS) decreases by less than in proportion with increases in (S/A). G. Short-Run Marginal Costs Short-run marginal costs are here defined as all those costs which vary directly with a marginal change in water consumption over a period of l/ Taking logs at both sides of the identity one obtains ln(C/S) a in(C/A) - in(S/A) which is similar to the last of the three regression equations above, except that the coefficient of ln(S/A) is here by definition minus unity. a day or week, rather than one or more years. Thus it is here assumed that only treatment and pumping costs vary in the short run with water consumption, and more particularly the costs of chemicals used in treat- ment and the cost of energy used in pumping. All other costs, such as labor, maintenance, land and capital rents are assumed to be invarient in the short run with a marginal change in water consumption. Given this set of assumptions which of the two plants currently producing water in Cali is the marginal plaht? According to information provided by the Superintendent of Water Production at EMCALIL the San Antonio Plant is always run at full capacity regardless which season. In contrast, the production of the Rio Cauca Plant is adjusted over the seasons to allow for variations in demand and in the supply of the San Antonio Plant. The reason for this production policy is that the short-run marginal costs at the San Antonio Plant are always below those of the short run marginal costs of the Rio Cauca Plant. This can immediately be verified from Table 8 which summarizes the short run marginal cost data for these two plants. Table 8 shows the monthly cost of chemicals, and of electricity required for pumping at the two plants in 1975. The cost of chemicals is computed at nominal costs of purchase, while energy costs are computed at the nominal price charged internally by EMCALI (which also produces electricity). This charge equals the industrial electricity tariff, which was set of Col$0.35 per kwh during January through October 1975, and was increased to Col$0.h5, effective November 1975. It has here been assumed that Col$0.45 is the economic opportunity cost of electricity,2/ during 1/ Strictly speaking this would need to be further justified by an investigation of the electricity cost structure. Table 8: SHORT-RUN MARGINAL COST WATER: RIO CAUCA PIANT AND SAN -ANTONIO PLANT Cost of Energy Total Cost 3/ Unit Cost Production Chemicals Use -Cost (Col$,000) (Col$,000) (Col.$ per m3) Month 1975 (m3,000) (Col,000) (K4h,0001 Nominal Real Y Nominal Real Nominal Real A. RIO CAUCA PLANT January 5,411 452 1,462.9 512 658 964 1,100 0.1782 0.2051 February 4,853 514 2,068.6 724 931 1,238 1,445 0.2551 0.2978 March 5,621 619 1,791.4 627 806 1,246 1,425 0.2217 0.2535 April 5,433 614 1,917.1 671 863 1,285 1,477 0.2365 0.2719 ,ay 6,092 634 1,928.6 675 868 1,309 1,502 0.2149 0.2466 June 5,723 567 2,531.4 886 1,139 1,453 1,706 0.2539 0.2981 July 5,797 600 2,442.9 855 1,099 1,455 1,699 0.2510 0.2931 August 5,731 702 2,362.9 827 1,063 1,529 1,765 0.2668 0.3180 Sepcember 5,594 600 2,631.4 921 1,184 1,521 1,784 0.2719 0.3189 October 5,693 787 2,345.7 821 1,056 1,608 1,843 0.2825 0.3237 November 5,362 698 2,026.7 912 912 1,610 1,610 0.3003 0.3003 December 5,753 929 1,673.3 753 753 1,682 1,682 0.2924 0.2924 Average 0.2521 0.2850 B. SAN ANTONIO PLANT 0% January 3,368 82 48.6 17 22 99 104 0.0294 0.0308 February 3,055 157 68.6 24 31 181 188 0.0592 0.0615 March 3,232 169 82.9' 29 37 198 206 0.0613 0.0638 April 3,243 259 77.1 27 35 286 294 0.0882 0.0906 may 3,240 315 74.3 26 33 341 348 0.1052 0.1074 *June 3,198 284 105.7 37 48 321 332 0.1004 0.1038 July 3,397 395 97.1 34 44 429 439 0.1263 0.1292 August 3,423 595 97.1 34 44 629 639 0.1838 0.1867 September 3,321 372 102.9 36 46 408 418 0.1229 0.1259 October 3,405 473 . 100.0 35 45 508 518 0.1492 0.1521 November 3,278 413 80.0 36 36 449 499 0.1370 0.1370 December 3,292 424 73.3 33 33 457 457 0.1388 0.1388 Average 0.1085 0.1106 1/ Computed on the basis of Col$0.35 per Kwh from nominal energy cost data for January through October; and Col$.045 per Kwh for November and December, 2/ Computed by multiplying the energy use figures of January through October by Coi$0.45 per Kwh. For November and December nor-inal equals real cost. 31 Sum of cost of chemicals and energy. 4/ Total cost divided by production of plant. Source: EMCALI, Superintendent of Water Production. -27 - all of 1975, implying that the electricity rate during the first ten months was below true opportunity cost.1/ For this reason the "real" (or economic) cost of energy had to be computed by applying the rate of Col$0.45 per kwh to the electricity use figures in the first ten months of 1975. As a result of this adjustment, the short-run marginal cost averaged across the twelve months is raised from a nominal level of Col$0.2521 per m3 to Col$0.2850 per m3 for the Rio Cauca Plant, and from Col$0.1085 per m3 to Col$o.1106 per m for the San Antonio Plant. These figures indicate that the short run marginal cost of running the San Antonio Plant is substantially below that of the Rio Cauca Plant. And despite substantial seasonal variations in the costs of tbe San A nio Plant these are never equal to or above those of the Rio Cauca Plant. Given this cost structure and the resulting policy of always running the San Antonio Plant at full capacity, and given the fact that all water supply system components in Cali are interconnected, the short run marginal cost of an additional unit of water used anywhere in the system equals the short run marginal cost of the Rio Cauca Plant. Seasonal variations in these costs are minori since the water quality at the Cauca River intake is virtually unchanged throughout the year and pumping costs do not vary over the year. For projection purposes it appears safe to assume that the short-run marginal costs at the new Puerto Mallarino plant will 1/ This is justified on the grounds that EMCALI is generally late in adjusting its user charges in response to changing costs. - 28 - approximately equal those of the Rio Cauca Plant. According to the Superintendent of Water Production at EMCALI less energy will be required for pumping of water from the intake to the treatment facilities at the Puerto Mallarino Plant as compared with the Rio Cauca Plant, due to the very much shorter transmission line from intake to plant in the former as against the latter case (50 meters versus 2,000 meters). But extra pumpig is required at the Puerto Mallarino Plant to dispose of waste water generated in the treatment process, while this is not the case with the Rio Cauca Plant, where the waste water can be disposed of by gravity. According to the estimates of the engineers, these two factors should approximately offset each other, leaving short run marginal costs approximately unchanged. It may furthermore be ncted that the amplification of the San Antonio Plant will not affect short run marginal cost either, since this plant will continue to be used at maximum capacity throughout the 1/ year, as costs will'.continue to remain below costs at the other plants. For those areas where secondary pumping is required because of special elevation, as in the case of Siloe Ileras and Terron Colorado (as well as in three other areas) further short run marginal costs are incurred on account of pumping. As in the case of energy costs for pumping at the production plants nominal energy costs are adjusted to reflect (estimated) opportunity cost of electricity. Table 9 shows that the "real" short run marginal cost varies between Col$0.2356 per m at Siloe Lleras and Col$0.3619 per m3 at Terron Colorado. 1/ In fact, while the expansion of the San Antonio plant increases the maximum flow during the wet season, it does not increase the flow during the dry season due to insufficient water flow in the Cali River. Taule 9: SIORT RUN MARCINAL COST: PUMPING STATIONS BELLAVIS TA B. MENCA Total Cost Unit Cost Total Cost Unit Cost Production Energy 1/ '. (Col$000) (Col$/m3) Production Energy 1/ (Col000 (C3ls/3) mlonth 1975 (m3,00) (Kwh,000) Noii[nal Real 1T Nolilal Renl -(m3.0o0) (Rwll,000) NoIinal Real 7 Non:iral ial Jar.uary n.a. n.a. n.a. n.a. x.a. n.a. 1.8 1.0 0.6 0.45 0.1S55 0.2500 February n.a. n.a. n.a. ma. n.a. n.a. 2.6 1.4 '0.5 0.63 0.1859 0.2423 March 141.3- 73.3 25.6 33.0 0.1815 0.2335 2.6 1.5 0.5 0.68 0.2095 0.2615 April 129.1 69.4 24.3 31.2 0.1882 0.2417 2.5 1.5 0.5 0.68 0.2110 0.2720 llay 122.7 68.7 24.0 30.9 0.1960 0.2518 2.7 1.6 0.6 0.72 0.2124 0.2667- Jur.e 119.6 67.8 23.7 30.5 0.1986 0.2550 2.8 1.7 0.6 0.77 0.2115 0.2750 July 123.1 71.3 24.9 32.1 0.2026 0,2608. 3.9 1.9 0.7 0.86 0.1732 0.2205 August 115.3 67.4 23.6 30.3 0.2037 0.2617 2.3 1.4 0.5 0.63 0.2097 0.2239 September 113.8 68.0 23.8 30.6 0.2120 0.2689 2.3 1.4 0.5 0.63 0.2112 0.2739 October 114.0 70.7 24.7 31.8 0.2169 0.2789 2.4 1.5 0.5 0.68 0.2108 0.2833 ove::ber 110.4 64.0 28.8 28.8 0.2609 0.2609 2.2 1.4 0.6 0.63 0.2714 0.2714 Dec,::ber 121.2 66.4 29.9 29.9 0.2466 0.2466 2.5 . 1.6 0.7 0.72 . 0.2923 0,2923 Average 0.2107 0.2560 0.2154 0.2652 C. T E'R R 0 N C 0 L 0 R A D 0 D. N AP O L ES Total Cost Unit Cost Total Cost Unit Cost Production Energy 1/ (C0l$000) (Col$a3) Production Energy !I (Col$000) (Col / -3 enth 1975 (.3,000) (Kwh,000) Nominal Real 2/ Nominal Real (m3,000) (Kdih,000) Nomini Roal2,/ Nominal ?nal Jauuary 57.4 46.2 16.2 20.8 0.2816 0.3624 51.5 25.5 8.9 11.5 0.1736 0.23 February 59.8 53.2 18.6 23.9 0.3114 0.3997 34.3 20.6 7.2 9.3 0.2102 0.11 Ma:ch 58.9 52.9 18.5 23.8 0.3i43 0.4041 30.2 18.3 6.4 8.2 0.2114 0.2715 April 56.4 50,3 17.6 22.6 0.3118 0.4007 34.4 20.9 7.3 9.4 0.2127 0.2733 May 57.4 51.0 17.8 23.0 0.3107 0.4007 36.4 21.9 7.7 9.9 0.2105 0.27^0 June 59.4 49.9 17.5 22.5 0.2940 0.3788 32,4 19.0 6.7 8.6 0.2052 0.2654 July 52.3 44.6 15.6 20.1 0.2985 0.3843 31.8 21.2 7.4 9.5 0.2333 0.2937 August 62.0 45.2 15.8 20.3 0.2552 0.3274 37.1 22.3 7.8 10.0 0.2100 0.2695 Septeher 65.4 47.9 16.8 21.6 0.2561 0.3303 38.3 23.6 8,2 10.6 0.2151 0.2768 Otober 63.7 46.2 16.2 20.8 0.2540 0.3265 42.5 25.6 9.0 11.5 0.2110 0.2706 oveaber 64.5 45.7 20.6 20.6 0.3187 0.3187 40.3 24.0 10.8 10.8 0.2677 0.2577 Lecc:-ber 62.3 46.2 19.2 19.2 0.3078 0.3078 39.2 22.9 10.3 10.3 0.2629 0.2629 Average 0.2928 0.3618 0.2136 0.2586 E. SILOE-LLERAS Total Cost Unit Cost Produccion Energy 1/ (Col.$000) Col$/n3) nonth 1975 (3,000) _Ktwh,000) Nominal Reall Nominal Reel jZnuary 54.3 34.7 12.2 15.6 0.2259 0.2873 Febr2.ry 37.0 21.1 7.4 9.5 0.1994 0.2568 .zrcn 47.9 28.7 10.2 12.9 0.2099 0.2693 41.6 16.8 5.9 .7.6 0.1410 0.1827 !-ay 58.2 2A.4 8.6 11.0 0.1471 0.1890 Notes: Same as for Table 8 June 55.2- 29.2 10.2 13.1 0.1851 0.2373 July 54.1 28.5 .9;9 12.8 0.1840 0.2366 Source: Same as for Table 8 August 52.7 29.1 10.2 13.1 0.1931 0.2486 Septe:,ber 51.2 26.5 9.3 11.9 0.1815 0.2324 Gctober 54.3 27.8 9.7 12.5 0.1794 0.2302 Noverber 52.3 27.3 12.3 12.3 0.2343 0.2352 Dece.ber 54.2 26.7 192.0 12.0 0.2213 0.2213 Éve::·age 0.1917 0.2356 - 30 - For the "La Normall" Pumping Station, which is planned to enter into operation for the High Level network in 1978, the short run marginal cost of energy is assumed to equal the mean unit pumping costs for the five secondary pumping stations discussed in the previous paragraph, i.e. Col$ 0.278 per m . Pumping at "La Normal" is expected to be required only during certain months of the year, increasing from zero in 1977 to 1 month out of the year in 1984, 2 months in 1990, and 3 months in 1995. - 31 - IV. AVERAGE INCREMENTAL COSTS OF WATER SUPPLY IN CALI The previous section presented detailed estimates of the incremental cost streams for water supply in Cali. In the present section I will translate these annual costs into a summary measure ccmmonly employed by the World Bank in its appraisals of water supply projects, the Average Incrdmental Cost (AIC). The general principles underlying this incremental cost measure were recently outlined by R. J. Saunders, J.J. l/ Warford, and P. C. Mann - and therefore need not be recapitulated here. They define AIC in period t as T AICt l (1 +i))- (1) ± Qt+t 1t t=1 (1 - i)t - where Pt = Operation and maintenance expenditure in yeal t Qt = Water produced in year t It = Capital expenditures in year t T = Number of years for which water expenditures and outputs are forecast. i = The discount rate. 1/ Robert J. Saunders, Jeremy J. Warford, and Patrick C. Mann, "The Definition and Role of Marginal Cost in Public Utility Pricing: Problems of Application in the Water Supply Sector," P. U. Report No. RES6, World Bank, July 30, 1976. -32- In words, to quote the authors: "AIC is calculated by discounting the incremental costs which will be incurred in the future to provide the estimated additional amounts of water which will be demanded over a specified period, and dividing that by the discounted value of incremental output over the period." 1/ Note that in the definitional equation of AIC any marginal cost component which remains in constant proportion relative to incremental water consumption may be factored out and added as a constant. If for instance short run marginal cost as defined in the previous section (treatment- and pumping costs) may be assued to .remain constant over the planning horizon, such that TPCt m cQt (2) where TPCt = total treatment and pumping costs in period t c = constant marginal treatment and pumping costs then equation (1) can be written as T [F(Rt,+ AICt c + (3) T t + t t t= L (1 + i) where RT = operations and maintenance costs minus TPC. t Equation (3) is useful for practical estimation purposes since it allows a simple adding on a constant short-run marginal cost component to the estimation of AIC, where the constancy assumptions is warranted, as indeed it appears to be in the case of Cali. 1! Ibid. p. 15. - 33 - In this section AIC for water supply in Cali is estimated by using a moving 15 year time horizon for every other year between 1969 and 1979, thus yielding six sequential AIC estimates for the decade. In terms of equation (1), T =1 t = 1969, 1971, 1973, 1975, 1977, 1979 Since capacity cost estimates are forecast only up to 1986, but it is known that further system expansion is required after 1986, it was necessary to assume for AIC estimation purposes that water consumption and all incremental operation and maintenance costs remain constant after that year rather than continuing to increase as projected in Tables 2 to 5 above. This is because the projected further increments in water consumption are attributable only to the further investments after 1986, rather than to the earlier investments which were costed out in the preceding section. A number of alternatiVe assumptions will be made in the estimation of AIC in order to test its sensitivity to parameter variations. The Standard Case (A) assumes no shadow pricing; Alternative 1 for operations, maintenance and overhead costs R' (i.e. they do not vary incrementally with output once capacity has been installed); a discount rate (i) of 9.7 percent is applied 1/ equally to costs and consumption figures;- and output is not lagged on costs. All alternative cases are defined as variants of the Standard Case: Case B assumes Alternative 2 for operations, maintenance, and overhead costs (i.e. they vary in proportion with output). Case C shadow prices 1/ This is the opportunity cost of capital (OCC) estimated for Colombia; see S. el Serafi, J. Linn, and L. Squire, Preliminary Report on Economic Analysis of Projects - Implementation Program FY197, World Bank, mimeo, September 1975. -3)4 - the foreign exchange component of costs at a shadow exchange rate of 1.053, Case D at 1.190. i Case E lowers the discount rate to 8 percent, while Case F raises it to 12 percent, thus bracketing the likely range of the opportunity cost of capital. Case G applies a discount rate of 9.7 percent to the cost stream, and a discount rate of 7.1 percent to the consumption stream. The latter reflects the estimated consumption rate of interest for Colombia which in pricniple must be applied to discount the consumption stream. Case H finally lags output by two years on investment expenditure, on the assumption that there is on average a two year lag between expenditure on new capacity and its use in production. This reflects the fact that construction of new capacity is generally spread out over a number of years, rather than instantaneous. The remainder of this section presents AIC estimates given these alternative sets of assumptions and treats in succession the various cost streams presented in the previous section, i.e. System costs, High Level costs, Low Level costs, Yumbo costs, and secondary pumping costs. In a final step the sum of all AIC components for each supply area will be compared with an overall estimate of AIC which does not differentiate by supply area. I/ This brackets the most likely range of shadow exchange rate estimates for Colombia; ibid. Other cost components are not shadow priced. A more refined analysis would have shadow priced labor,energy, etc. This was not attempted here given time and resource constraints. 2/ See Saunders, Warford and Mann, op. cit., p. 15. The consumption rate of interest, (CRI) is taken from El Serafi, Linn, and Squire, op. cit. A. Systems Costs Systems costs were previously defined as attributable to incremental water consumption anywhere in the Cali water supply system (see Table 2 above). Table 10 summarizes the AIC estimates over the period 1969 to 1979 and for the eight alternative sets of assumptions. Treatment and pumping costs (TPC) are not included in the AIC estimates presented in the first eight rows of Table 10, but are listed separately in the last row. From Equation 3 it may be inferred that total AIC attributable to annual increments of water anywhere in the system is the sum of any AIC estimate from the first eight rows and the constant TP.C. Beginning with the Standard Case (A) one finds that AIC (without TPC) is Col$1.177 per m3 in 1969. It drops slightly (by about 4 percent) in 1971, and in 1973 rises to a level approximately 9 percent above its 1969 value. Thereafter, AIC declines steadily, to a value of Col$ 0.175 in 1979, i.e. only 15 percent of its 1969 level. This pattern reflects the smoothened pattern of system-related investment expenditures shown in Table 2 above. Most investment takes place between 1974 and 1976, with only relatively minor spending forecast for after 1976. For Case B, where a proportionate rise in operations, maintenance, and overhead costs is forecast, AIC in 1969 is more than above the level calculated for the constant operating cost hypothesis (Case A). Further- more, after 1973 the AIC for Case B does not drop quite as drastically as in Case A. The lesson from this exercise is that it may be as important for pricing purposes to have available a reliable forecast of incremental operating and overhead costs as it is to predict correctly capital costs. -36 - Table 10: AVERAGE INCREMENTAL CCSTS OF WATER SUPPLY: SYSTEM (Col$/m3) 2/ Base Year (t) ASSUMPTIONS 1969 1971 1973 1975 1977 1979 A (Standard) 1.177 1.128 1.288 0.744 0.487 0.175 B 1.798 1.386 0.866 C 1.202 0.753 0.179 D 1.263 0.781 0.188 E E 1.132 0.691 0.165 F 1.238 0.820 0.189 G 0.951 0.781 0.147 H 1.359 0.949 TPC 0.285 0.285 0.285 0.285 0.285 0.285 I/ Blanks indicate that AIC wab not calculated. j A: Standard case, i.e. no shadow pricing, Alternative 1 for operations, maintenance, and overhead costs, equal discount rate of 9.7 percent for costs and water consumption, output is not lagged on costs. B: Alternative 2 for operations, maintenance and overhead costs. C: Foreign exchange component is shadow priced at 1.053. D: Foreign exchange component is shadow priced at 1.190 E: Discount rate is 8 percent. F: Discount rate is 12 percent. B: Discount rate for costs is 9.7 percent; for consumption 7.1 percent H: Output increment is lagged by two years after investments costs are incurred. 3/ Marginal treatment and pumping costs attributable to the system. They are not included in the AIC figures for assumptions A through H. Source:. Tables 2 and 3 -37 - In general, however, consultant and project appraisal reports largely neglect to estimate incremental operating and overhead costs. For Cases C and D, the foreign exchange component of costs is shadow priced at differentially higher rates as compared with domestic cost components. As might be expected the AIC is increased above its non-shadow price level. However, the increase is small: For 1969 it amonts to 2 percent in Case C, and to 7 percent in Case D (as compared with the Standard Case). Varying the discount rate within reasonable bounds, as was done in Cases $ and F, has also only minor impacts on the AIC. For the lower discount rate (Case E) the AIC falls as compared with the Standard Case (4 percent in 1969); for the higher discount rate (Case F) the AIC rises (by 5 percent in 1969). This is to be expected given that the incremental costs are high in early years, but low in later years, relative to incremental consumption. More significant is the reduction in the AIC resulting from attributing a lower discount rate to incremental water consumption than to incremental costs, as done in Case G. For 1969 AIC drops by 19 percent when comparing Case G with Case A. This will be the case whenever the CRI is below the 0CC, as one would expect to be the case in most developing countries wishing to accelerate economic growth above existing rates. Note,however, that this induces somewhat perverse incentives when AIC is taken as a pricing rule: With OCC>CRI, the price charged to consumers will be less than for 0CC = CRI, creating an incentive for higher rates of water consumption in a policy context where a premium is put on investment, rather than consumption. - 38 - Similarly noticeable is the increase in AIC recorded when one assumes that output increments lag investment expenditure by two years (Case H). In this case AIC for 1969 lies 15 percent above the Standard Case (A). Finally, treatment and pumping costs (TPC) are far from negligible relative to AIC under most of the alternative assumptions here considered. For the Standard Case, TPC varies from 24 percent of AIC (without TPC) in 1969 to 22 percent in 1973, and to 163 percent in 1979. Summarizing the results of this analysis of systems costs the following observations may be made for Cali: 1. AIC rises between 1969 and 1973, but declines thereafter, at least given present projections. 2. Alternative assumptions regarding the variation of operations, maintenance and overhead costs with. output have a considerable impact on AIC. A careful projection of these costs is therefore important. 3. Shadow pricing foreign exchange or varying the discount rate (for both costs and output equally) has only little effect on AIC. 4. Substituting the consumption rate of interest (CRI) for the opportunity cost of capital (OCC) in discounting consumption streams, leads to a significant reduction in AIC. 5. Specifying the exact lag structure for consumption increments resulting from particular investment expenditures may be important. Assuming output increments to be simultaneous with investment expenditure will lead to a downward bias in the estimate of AIC. 6. Short-run marginal cost (here defined as TPC) is a significant proportion of AIC, varying from 20 percent of AIC (without TPC) during the height of the investment cycle (1973), to over 150 percent during the later, lQw-investment years. B. High Level and Low Level Costs Turning then to area-specific costs of water supply in Cali, consider first the High Level costs; these are incurred on account of-the La Normal Pumping Station, and on account of incremental maintenance costs of the distribution system (under Alternative 2). AIC excluding short run marginal costs of energy used in pumping were computed for parameters of Case A, Case B, and CaseD (see Table 11). For 1969, High Level AIC are Col$0.193 per m3, given Case A assumptions, which amounts to 16.4 percent of System costs during the same year and given the same set of assumptions. High Level AIC increases steadily as the date of construction of the pumping station is approached and rises to as much as 4l.3 percent of System AIC in 1975. After construction, AIC drops off drastically, and becomes zero in 1979, since operations, maintenance and overhead costs are assumed to remain constant after 1978 (Case A). The patterns of High Level AIC for Cases B and D are quite similar, except that they lie generally above the values calculated for Case A, as is to be expected. For Case B the higher AIC is explained primarily by the assumed increases in distribution network maintenance costs and hence positive incremental operating costs. For Case C the reason for the higher AIC is found .in tI foreign exchange premium on the import. component of costs. 4 o - Table 11: AVERAGE INCREMENMAL COSTS OF WATER SUPPLY: HIGH LEVEL AND LOW LEVEL 1/ (C.o1$/ml) 2/ Base Year (t) ASSUMPTIONS 1969 1971 1973 1975 1977 1979 HIGH LEVEL A 0.193 0.196 0.296 0.307 0.046 0.000 (as % of System Costs) (16.4) (17.4) (23.0) (41.3) (9.4) (0.0) B 0.255 0.377 0.086 D 0.210 0.333 0.000 TPC/ 0.278 0.278 LOW LEVEL A 0.644 0.644 0.789 0.332 0.131 0.000 (as % of System Costs) (54.7) (57.1) (61.3) (44.6) (26.9) (0.0) B 0.739 0.41 0.139 D 0.644 0.322 0.000 1/ AIC figures for Cases A, B, and D do not include TPC. 2/ Same as for Table 10. / Energy costs for "La Normal" pumping station; pumping required only during dry season. Source: Table 3 and 9 As previously mentioned "La Normal" serves two objectives: First, it will provide supplemental water supply for the High Level network during the dry season, when the San Antonio Plant cannot provide enough capacity; second, it provides an added margin of safety for the High Level network, allowing for an alternative supply source in case the San Antonio Plant cannot be operated. The first of these two objectives would require a seasonal pricing strategy for "La Normal", while the second does not. For the calculations of AIC (net-.of short run operating costs), it has here been assumed that the -season related objective of "La Normal" can be neglected. However, in the calculation of short run marginal costs, incurred as a result of energy use in pumping during the dry season, it was assumed that these apply only during the seasonal use of "La Normal" (see Table 11). Turning then to Low Level costs, these consist of the installation of new distribution mains and the incremental maintenance costs of the distribution system (Case B). The AIC for these components are shown in Table 11. For Case A AIC-in 1969 amounts to Col$o.64h per m3, or 54.7 percent of System AIC. Low Level AIC rise until 1973, both in terms of absolute values and relative to System AIC. After 1973 they drop off, and as for High Level AIC become zero in 1969, due to the assumption of zero incremental maintenance costs in Case A. Low Level AIC for Case B is somewhat above that for Case A, reflecting the assumption of positive incremental maintenance costs for the new distribution mains and for the existing distributiQn system. Case D, introducing a premium on foreign exchange, is identical to Case A, since the foreign exchange component for 1/ Strictly speaking, the short run marginal cost ought to have been discounted back also to years before "La Normal" enters into operation. However, this aspect was neglected here as only of minor importance. Low Level works is zero. There are no additional short run marginal costs specific to the Low Level network. C. Yumbo, Siloe-Lleras, and Terron Colorado While different types of capital investment are involved in Yumbo, Siloe-Lleras and Terron Colorado, (a main for the first, and secondary pumping stations for the second and third), all three components of the investment program present a similar problem when attempting to compute the AIC for these areas. In all three cases a new area of supply is opened up during the planning period--the municipality of Yumbo, which had its own supply previously, and two low-income neighborhoods, previously without piped water . These investment projects provide capacity increments for particular segments of users, previously not hooked up to the system, and for the foreseeable future are once-and-for-all expenditures. Before the investment is made, no consumption on account of these future beneficiaries is registered and while an AIC estimate can be made, this is probably not of usefulness for pricing purposes. Once the capacity is constructed and consumption begins, past expenditures of course represent sunk cost and are no longer captured in the AIC calculation. Only incremental maintenance and operations costs will then be counted. In other words, the AIC concept cannot deal effectively with cases where entirely new segments of users are served rather than where capacity is increased to serve increased demand in areas already served. This statement is based on the understanding that AIC pricing aims at providing an indication of the opportunity cost to society of increased (water) consumption, a price signal to consumers to adjust their consumption in line with this opportunity cost; and a signal to investment planners as to when new capacity should be installed. 1 But where an entirely new area or group of consumers is served, the AIC pricing principle cannot serve any of these three goals. These conclusions are reflected in Table 12. AIC have been calculated only for those years after the works have come into operation. Yumbo drops entirely out of the picture, since it enters into the supply net only in 1980. Estimates of AIC for Siloe-Lleras and Terron Colorado are possible, however, beginning in 1975. Since after the initial investment no subsequent capital expenditures are envisaged in these areas during the planning horizon, and since incremental maintenance costs are assumed to be zero, AIC for Cases A and D are zero for all components in each year. Only for Case B, where it is assumed that maintenance costs increase with consumption, thus generating positive incremental costs, are positive AIC encountered. Relative to Systems AIC for Case B (Table 10) they remain small. Short run marginal costs of energy (TPC) required for secondary pumping in Siloe-Lleras and Terron Colorado are, however, quite substantial. While TPC for both are in the general range of magnitude as TPC for the entire system (which also includes treatment costs), TPC for Terron Colorado exceeds system TPC while TPC for Siloe-Lleras falls short of System TPC. 1/ See Saunders, Warford and Mann, op. cit. - 44 - Table 12: AVERAGE INCREMENTAL COSTS OF WATER SUPPLY: SILOE-LLERAS AND TERRON COLORADO i/ (c ol$/m3) 2/ Base Year (t) ASSUMPTIONS 1975 1977 1979 SILOE-LLERAS A 0.000 0.000 0.000 B o.o8 0.051 D 0.000 0.000 TPC 0.236 0.236 0.236 TERRON COLORADO A 0.000 0.000 0.000 B 0.064 0.069 D 0.000 0.000 TPC 0.362 0.362 0.362 1/ AIC figures for Cases A, B and C d not include TPC. 2/ Same as for Table 10. Source: Tables 4 and 9. D. Total Costs By way of contrasting the area-specific estimates of AIC with the more conventional approach of taking total projected incremental costs and relating them to total increments in consumption, Table 13 summarizes the estimates of AIC thus derived. The entries in this table are derived from applying the AIC calculus to the total cost streams in Table 5 above and the System consumption data shown in Table 2. Similar to System Costs (see Table 10) total costs increase between 1970 and 1974 and decrease thereafter. Furthermore, testing for two of the major alternative assumptions (Alternative 2 for Operations, maintenance and overhead costs, and shadow pricing of foreign exchange, respectively ) similar impacts are found as shown earlier for the System costs. More of interest, however, is a comparison of AIC in particular areas, when summing area specific AICs with System costs for each area, to determine the bia8 which is introduced when approximating such area-specific AICs by estimating, as conventional, only one summary AIC for total incremental cost as in Table 13. 1 Table 14 shows that for the first three periods (1969, 1971 and 1973) Low Level AIC differ only insignificantly from Total Cost AIC. In contrast, High Level AIC lie by about 25 percent below Total Cost AIC (and thus Low Level AIC). The main reason for this divergence is the fact that the cost of new mains in the Low Level area are not attributed to High Level consumers in the calculation of area specific costs. For the last three 1/ Short run marginal costs are compared separately further below and are not included in the AIC estimates discussed here. Table 13: AVERAGE INCREKENTAL COST CF WATER SUPPLY: TOTAL COSTS (col$/m3) 1/ Base Year (t) Assumptions 1969 1971 1973 1975 1977 1979 A 1.854 1.776 2.o3 1.177 0.697 O.2o B 2.535 1.961 1.ol D 1.966 1.248 0.217 1/ Same as in Table 10 Source: Tables 2 and 5 Table 14: AVERAGE INCREMENTAL COST OF WATER SUPPLY: COMPARISON OF AREA SPECIFIC COSTS WITH TOTAL COSTS (STANDARD CASE ASSUMPTIONS) 1 (Col$/m3) Base Year (t) Area 1969 1971 1973 1975 1977 1979 High Level 1.370 1.324 1.584 1.051 0.33 -0,175 (% of Total) (73.9) (74.5) (77.2) (89.3) (76.6) -(85.4) Low Level 1.821 1.772 2.077 1.076 0.618 0.175 (4 of Total) (98.2) (99.8) (101.2) (91.4) (88.7) (85.4) Siloe-Lleras l.o1 0.33 0175 ( of total) (89.3) (76.5) (85.4) Terron Colorado l.ol 0.33 0.175 ( of Total) (89.3) (765) (85.4) 1/ Not including short run marginal costs (TPC) Source: Table 10 and 13 -47 - periods (1975, 1977 and 1979) High and Low Level AIC approach each other and are in fact identical for 1979. They both fall short of Total Cost AIC since the latter incorporates the costs attributable to Yumbo, Siloe- Lleras and Terron Colorado. Indeed, this factor also explains why in the earlier years Total Cost AIC do not in general fall between High and Low Level AIC. As was argued earlier it is not appropriate to calculate area specific AIC for Yumbo, Siloe-Lleras, and Terron Colorado for the years before the operation of these components. But this is precisely what is being done implicitly in calculating the Total Cost AIC. The failure to analyze separately the AIC for individual areas therefore leads not only to an averaging of incremental costs across areas, but also appears to lead to a failure to exclude costs which should not be incorporated into the analysis in the first place. Short-run marginal costs (TPC) differ by area mainly because of supplementary pumping requirements in the higher-lying areas of the city, especially such areas of the city, as Siloe-Lleras and Terron Colorado. These cost differentials are summarized in Table 15. Furthermore, as the "La Normal" pumping station will increasingly be required to transfer water from the Low to the High Level network in future years, a higher short run marginal costs will become more prevalent in the latter area, particularly during the dry season. Combining finally our short run marginal cost estimates with the earlier AIC cost estimates we may derive total AIC for each area and compare them in order to determine the spatial cost variations for water supply in Cali-(Table 16). Taking High Level AIC (including short run marginal costs, Table 15: SHORT RUN MARGINAL COSTS BY AREA (C ol$/m3) SHORT -RUN AREA MARGINAL COST High Level 0.285 (+ 0.278) 1/ Low Level 0.285 Siloe-Lleras 0.521 (+ 0.278) 1 Terron Colorado o.647 (+ 0.278) 1/ 1/ Short run marginal cost limited to dry season when supplementary pumping at "La Normal" is required, applicable only from 1977. Source: Tables 10 and 13 44 Table 16: AVERAGE INCREMENTAL COST OF WATER SUPPLY: SUMMARY (Co1$/m3) BASE YEAR (t) AREA 1969 1971 1973 1975 1977 1979 High Level 1.655 1.609 1.869 1.336 0.818 0.460 Low Level 2.106 2.057 2.362 1.361 0.903 0.460 (% of High Level) (127.3) (127.8) (126.4) (101.9) (110.4) (100.0) Siloe-Lleras 1.572 1.054 0.696 (% of High Level) (117.7) (128.9) (151.3) Terron Colorado 1.698 1.080 0.622 (% of High Level) (127.1) (132.0) (178.7) 1/ Including short run marginal costs, except for seasonally higher short run maiginal costs in High Level areas. Source: Tables 14 and 15 ..).. . 5o with the exception of the seasonally higher short run costs due to pumping at "La Normal") as the basis of comparison, it is found that during the early Seventies Low Level costs exceeded high level costs by just over 25 percent due to higher investment activities attributable to low area subscribers. This difference virtually disappers in the mid-and late seventies. During this latter period, however, the higher short run marginal costs due to secondary pumping in the high lying areas lead to a growing relative difference between AIC for these peripheral areas and the rest of the city. In 1979, AIC in Siloe-Lleras and Terron Colorado exceed High and Low Level Level AIC by more than 50 percent and 7 percent respectively. In concluding this section is is of interest to observe that over the whole period of 1969 to 1979 AIC initially rise and then decline. Unfortunately, however, the period is too short to determine whether in the long run AIC of water supply in Cali show a rising or falling trend. Only if it were possible to cost out and incorporate system increments substantially beyond 1986 could a reliable judgment on this issue be made.