Water allocation under alternative water supply conditions

Soeio-Econ. Plan. Sci. Vol. 16, No. 5, pp. 217-221, 1982 00384)121[82/050217415503.0010 Printed in Great Britain Pergamon Press Ltd.

WATER ALLOCATION UNDER ALTERNATIVE WATER SUPPLY CONDITIONS

MARK S. HENRY and ERNEST J. BOWEN Department of Agricultural Economics and Rural Sociology, Clemson University, Clemson, SC 2%31, U.S.A.

(Received 10 January 1982)

Abstract--The purpose of this paper is to develop a method for estimating the marginal value of water to alternative users. The framework developed employs input-output and linear programming methods and is applied to a rural area of South Carolina. Pricing water resources according to their value of marginal product will lead to efficient allocation of the water and maximum Gross Regional Product under alternative water supply conditions.

Early research into the question of imputing a monetary value to water resources in a region estimated the value of additional units of water to the entire regional economy[I,2]. These studies employed a combined input-output, linear programming (IOLP) framework so that regional exports and final demand sales and interindustry requirements were all considered in estimating the demand for water resources. Of course, as water supplies dwindled and/or final demand deliveries increased, the IOLP solution yielded rising shadow prices for water resources[l]. These results are interesting and useful to decision makers concerned with providing additional water to users. In a cost benefit framework, estimation of the benefits stream of additional supplies of water was greatly improved relative to partial equilibrium methods.

For the past decade, water supplies have become increasingly limited in parts of the United States while irrigation has increased and population pressures in- tensified. With these trends, intersectoral allocation of water has become a hotly debated issue[3, 4]. The early empirical IOLP research yields little insight into the basis for arriving at optimal water resource allocation under realistic scenarios because estimates of the value of water to alternative sectors is absent.

This study will show an input-output model and linear programming can be combined to create a useful tool for water resource planning. The need for such a tool, capable of determining the marginal value of water to different uses, stems from the lack of a viable market mechanism for water. With regional growth and increased agricultural demands for water a reasonable likelihood, competing demands on water could conceiv- ably result in gross inefficiencies and suboptimal incomes for the region. Huffman argued that since the established market structure does not provide an adequate means of apportioning and regulating control and use of water on an equitable basis, water rights will become more and more a result of rationing transactions by units of government. While some critics[5] contend that the proper approach to solving our water problems is by means of a system of explicit ownership rather than by a system of government allocations, political and social difficulties render their proposal to a status of long-term institutional evolution. Moreover, there are very few a priori grounds for judging how water transfers increase or decrease regional income. It is obviously an empirical

question for specific situations. The model presented in this study, therefore, offers an empirical basis upon which to establish an allocation scheme. Fox and others argue that general equilibrium models are preferable to partial equilibrium analysis in water resources questions. Henry and Bowen[6] further argue that IO models are inadequate for establishing the marginal value of water between sectors of a regional economy.

Our objective is to demonstrate how the value of water to each sector of a regional economy may be estimated using the IOLP approach. It will be demon- strated for a case study that the value of additional units of water to a region varies between sectors and the sector valuations vary considerably with water availability conditions and the irrigation practices in the region. The sector valuations can provide the basis for a pricing mechanism for water resources that reflects the sector value of marginal product (VMP) of the water resource. VMP pricing of water yields efficient water use within the region; i.e. it allows maximum Gross Regional Product given the resources constraints, interindustry linkage and final demand patterns.

MODEL STRUCTURE The model employed in this study consists of a non-

survey regional input-output model (64 sectors) and a linear programming framework (IOLP) which may be written as:

Maximize the objective functions:

Z = CX (1)

subject to

~" >- ( I - A ) X >- y (2)

wiXi >- Wo (3) i=1

~"~ liXi -> Lo (4) i = l

where i (or / )= number of the sector; n = number of sectors; Z = Gross Regional Product; C = value added coefficients vector (dimension 1 × n); X = gross output vector (dimension n x 1); ( l -A)= I is the identity matrix; A is the technical coefficient matrix; Y = final

217

218 M. S. HENRY and E. J. BOWEN

demand vector (dimension n x 1), current level; I;" = final demand vector (dimension n × 1), projected level; w~ = sectoral water intake per dollar of gross output; Wo = total water availability; l~ = sectoral labor requirements per dollar of gross outputs; Lo = total labor availability.

The IOLP framework may be used to solve for the optimal allocation of activity between alternative sectors of the regional economy in several different ways. First, by specifying the Y vector, Wo and Lo constraints, a base level of activity (X) is established. By next allowing the Lo constraint to increase, the desired level of new activity may be estimated for each sector in order to maximize gross regional product (GRP).

Second, the marginal value of additional water supplies to the region can be assessed by varying the Wo parameter and solving for the shadow price of water. Finally, the relative marginal value of allocating water to alternative sectors may be estimated by comparing the optimal output levels of each of the 64 sectors of the model under different levels of Wo.

This IOLP model description differs from the earlier studies by imposing a constraint that final demand (FD) sales must be less than or equal to some prespecified level, Y. This amounts to assuming that the growth in final demand sales for a given sector has an upper limit. These constraints are akin to sectoral capacity constraints that exist in the short run but of course are not recognized in the traditional input-output model.

Prior to using the model results for policy making, the analyst would be required to estimate the actual limits to final demand sales that exist for each of the sectors in the re#on. These could be either production capacity or demand side limits. For our purpose, we have chosen a rather arbitrary FD limit of 10% annual growth in real final demand sales. This procedure effectively limits expansion in any given sector even if that sector has a low water use requirement and is a very high contributor to regional value added. The Y limit has important consequences for the model solution and in an actual policy analysis, the 1 ? limit wbuld have to be determined with considerable care.t In sum, the value of water between sectors is determined over the range of sector output that is obtained from the need to deliver between Y and Y levels of final demand.

A CASE STUDY OF CENTRAL SOUTH CAROLINA A multicounty region in central South Carolina was

used as the study area. The IO model was constructed using the 1972 United States model updated to 1980 prices and modified by the simple location quotient approach. (See Mulkey-Hite[7] for a complete description of the non-survey technique used.) Regional sectoral exports were estimated as the residual between in-region requirements and total output by sector. The IO model is closed with respect to households and government. Ac- cordingly, the export vector is used as the estimated Y vector in the IOLP model. The wl coefficients were obtained from a survey of all manufacturing firms in the region[8], published data on agricultural output[9, 10], a survey of South Carolina farm irrigation practices [l l], and from United States Geological Survey data[12] for

?Of course, the forecast of the Y vector is the critical issue in forecasting the overall level of economic activity in the re#on using standard IO techniques.

most other sectors. Wo and Lo were obtained from the South Carolina Water Resources Commission[13] and the Employment Security Commission[14], respectively.

EMPIRICAL RESULTS

To demonstrate the usefulness of the IOLP framework, alternative water use coefficients (wi) and water supply conditions (Wo) are used to estimate the marginal value of water under each scenario. For each scenario, the model is solved for a baseline (Wo) optimal output distribution, then solved for the case where 1000 acre feet of water are added to the Wo restriction, Wo = Wo+ 1000 acre feet.

Low agricultural water demand scenarios With the assumption of constant direct water use

coefficients, we are not allowing the agricultural sectors to increase their use of irrigation. Under this condition, water becomes a constraint of GRP growth at about the 45,000 acre foot level of water availability where the shadow price of water is $700 per acre foot. The marginal value of an acre foot of water to GRP is constant over a long range (17,200 acre feet to 46,000 acre feet). If water availability falls below the 17,200 level the shadow price varies from $8,000 to $37,000 per acre foot at the 16,200 level. The current surface water capacity (systems in place) is about 30% above the level where water becomes an effective constraint. Thus, for this aggregate region, on average, water availability is not currently a serious problem.

We solved for optimal sector output levels (to maximize GRP) under four alternative Wo constraints: 45,148; 44,148; 17,200 and 16,200 acre feet. Data in Table 1 show how the relative value of water between sectors changes as water becomes increasingly scarce.

The results in Table 1 provide a basis for valuing water among competing users. The AQt column gives the desired increase in output for each sector in order to maximize GRP as the level of water available increases by 1000 acre-feet. To obtain this result water could be allocated to each of the 64 sectors by using the w, coefficients and the X~ value from the AQI column. Similarly, when the level of water available falls to 16,200 acre-feet and an additional 1000 acre feet becomes available, AQ2XI values could be used in the same way.

This analysis shows that when water availability is increased from 44,148 acre-feet to 45,148 acre-feet, several of the agricultural sectors are primary beneficiaries. The top ranking nonhousehold sectors are livestock and livestock products (sector 6), feed grains (sector 3), trade (sector 49), and miscellaneous food and allied products (sector 22).

The AQ2 column presents the results of a situation with exceptionally low water supply. In this situation the water available is set at 17,200 acre-feet with the shadow price of $7,934. Reducing the supply by a thousand acre-feet to 16,200, the shadow price of an additional unit of water increases sharply to $37,471. The nonhousehold sectors with the largest desired output changes are fabricated metals manufacturing (sector 40), miscellaneous food and kindred products manufacturing (sector 22), other food and kindred products" (sector 23), and wholesale and retail trade (sector 49). It is interesting to note that at this level of water availability, there are no agricultural sectors represented in the top five sectors. However, oil bearing crops (sector 4) improved its rank from 27th (Rank AQ1 column) to 7th. Cotton (sector 1)

Water allocation under alternative water supply conditions 219

also improved its rank position (42th to 14th) while livestock (sector 6) falls from 1st to 55th.

For the agricultural sectors, given the interdependen- cies within ths local economy and current water use coefficients, the major implications of the analysis are that water resources will yield the highest GRP, in times of severe drought, if soybean production increases relative to livestock production. Of course, changing irrigation practices and the composition of the manufacturing sector (high versus low water intensive sectors) can affect the water use coefficient and the resulting ranking of each sector.

High agricultural water demand We investigated the sensitivity of the model results to

changes in irrigation patterns by estimating a second set of wl coefficients for the agricultural sectors. We assumed that all farms would irrigate using water at the same rate as farms currently irrigating in the region.

The model was run for the case of low water supply with high agricultural water demand. The results reflect a situation where there was universal adoption of irrigation with no expansion of current water systems. Using a base supply of 66,628 acre-feet, a reduction to 65,628 acre-feet borders on infeasibility. The model results in-

Table 1. Low agricultural water demand scenarios

$ Rank $ Sector IO Name AQI AQI hq2

1 COTTON PRODUCTION 821 40 616824 2 FOOD GRaIN PRODUCTION 5804 19 174762 3 FEED GRAIN PRODUCTION 354583 3 1198069 4 OIL BEARING CROPS 2887 27 3931222 5 TOBACCO 0 55 0 6 LIVESTOCK & LIVESTOCK PROD. 1204406 I 0 7 OTHER AGRIC. PRODUCTION 2774 29 424833 8 FORESTRY 1138 41 37074 9 FISHERIES 0 55 0

i0 AGRIC.,FORESTRY & FISHERIES 52364 7 305225 ii IRON & FERROALLOY ORES MINES 0 55 0 12 NONFEROU8 METALS MINING 0 55 0 13 COAL 0 55 0 14 CRUDE PETRO & NAT'L GAS EXT. 0 55 0 15 STONE & CLAY MINING 1797 36 29454 16 CHEM. & FERTILIZER MINING 0 55 0 17 NEW CONSTRUCTION 0 55 0 18 MAINTENANCE & REPAIR COHST. 0 55 0 19 ORDINANCE & ACCESSORIES MFG. 0 55 0 20 GRAIN MILL PROD. MFG. 8403 16 413115 21 BAKING PROD. MEG. 150 44 6668 22 MISC FOOD & KINDRED PROD MEG 98693 5 9915086 23 OTHER FOOD & KINDRED PROD MEG 4738 21 8049647 24 TOBACCO 0 55 0 25 BROAD-NARROW FABRI CS-YARN-TD 6683 18 6789202 26 MISC TEXTILE GOODS & FLOOR COV 1249 39 54338 27 APPAREL 19313 ii 788133 28 MISC FABRICATED TEXTILES 1666 37 75849 29 LUMBER-WOOD PROD MEG. 2465 31 286590 30 FURNITURE & FIXTURES MFG. 4223 25 168144 31 PAPER & ALLIED PRODS. 0 55 0 32 PRINTING & PUBLISHING 2073 33 135755 33 CHEMICALS MEG. 9227 14 400232 34 PLASTICS & SYNTHETICS MFG. 4573 23 2836476 35 PETRO REFINING & RELATED PRD 1166 40 28827 36 RUBBER & MISC. PLASTICS pROD 2631 30 393737 37 LEATHER, TANNING, ETC. 0 55 0 38 GLASS, STONE b CLAY PRODS. MEG. 2027 35 276383 39 PRIMARY METALS MFG. 130 45 183141 40 FABRICATED METALS MEG. 6813 17 22768171 41 MACHINERY, EXCEPT ELECTRICAL 356 43 22683 42 ELECTRICAL MACHINERY MEG. 13245 13 629233 43 TRANSPORTATION EQUIP. MFG. 2131 32 82900

44 SCIENTIFIC INSTRUMENTS MEG. 2800 28 157327 45 MISCELLANEOUS MFG. 1378 38 56972 46 TRANSPORTATION & WAREHOUSING 8813 15 376550 47 CO}~4UNI CATIONS 14700 12 539052 48 UTILITIES 32574 9 1239576 49 WHOLESALE & RETAIL TRADE 151329 4 5751699 50 FINANCE INSURANCE REAL EST. 45472 8 1356679 51 HOTELS, LODGING PLACES 21822 I0 1003754 52 BUSINESS SERVICES 5492 20 328877 53 RESEARCH & DEVELOPMENT 0 55 0 54 AUTO REPAIR 4679 22 161296 55 AMUSEMENTS 2967 26 118029 56 MEDICAL, EDUC., NON PROFIT SER 72738 6 2618101 57 FEDERAL GOV'T ENTERPRISES 4487 24 208486 58 STATE b LOCAL GOV'T ENTERPR. 2043 34 84967 59 GROSS IMPORTS 0 55 0 60 DU~ ENTERPRISES 0 55 0 61 OOV'T (GENERAL) 0 55 0 62 REST OF WORLD 0 55 0 63 HOUSEHOLD INDUSTRIES 0 55 0 64 CONSUMERS 650149 2 25726004

X~o- X~o where W o = 44,148 acre feet, W O = 45,148 acre feet

AQ i - X~o- ~ where W o - 16 ,200 . . . . f e e t , W o ffi 17 ,200 . . . . f e e t o

Rank ~Q7

16 29 12

7 55 55 18 4o~ 55 24 55 55 55 55 55 55 55 55 55 19 44

3 4

55 5

39 14 37 25 30 55 33 20

8 42 21 55 26 28

2 43 15 36

32 38 22 17 i i

6 i o 13 23 55 31 34

9 27 35 55 55 55 55 55

1

220 M.S. HENRY and E.J. BOWEN

dicate that for the agricultural sectors, cotton production and livestock production should be drastically reduced.

Data in Table 2 indicate that the highest ranking manufacturing sectors are Miscellaneous Food Products (sector 22), Apparel (sector 27), Electrical Machinery (sector 42), Broad and Narrow Fabric (sector 25), Fur- niture and Fixtures (sector 30) and Plastics and Syn- thetics (sector 34). All of these manufacturing sectors, and three others, 32, 33, and 36, all rank higher than the highest agricultural sector, Oil-Bearing Crops (sector 4). Under this situation, the highest ranking sectors are in the trade and service sectors. Their low water use

coefficients indicate that they use little water in relatiov to the value of their output. With water at such limitin~ levels, encouragement of tlaese sectors would be highly beneficial in increasing gross regional product.

DATA PROBLEMS

The IOLP offers a basis upon which to analyze water supply and distribution issues. It is not, however, without major shortcomings. The most important of these is the problem of data. The construction of an input-output model requires massive amounts of data. The nonsurvey

Table 2. High agricultural water demand

$ Rank $ Rank Sector I0 Name AQI AQI AQ 2 AQ 2

I COTTON PRODUCTION 79 40 0 51 2 FOOD GRAIN PRODUCTION 134365 i 7866 38 3 FEED GRAIN PRODUCTION 1004 15 71593 34 4 OIL BEARING CROPS 203 33 356070 24 5 TOBACCO 0 55 0 51 6 LIVESTOCK & LIVESTOCK PROD. 3234 7 0 51 7 OTHER AGRIC. PRODUCTION 195 34 343943 25 8 FORESTRY 92 39 137081 30 9 FISHERIES 0 55 0 51

i0 AGRIC., FORESTRY & FISHERIES 2143 i0 102327 32 ii IRON & FERROALLOY ORES MINES 0 55 0 51 12 NONFEROUR METALS MINING 0 55 0 51 13 COAL 0 55 0 51 14 CRUDE PETRO & NAT'L GAS EXT. 0 55 0 51 15 STONE b CLAY MINING 308 26 18838 35 16 CHEM. & FERTILIZER MINING 0 55 0 51 17 NEW CONSTRUCTION 0 55 0 51 18 MAINTENANCE & REPAIR CONST. 0 55 0 51 19 ORDINANCE & ACCESSORIES MFG. 0 55 0 51 20 GRAIN MILL PROD. MFG. 42 42 0 51 21 BAKING PROD. MFG. 16 44 0 51 22 MISC FOOD & KINDRED PROD MFG. 7172 4 12999539 4 23 OTHER FOOD & KINDRED pROD MFG. 495 20 0 51 24 TOBACCO 0 55 0 51 25 BROAD-NARROW FABRICS-YANN-TD 709 16 1318305 ii 26 MISC TEXTILE GOODS & FLOOR COV 73 41 118318 31 27 APPAREL 2109 ii 3953039 7 28 MISC FABRICATED TEXTILES 173 36 313952 27 29 LUMBER-WOOD PROD MFG. 215 31 337040 26 30 FURNITURE & FIXTURES MFG. 463 22 865692 12 31 PAPER & ALLIED PRODS. 0 55 0 51 32 PRINTING & PUBISHING 211 32 361197 23 33 CHEMICALS MFG. 1258 14 476733 19 34 PLASTICS & SYNTHETICS MFG. 492 21 861939 13 35 PETRO REFINING & RELATED PRD 142 38 90869 33 36 RUBBER & MISC. PLASTICS PROD 255 28 423703 20 37 LEATHER, TANNING, ETC. 0 55 0 51 38 GLASS, STORE & CLAY PRODS. MFG. 175 35 290284 28 39 PRIMARY HETALS MFG. 12 45 12157 37 40 FABRICATED METALS MFG. 511 19 0 51 41 MACHINERY, EXCEPT ELECTRICAL 31 43 12383 36 42 ELECTRICAL MACHINERY ~fFG. 1413 12 2431911 9 43 TRANSPORTATION EQUIP. MFG. 227 29 413590 21

44 SCIENTIFIC INSTRL~NTS MFG. 293 27 529213 17 45 MISCELLANEOUS MFG. 148 37 269543 29 46 TRANSPORTATION & WAREHOUSING 614 17 737320 15 47 COF~I/NICATIONS 1378 13 2060415 i0 48 UTILITIES 2616 8 3735658 8 49 WHOLESALE & RETAIL TRADE 13475 3 21505309 2 50 FINANCE INSURANCE REAL EST. 6309 6 5277362 5 51 HOTELS, LODGING PLACES 2326 9 4251696 6 52 BUSINESS SERVICES 540 18 486871 18 53 RESEARCH & DEVELOPMENT 0 55 0 51 54 AUTO REPAIR 428 24 637926 16 55 AMUSEMENTS 319 25 0 51 56 MEDICAL, EDUC., NON PROFIT SER 7092 5 13135481 3 57 FEDERAL GOV'T ENTERPRISES 451 23 763479 14 58 STATE & LOCAL GOV'T ENTERPR. 218 30 399181 22 59 GROSS IMPORTS 0 55 0 51 60 DUMMY INTERPRISES 0 55 0 51 61 GOV'T (GENERAL) 0 55 0 51 62 REST OF WORLD 0 55 0 51 63 HOUSEHOLD INDUSTRIES 0 55 0 51 64 CONSUMERS 71258 2 133732376 1

i I i AQI = XWo- ~o where W o ffi 135,521 W o ffi 134,521

i i i AQI ffi XWo- ~o where W o = 66,268 Wo ffi 65,268

Water allocation under alternative water supply conditions 221

technique used in creating the regional model fortunately provided an acceptable alternative. However, the large degree of indexing required to update the model was far from ideal.

Water use data are also not readily available for all sectors. The water used in agriculture may, in fact, be slightly underestimated since farmers who have irrigation are likely to overwater once the system is in operation. As the need for improved data on water use is realized by both manufacturing and agriculture and advances are made in this area, economic analysis relating to water will become more reliable and precise.

CONCLUSIONS

Assuming that there is no change from current irrigation practices in the region, we can draw several conclusions from our results regarding the value of water within agriculture, and between agriculture and other sectors. Looking at agriculture first, we find that with adequate water supplies, the livestock and feed grains sectors would make the greatest contribution to GRP from additional water supplies in the region. However, if water is in short supply, then livestock is no longer desirable, and among the agricultural sectors a shift to soybeans would most improve the region's GRP. There are two reasons for this shift. First, livestock contributes more to GRP directly and indirectly than soybeans when water is not an important limiting factor. Second, livestock requires more direct and indirect water inputs per dollar of final sales than soybeans and as water becomes increasingly scarce, the water niggardliness facet of soybean production becomes more important to in- cremental GRP growth than the larger value added aspects of livestock production.

Looking next at the competition for water between agricultural sectors and other sectors, we find that with adequate water supplies, livestock and feed grains are the top ranking nonhousehold sectors. Accordingly, additional water allocations to these agricultural sectors are clearly justified from a GRP perspective given current irrigation practices in the region. Even in periods of short water supply, the soybean sector and feed grain sectors rank among the top 15 sectors in contribution to GRP from additional water resources availability. Thus, if water resources become scarce in the region, agriculture can continue to compete effectively with other sectors for water resources by reallocating their resources into oil bearing crops and away from livestock.

If we next assume that irrigation becomes pervasive in the region, then agriculture no longer competes as effectively with manufacturing for water resources under conditions of water scarcity. Data in Tables 1 and 2 imply that the relative value of water to the region in alternative use depends on the level of water available. The IOLP model yields appropriate measures of relative water value between sectors under conditions where the water shadow price is nonzero. These conditions may prevail in the face of short-term drought conditions or as future growth increases the aggregate demand for water resources.

REFERENCES

1. E. M. Lofting and P. H. McGauhey, Economic Evaluation o[ Water, Part III: An lnterindustry Analysis o[ the Cali[ornia Water Economy. Water Resources Center Contribution No. 67. Berkeley: Sanitary Engineering Research Laboratory, University of California (1%3).

2. E. M. Lofting, and P. H. McGauhey, Economic Evaluation of Water, Part IV: An Input-Output Linear Programming Analysis o[ Cal[ornia Requirements. Water Resources Con- tribution No. 116. Berkeley: Sanitary Engineering Research Laboratory, University of California (1968).

3. C. Kiker, Water allocation in the east. Econ. Rev. Federal Reserve Bank of Atlanta pp. 27-31 (1981).

4. C. Kiker and G. D. Lynne, Water allocations under ad- ministrative regulations: some economic considerations. Southern J. Agricultural Econ. 8, 57-63 (1976).

5. J. H. Dales, Land water and ownership. Economics of the Environment (Edited by R. and N. Dorfman). W. W. Norton and Company, (1972).

6. M. Henry and E. Bowen, A Method for estimating the value of water between sectors of a regional economy. Southern J.. Agricultural Econ. 13 (2), (1981).

7. D. Mulkey and J. Hite, A Procedure [or estimating inter- regional input-output matrices from secondary data. Tech. Bull. 1072, South Carolina Argicultural Experiment Station, March 1979.

8. South Carolina Department of Labor, Unpublished annual wage and salary survey of manufacturing firms in South Carolina, Columbia, South Carolina, 1977.

9. South Carolina Crop Statistics. AE 413, South Carolina Crop and Livestock Reporting Service, Agricultural Experiment Station, Clemson University, Clemson, South Carolina, June 1980.

10. South Carolina Livestock and Poultry Statistics. AE 414, South Carolina Crop and Livestock Reporting Service, Agricultural Experiment Station, Clemson University, Clemson, South Carolina, June 1980.

11. Clemson University, Department of Agricultural Engineer- ing, Unpublished survey. Irrigation by Crop and County in South Carolina, 1980.

12. U.S. Department of Interior, Geological Survey. 13. South Carolina Water Resource Commission, Unpublished

water resource maps for Claredon, Kershaw, Lee, and Sum- ter Counties. Columbia, South Carolina, 1980.

14. South Carolina Employment Security Commission, Man- power In Industry. Columbia, South Carolina, 1980.

15. K. I. Fox, We can solve our water problems. Water Resources Res. 2, 617-623 (1966).

16. C. W. Howe, Broad horizons in water resourses planning and investment. Water Resources Res. 2, 843-847 (1%6).

17. R. E. Huffman, Irrigation Development and Public Water Policy. The Ronald Press Company, New York (1953).

18. M. M. Kelso, W. E. Martin and L. F. Mack, Water Supplks and Economic Growth in an arid Environment: An Arizona Case Study. University of Arizona Press, Tuscon (1973).

19. H. H. Stoevener and E. N. Castle, Input--output models and benefit cost analysis in water resources research. J. Farm Econ. 47, 1572-1579 (1%7).

20. A. G. Tijoriwala, W. E. Martin and L. G. Bower, The Structure of the Arizona economy: output interrelationships and their effects on water and labor requirement, Part I. Arizona Agricultural Experiment Station Tech. Bull. 180. University of Arizoa press, 1%8

21. South Carolina Cash Receipts From Farm Marketing. AE 416, South Carolina Crop and Livestock Reporting Service, Agricultural Experiment Station, Clemson University, Clemson, South Carolina, September 1980.

Documents

Water allocation under alternative water supply conditions