By Charles W. Howe
With the assistance of D. Jay Goodman, Graduate Student in Economics
INTRODUCTION
Background on the Drought Issue
Reliability, along with water quality and cost, is a key feature of urban water supply. In extreme cases found in the Third World, supply failure may result in infiltration of polluted groundwater into the supply system with dire health consequences. In the industrialized countries, water supplies of high reliability and quality are taken for granted, but events like weather induced supply reductions and demand increases can result in water rationing, mandatory conservation, and sometimes in pressure drops with consequent inability to provide adequate fire flows or service to higher elevation areas.
Water system planning and management consists of two parts: (1) the long-term planning of raw water supply, treatment and distribution capacities to serve users' demands adequately; and (2) the management of short-term shortage events (usually occasioned by droughts) when they occur to avoid unnecessary social costs. Rational long-term raw water and capacity planning subsumes a rational approach to (2) since poor handling of shortages when they occur can appear to justify unneeded long-term capacity investments.
Urban water utilities are beginning to use explicit risk-based decision procedures for both long-term planning and short-term shortage (drought) management (Moreau and Little, 1989), but the majority still use outmoded traditional planning based on meeting the needs of their "drought of record" and are frequently inadequately prepared to deal with drought when it occurs (Moreau and Little, pp. 113-115). The key question is whether or not water users would be willing to pay at least the needed storage and capacity costs to avoid the consequences of a specified shortage event.
The classic study by Russell, Arey, and Kates (1970) was the first to try to estimate the losses to public and private parties from drought (the New England drought of the mid 1960s). That study measured the dollar value of observable losses to water users: garden and lawn losses to residential customers; commercial and industrial losses in the form of lost profits and additional drought-related expenditures (e.g., water recirculation systems); and emergency expenditures of public agencies (e.g., for new supplies). They concluded that the measurable losses to the studied supply systems were surprisingly small even for that severe drought that had an estimated recurrence interval of 150 years and that, for many urban systems, the balancing of mathematically expected drought-related losses against additional expenditures on supply capacity could result in substantial system savings.
While risk-related system standards and drought strategies have been adopted by some urban systems, these strategies have never been based on objective measures of shortage-related losses. Boulder, Colorado has established the following policy regarding its raw water availability (City of Boulder, 1988):
The City shall strive to balance the costs of increased reliability and the consequences of temporary water shortages. It is recognized that no municipal system can ever be 100% reliable against all risk facts and that the economic and environmental opportunity costs of reducing the risks of occasional water shortages are significant.
Having said that, Boulder proceeded to set exceedingly stringent standards of reliability:
One reason Boulder may have chosen such high reliability standards is that the City had obtained extensive water rights at low prices in earlier years, water rights with a total reliable yield far in excess of forecasted water demand, even under drought conditions. However, no loss estimates were used in arriving at these standards. The probability of a shortage severe enough to require lawn watering restrictions for the months of July, August and September is projected to be 1/300 under conditions of full build-out for the service area (Howe et al, 1991).
Other cities have set risk or reliability standards, but usually as trigger criteria for drought-related conservation measures. For example, the Orange Water and Sewer Authority of North Carolina has set the following trigger levels for the purchase of water from other systems (Moreau and Little, 1988, p. 56): (a) when the probability of storage depletion over the remainder of the year exceeds 1/60; or (b) the probability of having to invoke any form of conservation for more than four weeks over the remainder of the year exceeds 3/60; or (c) the probability of storage being less than 75% full on April 1 of the following year exceeds 2/60; or (d) the probability of having to purchase water from other entities for more than 10 weeks exceeds 6/60.
Moreau notes similar drought action trigger levels for Durham, NC, Seattle, WA, and the Washington Suburban Sanitary Commission (WSSC) of the Washington, DC metropolitan area. The schedule adopted by WSSC nicely illustrates the definition of "standard shortage events" namely shortages sufficient to require:
Event 1 that restrictions on outdoor use would be in effect for 30 days or less during a year.
Event 2 that air conditioning restrictions and swimming pool closures would be in effect for 30 days or less.
Event 3 that sequentially more severe restrictions would be in effect for 30 days or less in the coming year (Moreau and Little, 1988, p. 65).
System reliability can be defined in a simplified manner by determining the probabilities of one or more of these events under current or projected system conditions -- a strategy followed in the present study. It must be noted that none of these systems of drought action trigger levels was based on an assessment of actual losses that would result from the occurrence of a shortage event.
The 1976-77 drought was particularly severe in Northern California, especially in Marin County. There is no question that the costs of that event were quite high -- in the forms of emergency supplies, the inconveniences of severe rationing schemes, and commercial and residential losses (see Nelson, 1979). Unfortunately, no one essayed to estimate the inconveniences to water users in terms of a willingness to pay to avoid recurrence of such events.
Determining Optimal Supply System Reliability, R*
The uniqueness of water supply system management comes from the significant levels of randomness that attach to both supply and demand. The hydrology and hydraulics of water supply are well known and widely utilized in system design (e.g., Goodman, 1984). The stochastic components of demand have been much less studied (for an exception, however, see Maidment, et al, 1985) and even less frequently applied to system design. Over time, the interactions of stochastic supply (S) and demand (D) result in a probability distribution of shortages, i.e. a distribution of excess demand X = (D-S) that is defined as
The discrete-continuous density function of X would appear as in Figure 1 below.
Figure 1. The Probability Distribution of Excess Demand.
If (D,S) is the joint probability density function of annual supply and demand, the probability of an annual shortage is given by:
If losses are known as a function of the shortage, say L(X), one can compute expected annual losses:
Since the probability distribution of supply depends on the raw water system, and storage and treatment capacities (call that capacity K), and since the distribution of demand depends on adopted policies, P (pricing, conservation, etc.), we can write f (D,S: K,P) so that (3) becomes a function of K and P. One can then optimize the system with respect to K and P, i.e. one could choose values so as to minimize the present value of combined capacity costs, expected shortage-induced losses, and policy-related costs (objective and subjective). Letting r be the appropriate annual discount rate and assuming f to be unchanged over time, one seeks to
The major inputs needed to execute any such policy include demand studies to identify causal factors and responses to policy, hydrologic-climate models and estimation of the loss function, L, for the whole range of shortages, D-S. It should be noted that system costs included in (4) subsume appropriate short-term steps to minimize costs during shortage events. These costs depend not only on the hydrology of the system and the demand pattern but also on the utility's decision rule concerning the allocation of shortages among residential, commercial, industrial and public water users. For typical urban utilities within which residential demands comprise, say, 50% to 80% of total demands, and in light of empirical findings that typical residential drought losses are small (Russell, Arey and Kates, 1970), it seems reasonable that outdoor residential uses would be required to absorb initial shortages.
The optimization process described above can be thought of as choosing the optimal "reliability" for the water system, R*, by observing the marginal present value of expected benefits from greater reliability, MB(R), (that include lower objective and subjective losses due to shortages) and the marginal present value of expected costs of reliability, MC(R) that include investment and policy-related costs. R* should be chosen to equate MB(R) to MC(R). This can be graphed as in Figure 2.
Figure 2. Selecting the Optimal Reliability
Drought Response Measures As A Component of Statewide Systems Reliability Optimization
The discussion above concerning drought policy as a component of overall water system optimization (including demand influencing measures) was cast in terms of a typical urban water utility: the utility managers set policy, perhaps following policy guidelines set forth by the City Council, including the design of steps to be followed during drought, e.g., the allocation of shortages among different groups of users, rationing and use restrictions, conservation, education, etc. The steps usually are intuitively designed, although some cities (e.g., Boulder, CO and Seattle, WA) have undertaken research to evaluate the efficacy of steps like price increases, conservation subsidies, metering, etc.
The selection of optimal drought response becomes much more complex at the state level. There users are more diverse and include all types of agricultural and other primary activities. Supplies come from many sources and have different types of claims on them (e.g., contracts, appropriative rights, ditch company shares). Relevant climate conditions are not known in sufficient detail to allow the formulation of region-wide plans because of the existence of micro-climates that can cause vastly different water conditions to exist within small sub-state regions (in southwestern Colorado during the 1976-77 drought, some Districts had nearly normal supplies while others experienced record dry conditions (Howe, et al, 1982). Shortage loss functions are seldom known, nor are the values of water in different uses. All of these informational problems point to the difficulties that a centralized water bureaucracy would have in allocating water rationally during drought.
Institutions for water allocation should exhibit five characteristics (Howe, et al, 1986): (1) flexibility in allocation over time to respond to changing supply and demand conditions; (2) security of tenure of title so that longer term investment plans will be secure; (3) fairness to all affected parties (no one worse off); (4) they should present the water users with the real opportunity cost of the water allocated to them; and (5) the outcome of the allocation process should be predictable to the participating parties. In the case of drought, flexibility in the allocation of water among uses is critical so that longer term investments can be protected and so that fairness can be sustained, but in a framework that forces drought period water users to take account of the opportunity costs of their water supplies.
The information requirements for carrying out the optimization depicted by equation 4 and Figure 2 are so formidable that the use of decentralized or market-type allocation mechanisms during droughts must be seriously investigated. Thus the remaining parts of this paper will consider the actual and potential role of markets and market-like mechanisms in dealing with drought, with emphasis on California's drought experiences and some of the institutions that have been developed in the United States.
INNOVATIONS IN DROUGHT RESPONSE
The Colorado Drought Experience of 1976-77
The drought of 1976-77 inflicted hardships on small rural towns and agricultural water users in Colorado (Howe, Alexander and Moses, 1982). Small towns experienced the greatest difficulties because of inadequate supply systems, exhaustion of groundwater, or inadequacies of water rights. In a number of cases, emergency supplies had to be provided by tank truck.
Among the lessons learned were that (1) there was great variation in the conditions facing adjacent drainage basins, some having nearly normal supplies and others facing record low flows; and (2) flexibility in allocation among agricultural users and water suppliers (e.g., mutual ditch companies) occurred frequently and served greatly to reduce drought damages. These findings point once again to the desirability of decentralization of decision-making regarding reallocation of water to take advantage of information on local conditions. Uniform state-wide policies would over-assist some areas and under-assist others.
Fortunately, water use patterns are not as rigid as a strict interpretation of appropriation doctrine would imply. Rights holders can share or sell some of their water on a short-term or "rental" basis. Additional drought mitigation practices observed in Colorado in 1976-1978 included those listed below. Significantly, most of these were locally initiated.
Proposed Agricultural "Lease-Outs"
The concept of the agricultural "lease-outs" is similar to that of short-term rentals but within a formal long-term contractual arrangement. A water user (a town, industry, or hydro-power company) contracts with agriculture (an irrigation district or large irrigated farm) to take some agricultural water under drought conditions. Notice of intent to take the water must be given in early Spring so that farmers can plan not to plant certain acreage. Land is then dried up for the season and the water left in the stream for use by the purchaser. Proposed payment arrangements differ but usually include an annual "retainer" payment plus a larger payment in years when water is actually taken.
Two careful studies have been made of such lease-outs. Whittlesey, Hamilton, and Halverson (1986) have studied the potential for a sizeable lease-out of agricultural water on the Snake River by Idaho Power Corporation. Idaho Power would take the water during low flow years to firm up its base power load -- an action of great value to the Company since reliable (firm) power is worth much more than interruptable power. Michelsen (1988) has modelled and evaluated a similar lease-out of Poudre River agricultural water by the town of Fort Collins, CO.
The conclusions of both studies are that (1) the fundamental hydrologic, economic and institutional conditions needed for beneficial lease-outs already exist in most of the semi-arid areas of the western U.S. and (2) net economic benefits from these arrangements can be quite large over a wide range of conditions, up to as much as $2,800 per acre-foot leased in the Fort Collins cases.
Conservation Lessons from the 1976-77 California Drought
Meral (1979) has analyzed various programs of water conservation that were introduced in California during the 1976-78 period. In February and March of 1977, the California Department of Water Resources (DWR) surveyed water users in the Marin Municipal Water District (MMWD) because of the severity of drought impacts on Marin County. The results showed that in 1976, normal water withdrawals were reduced by 25% and, in 1977 the reduction reached 64%, more than the 57% goal that was mandated. Part of the reduction was attributable to the use of reclaimed water applied to nonresidential landscapes: 12 million gallons in 1976-77 and 50 million gallons in 1977-78. Regarding losses from reduced water use, the Marin County survey showed that 2/3 of single family dwellings indicated some loss of landscaping averaging $570 per dwelling.
Emergency supplies were provided by a temporary pipeline from the East Bay Municipal Utility District (MUD), water provided by a complex institutional and physical scheme. The water was part of the Metropolitan Water District (MWD) State Water Project entitlement, delivered to a San Francisco reservoir, transferred to the City of Hayward, and hence to East Bay MUD. Institutional flexibility again came to the rescue!
San Francisco reduced per capita use by 30%, while East Bay MUD's use fell by 36%. Los Angeles reduced per capita use by only 12%. In each case, reduced per capita use persisted into 1978 and probably much longer (no current date). 35 surveyed urban areas showed a 21% average reduction in water use between 1976 and 1977. Some minor sewage treatment problems occurred because of reduced flows but were solved by changes in plant operation. Table 1 exhibits the participation rates in 10 common conservation measures.
Agricultural Conservation Lessons
Recognizing that 85% of all consumptive use is in agriculture, the State of California has taken some steps to stimulate conservation in that sector. Following the 1976-77 drought, a large program of both research and information was undertaken by DWR to promote sprinkler and drip irrigation. The Katz-Bates bill of 1984 expressly guarantees that water thus salvaged belongs to the farmer for other uses or for sale.
Fullerton (1979) made the following observations regarding agricultural drought:
Table 1
Drought-induced Conservation
MarchMarch
19791977
Washing the car less often56%68%
Cutting back on water outdoor gardens,
lawns, shrubs 46% 67%
Taking shorter showers, using less water
in the bathtub43% 70%
Running the washing machine less often35% 54%
Using less water for brushing teeth,
shaving & other personal washing34% 10%*
Using devices in toilet tanks to reduce
the amount of water in each flush27%25%
Using dishwasher less often25% 39%
Using reducers on showers to slow down
water flow 21% 16%
Flushing the toilet less often15% 53%
Saving waste water from kitchen or
bathroom to use elsewhere13% 25%
None of These 12% 7%
Base: (1006)(962)
* Not included on list in 1977, but volunteered by 10% of respondents
Source: Meral, 1979, Chart 2, p. 41.
Water Banking: California and Idaho
Within the past year in Southern California, desalination plants and towing icebergs from the Arctic have been considered as alternative water supplies. The City of Santa Barbara has committed itself to building a desalination plant which will produce water directly from the ocean, and Goleta, California will import water from Canada by tanker. Santa Barbara also is considering connection with the State Water Project. This would involve tunneling through the Coast Range, which would provide water at approximately $1,000/AF (Priest, July 30, 1991). Vaux and Howitt (1984) found marginal values of raw water of $210/AF in Southern California municipal uses while Central Valley farmers continued to grow crops such as alfalfa and rice which produce marginal water values of less than $15/AF (DWR, 1978).
The presence of such differences in water values suggests the gains to be made in economic efficiency by introducing market-like mechanisms. In light of this, "water banking" has been reintroduced. This section will look at some of the experiments with it in recent times, what went wrong and how those problems could be corrected.
In some regions there is no market for water. Farmers, especially in California, receive allotments for irrigation at a certain price which has been determined through a non-market process, usually contractual agreements. If the M & I users wish to purchase agricultural water by offering a price between their marginal valuation of additional water and that of the farmers, there are often constraints preventing them from doing so. Several constraints can prevent an effective water market from developing are: poorly defined ownership rights, large third-party effects, interstate compacts and international treaties, and the desire to protect agricultural interests.
Water banks require some physical factors to be in place: sufficient storage space such as reservoirs or natural lakes and a reasonably extensive network of canals or streams in order to facilitate transfers among a large number of buyers and sellers over a sizeable area. Most areas in the West, including the Colorado River Basin, have satisfied these requirements in the past fifty years through Bureau of Reclamation and other federal agencies' water supply projects for agriculture. Needed institutional accommodations include relaxation of usual water law and regulations, especially under drought conditions.
U.S. Bureau of Reclamation Water Bank (1976-77)
The first water bank was established by the Federal Emergency Drought Act of 1977 in response to the drought of 1976-77 in Southern California (Stavins, 1983). This water bank was administered by the Bureau of Reclamation in order to facilitate transactions between willing sellers and buyers. The Bureau was authorized to purchase water only from Federal water contractors, Indian irrigation projects and State-authorized agencies. Thus, it could hardly have been considered a free market. Sellers were paid according to the opportunity costs of their water and prices paid ranged from $15/AF to $85/AF. 46,438 acre-feet were purchased in this way. The Bureau sold the water to various qualified entities according to their willingness to pay, or the marginal benefit expected from the use of that water. The prices received for this water ranged between $55/AF and $142/AF (Wahl, 1989). See Table 2.
Table 2
Federal Water Bank in California, 1976-77 Drought
Sellers Amount (AF)Price ($/AF)
Dept of Water Resources 818585.51
Pleasant Grove-Verona Mutual
Water Company 1575270.00
Chaplin-Lewis-Lewis 127935.00
Reclamation District No. 108 500025.00
Pelger Mutual Water
Company 442515.00
Natomas Central Mutual
Water Company 600015.00
Sacramento River Water
Contractor's Assn. 579715.00
TOTAL 46438
Purchasers Amount (AF)Price ($/AF)
Hills Valley ID 76142.44
Stone Corral ID 124138.97
Tri-Valley WD 135138.43
Delano-Earlimart ID 200136.47
Lindsay-Strathmore ID 300135.15
Terra Bella ID 503134.10
Mustang WD 112 64.00
San Luis WD 1144 63.59
Romero WD 120 62.35
Broadview WD 120 61.57
Hospital ID 1389 61.39
Kern Canyon WD 605 60.98
Arvin-Edison WSD 500 60.71
Davis WD 304 59.90
Salado WD 500 59.08
Centinella WD 329 58.98
Orestimba WD 843 58.72
Quinto WD 435 58.18
Del Puerto WD 1326 58.06
Westlands WD 22362 57.92
Foothill WD 1100 57.76
San Luis WD 5469 57.58
Plain View WD 1180 57.52
Sunflower WD 1205 57.03
Contra Costa WD 1250 56.20
Panoche WD 891 55.96
Glenn-Colusa ID 22 54.93
TOTAL 42544
Source: Records of Contracts and Repayments Branch, Bureau of Reclamation, U.S. Department of the Interior, Washington, DC.
Snake River Water Bank (1980-Present)
The second case concerns a water bank which has been operating in southern Idaho along the Upper Snake River (Miller, 1989). In 1980, the State Legislature of Idaho formally authorized rental agreements which had existed since the 1930's. The water banking legislation assured the owners of stored water that seasonal transfers through the water bank would not jeopardize their rights to the water (Miller, 1989). In this water bank, federally supplied water is obtainable at $0.19/AF by holders of contracts for water from the Bureau of Reclamation reservoirs on the Upper Snake River within Idaho Water District No. 1. Many holders keep large water holdings as insurance against drought (Wahl, 1989). The Bureau of Reclamation offers $2.50/AF for undesired water, to be leased for one year and in turn sells this water to others in the basin who desire it, principally other irrigators and the Idaho Power Company, charging only this price which includes a transaction cost of $0.50/AF involved in completing the transfer. See Table 3.
Table 3
Water Transfers on the Upper
Snake River Idaho, 1979-86
Water WaterIrrigationPowerPrice
Year Suppl (AF)Purch (AF) Purch (AF) Purch(AF) ($/AF)
1979 88,870 73,96023,960 50,0001.19
1980 72,190 14,57514,575 01.19
1981 170,107149,03924,039125,0001.60
1982 290,426203,515 3,515200,0002.30
1983 540,606353,084 3,084350,0002.40
1984 806,400277,433 2,433275,0002.50
1985 497,302362,16912,169350,0002.50
1986 895,642159,735 9,735150,0002.50
Source: Records of Idaho Water District No. 1 (Idaho Falls) and the Bureau of Reclamation, U.S. Department of the Interior, Boise, Idaho.
An obvious problem with this water bank is that the buying and selling prices are arbitrarily set, so they don't reflect market conditions. In 1986, for instance, the Water Bank bought in 895,642 AF and sold only 159,735 AF. Clearly, the price was too high for the given conditions. In wet years, there will be an oversupply and in dry years excess demand will occur, because those who value water most highly will not be able to bid it away from those who value it only slightly above $2.50/AF.
California Drought Water Bank (1991)
Most recently, the California State Legislature has turned to the water bank as a way to mitigate drought conditions (Macaulay, 1991). The State Water Resources Control Board (SWRCB) has been authorized to purchase water from any entity with proven water rights. Individual water rights holders can sell their rights freely at a schedule of prices reflecting foregone incomes in agriculture, but regional and local water districts must satisfy some stringent conditions required to prove that surplus conditions exist in their district. The State Water Code, as amended in 1971, prohibits districts from transferring water outside district boundaries unless it is declared "surplus." The test of "surplus" requires the water to be unwanted at any price by any member of the district (Saliba, 1987). This substantially reduces the supply available for reallocation.
Through this water bank the SWRCB has purchased a significant amount of water at $125 per acre-foot and plans to sell it at $175/AF. The problem is that both the purchase and sale prices were administratively set, with no idea how much water would be desired at $175/AF. The result was the purchase of 836 thousand acre-feet. See Table 4. DWR has agreed to buy all unsold water for use by the State Water Project.
Table 4
California Drought Water Bank, 1991
Sellers Amount (AF)Price ($/AF)
Delta 333,814125
Sacramento River 83,419125
Yolo Bypass 56,856125
Yuba/Feather/Elsewhere354,781125
Above Shasta Reservoir 6,709125
TOTAL 835,579
Purchasers Amount (AF)Price ($/AF)
Alameda County WD 14,800175
American Canyon
County WD 370175
Dudley Ridge WD 17,205175
Kern County WA 73,997175
Oak Flat WD 925175
San Francisco 50,000175
Santa Clara Valley WD 39,500175
Metropolitan WD215,000175
Westlands WD 13,820175
San Luis WD 1,567175
Crestline-Lake Arrowhead 236175
Contra Costa WD 6,717175
Alameda County-Zone 7 500175
TOTAL 434,647
The initial agreement establishing a 1991 California Drought Emergency Water Bank stipulates that the water bank should exist only for the 1991 dry season (Macaulay, 1991). A number of water districts and other state agencies are the only ones eligible to make purchases from the water bank, and are required to deposit 75% of the purchase price for the water in advance to provide the funds for the water bank to purchase water. The sellers, however, are not restricted to particular groups as was the case in the 1977 water bank, but the targeted group is farmers in the Delta region. Any farmer who agrees to fallow land will be reimbursed according to a preset schedule of crop irrigation water requirements. Other sellers are paid $125/AF.
Numerous constraints are placed on which districts can purchase from the water bank. It appears that the goal of the water bank organizers was to prevent any appearance that certain communities, especially wealthier ones, would be able to hoard water resources while poorer communities fail to have the minimum safe level to maintain health and safety requirements. The enabling legislation allowed the governor to make emergency appropriations for these types of requests. In addition, no district is permitted to purchase water if it is shown to be using resources inefficiently, not using all conservation practices in common usage, or can not show a shortage of supplies given current demand levels (Water Bank Contract form, 1991).
These restrictions on the functioning of the California Water Bank appear to reflect continued skepticism among government and other leaders concerning the ability of a water transfer market to allocate the resource effectively during drought. But in so restricting Bank activities, potentially large benefits from a more flexible market system are lost. If farmers and other low marginal value users knew they could count on a market for their excess water on a regular basis, they would be encouraged to engage in conservation practices or to produce those crops which use water more effectively. Lining of irrigation canals, patching of leaks in these systems, and allowing cropland to lie fallow would become potentially profitable for farmers, rather than raising the risk of potential loss of water rights or contract water. In the case of the 1991 bank, the farmers are able to sell their water only for 1991. The motivation for permanent improvements in water conservation thus is absent.
Colorado River Basin Water Bank (Proposed)
The final water bank to be considered is one which has been proposed as a partial solution to the problems of both short term and long term water allocation in the Colorado River Basin. The proposal is for an interstate water bank involving the riparian states of Wyoming, Colorado, Utah, Arizona, New Mexico, Nevada and California to be established to help each of the seven states meet their individual and collective water supply needs during critical water supply periods (Colorado River Board of California, 1991). All transactions would take place through the Water Bank with state agencies acting as agents for buyers and sellers in their respective states. Each state would be limited to purchases of 1,000,000 AF and price limits might be established (a poor idea, as seen in past water banks).
Under the Colorado River Water Bank proposal, water currently being consumptively used by the Upper Basin states could be leased to Lower Basin states on an annual basis. To reduce the attractiveness of long-term transfers, the Lower Basin states propose to engage in conservation practices which would lower their consumption of Colorado River water to the levels they are entitled to under the Colorado River Compact and the U.S. Supreme Court case of Arizona v. California. It also would eliminate the urge to build new projects in the Upper Basin for the sole purpose of protecting the rights to water entitled to them.
Water Valuation Estimates in the Colorado Basin
The main rationale for a Colorado River Water bank is the large discrepancy between marginal values of water in the Upper and Lower Basins. Vaux and Howitt (1984) estimated marginal values in Southern California of $210/AF for municipal uses and $45/AF in Imperial Valley agriculture, clearly implying that there are significant gains to be had from within-state transfers. They also estimate increasing marginal valuations as growth places a premium on water use, going up to $360/AF and $60/Af respectively in 1995. Upper Colorado River Basin water values at the margin were estimated by Booker (1989) at $18/AF, suggesting large gains in economic efficiency from interbasin transfers.
Booker (1989) estimated maximum annual benefits achievable through interstate water transfers of Colorado River water at $140 million annually, with $74 million for consumptive use, $35 million for hydropower generation and $31 million in reduced salinity damages. On the other hand, large benefits could be experienced simply through intrastate transfers in the Lower Basin. Transfers between MWD and the Imperial Valley Irrigation District could provide gains of $69 million, with $98 million in consumptive use gains, less added salinity damages of $24 million (Booker, 1989).
Progress toward efficient water allocation in the Colorado River Basin should first move toward water banks in each state, so that each state individually could move toward more efficient water usage. Then a Colorado River Basin water bank could be organized to facilitate water transfers among states. From a legal standpoint, this would be the form least likely to violate the Colorado River Compact and other agreements and court decisions- the "law of the river". Within-state transfers can establish a market price in each state, after which a state would be allowed to put its surplus water on the basinwide market. For example, if Colorado's water bank established a market price of $30/AF in the state, and California's water bank established a price of $50/AF, sellers in Colorado with marginal valuations below $30/AF could sell to the Colorado bank, which could sell its excess water to the California bank.
Conclusions Regarding Water Banks
Each of the water banks studied had both positive aspects and some shortcomings. In the case of the 1977 Water Bank, the participants were limited to Federal and State users, thus excluding a large group of inefficient water users, the farmers. On the other hand, prices were allowed to reflect opportunity costs and buyers' willingness to pay, thus ensuring efficient transactions between those who participated in the water bank. In the Snake River situation, all contractors of reservoir water in the region can participate but water prices are not allowed to vary to reflect market conditions. In the California Water Bank of 1991, the buyers had to be water districts demonstrating need, while sellers were targeted as farmers in certain regions of the state. The fixed price in this case is inefficient.
As long as water banks are thus constrained in their operation, they will not function as well as they might. Appropriate legislation is needed to guarantee title to water that is leased through water banks and to allow supply and demand conditions to determine appropriate prices.
RECOMMENDATIONS
The first need is that public decision-making processes be opened up to the input of objective data, research and analysis. So often, the range of alternative solutions to a problem gets narrowed prematurely, without the utilization of available information. At times, this narrowing of the decision process occurs because of traditional engineering approaches that omit considerations of institutional and individual behavior. At times, it occurs because special interests dominate the decision process and desire to secure benefits that attach to specific policies or programs (e.g., the cost-free features of federally provided flood control or the huge subsidies attached to federally provided irrigation water). The narrowing may also result from long-term political commitments to projects that may represent the capstone of a politician's career (e.g., Senator Carl Hayden's 40-year campaign for the Central Arizona Project. See also Imhoff, 1991, for insights into the political pressures that are exerted on planning teams.).
A second recommendation is that much more serious consideration be given to water banks as permanent components of the institutional framework for water allocation. It will be necessary to allow for greater price responsiveness to actual supply and demand conditions if the efficiency of water reallocation through water banks is to be assured. The actual mechanisms for setting prices in water banks (e.g., auctions, pre-set schedules of estimated opportunity costs, one-on-one matching of buyers and sellers) need to be much more carefully analyzed.
Thirdly, the possibility of interstate water banks needs to be given further consideration. While the August, 1991, proposal by California has been rejected by the Upper Basin states, the potential benefits to all participants (see "Water Valuation...in Section III.E.) warrant further co-development of linked state and interstate water banks.
Finally, further study of the objective and subjective costs of drought-induced water shortages in both urban and agricultural settings is warranted. It is time that the level of water system reliability be based on objective risk analysis rather than outdated rules of thumb.
REFERENCES
Booker, James F., 1990, "Economic Allocation of Colorado River Water: Integrating Quantity, Quality and Instream Use Values," Unpublished Ph.D. dissertation, Colorado State University, Fort Collins, CO, p.14.
California Department of Water Resources, 1978, "The 1976-77 California Drought, A Review," Sacramento, CA.
City of Boulder, 1988, "Raw Water Master Plan," September.
Colorado River Board of California, 1991, Conceptual Approach for Reaching Basin States Agreement on Interim Operation of Colorado River System Reservoirs, California's Use of Colorado River Water Above Its Basic Apportionment, and Implementation of an Interstate Water Bank.
Cummings, Ronald G., David S. Brookshire and William D. Schulze (eds.), 1986, Valuing Environmental Goods: An Assessment of the Contingent Valuation Method, Totowa, NJ: Rowan and Allanheld.
Fullerton, Herbert H., 1979, "Drought Lessons from Agriculture," in Proceedings of the Conference on Water Conservation: Needs and Implementing Strategies, American Society of Civil Engineers, New York.
Goodman, Alvin S., 1984, Principles of Water Resources Planning, Englewood Cliffs, NJ: Prentice-Hall, Inc.
Howe, Charles W., Paul K. Alexander and Raphael J. Moses, 1982, "The Performance of Appropriative Water Rights Systems in the Western United States During Drought," Natural Resources Journal, Vol. 22, April, p. 379-389.
Howe, Charles, et al, 1986, "The Potential for Water Markets," Water Resources Research, Vol. 22, No. 4, April, p. 440.
Imhoff, Edgar A., 1991, "Management of the San Joaquin Valley Drainage Program: The Dichotomy Between Practice and Theory," Chapter 44 in (eds.) Ariel Dinar and David Zilberman, The Economics and Management of Water and Drainage in Agriculture, Boston: Kluwer Academic Publishers.
Knetsch, J. S. and J. A. Sinden, 1984, "Willingness to Pay and Compensation Demanded: Experimental Evidence of an Unexpected Disparity in Measures of Value," Quarterly Journal of Economics.
Knetsch, Jack L., 1990, "Environmental Policy Implications of Disparities between Willingness-to-Pay and Compensation-Demanded Measures of Values," Journal of Environmental Economics and Management, 18, 227-237.
Macaulay, Steve, Drought Water Bank Manager, 1991, "Water Marketing in California as a Strategy to Meet Future Urban and Irrigation Demands," California Department of Water Resources, September, draft, p.2.
Maidment, D. R., S. Miaou and M. M. Crawford, 1985, "Transfer Function Models of Daily Urban Water Use," Water Resources Research, 21 (4).
Meral, Gerald H., 1979, "Local Drought-Induced Conservation: California Experiences," in Proceedings of the Conference on Water Conservation: Needs and Implementing Strategies, American Society of Civil Engineers, New York.
Michelsen, Ari M., 1988, "Economics of Optioning Agricultural Water Rights for Urban Water Supplies During Drought," unpublished Ph.D. dissertation, Department of Agricultural and Resource Economics, Colorado State University, Fort Collins, CO.
Miller, Kathleen A., 1989, "Hydropower, Water Institutions and Climate Change," Water Resources Development, Vol. 5, No. 2, June, pp.71-83.
Mitchell, Robert Cameron and Richard T. Carson, 1989, Using Surveys to Value Public Goods: The Contingent Valuation Method, Washington, DC: Resources for the Future, Inc.
Moreau, David and Keith W. Little, 1989, "Managing Public Water Supplies During Droughts: Experiences in the United States in 1986 and 1988," Water Resources Research Institute of the University of North Carolina, Report No. 250, September.
Nelson, John Olaf, 1979, "Northern California Rationing Lessons," in Proceedings of the Conference on Water Conservation: Needs and Implementing Strategies, American Society of Civil Engineers, New York.
Russell, Clifford S., David G. Arey and Robert W. Kates, 1970, Drought and Water Supply: Implications of the Massachusetts Experience for Municipal Planning, Baltimore: Johns Hopkins Press for Resources for the Future, Inc.
Saliba, Bonnie and David B. Bush, 1987, Water Markets in Theory and Practice, Westview Press, Boulder, CO, p. 111.
Schulze, William D., G. H. McClelland and G. W. Russell, 1988, "Pretesting Plan for the Niagara-Mohawk Residential Survey on the Value of Service Reliability" (xerox), Center for Economic Analysis, University of Colorado, Boulder.
Smith, V. Kerry and William H. Desvousges, 1987, "An Empirical Analysis of the Economic Value of Risk Changes," Journal of Political Economy, 95-1, 89-111.
Stavins, Robert, 1983, "Trading Conservation Investments for Water," Environmental Defense Fund, pp. 29-36.
Vaux, H.J. and R. E. Howitt, 1984, "Managing Water Scarcity: An Evaluation of Interregional Transfers," Water Resources Research, Vol. 20, pp. 785-792.
Wahl, Richard., 1989, Markets for Federal Water, Resources for the Future, Inc., Washington, DC, pp. 136-139.
Whittlesey, Norman, Joel Hamilton and Philip Halverson, 1986, "An Economic Study of the Potential for Water Markets in Idaho," a study for the Snake River Studies Advisory Committee, Office of the Governor, State of Idaho, Idaho Water Resources Research Institute, Moscow, December.
Willig, R., 1976, "Consumer's Surplus Without Apology," American Economic Review, Vol. 66, No. 4, September.