A modeling approach that relies on analyzing the individual farm operation as the unit of interest, as proposed in the types of models described in Rosen and Sexton, and Zusman and Rausser, has two problems. The first is that it misses the influence of non-farm voters on district decisions, particularly in popular-vote and board-appointed selection systems. The second is that the data requirements for a sufficiently broad empirical analysis quickly overwhelm the available resources for most studies of this type. A farmer proceeds through several decision-making stages in deciding what to plant, production levels, investment and water use. The initial choice is the size of the operation. The decision as to how much land to put under cultivation and irrigation is dependent on many factors such as how it is acquired , available financial resources, which crops are appropriate, past resource usage, variation in land quality, and distance to markets. Once this choice is made, a farmer chooses to plant and irrigate on their most “flxed” asset, land,wholesale plant containers to the maximum extent possible and selects that appropriate crops, water use and irrigation technologies on that basis.
Next the farmer selects the crops to be grown on this land. This choice drives other factor choices, particularly for water. Most crops require a fairly narrow range or “effective” water application as determined by local evapotranspiration requirements and land quality factors such as permeability, drainage and nutrient levels . The amount of effective water, e, is a product of the amount applied, a, and the technical efficiency of the irrigation method, h. The farmer then adjusts either irrigation technology/source or amount of applied water to compensate for changes in the other factor. As a result, the farmer faces a two-stage problem-first choose either water applied or irrigation efficiency, then select the other given conditions that dictate effective water requirements . Thus, the farmer first chooses optimal input levels for a particular mode and efficiency of irrigation, hi’ and selects the irrigation method that provides the largest net profits to the farm.The decision on how much water to apply can be a long-term commitment. Historically, only a few opportunities have arisen to acquire surface water supplies with the initiation or expansion of water projects . These water “markets” only opened for short periods and only offered long-term contracts. Water diversion is capital intensive and can require commitments up to 40 years with payments relatively invariant with actual usage. While water market opportunities now are expanding and environmental regulations are constraining supplies, even in these cases farmers face long-term choices.
Because of this time frame, the amount of water to apply from water district sources appears to be the dominant variable in choosing how to meet effective water requirements, and efficiency is a residual of these choices; thus we can leave a choice variable, h, to the second stage. The amount of effective water as a result is based on an expectation about the amount of land under cultivation, the price of water and of irrigation technologies, and the price and availability other inputs. The water-use efficiency variable, h, can be interpreted in several ways, either as improved irrigation technologies or as greater reliance on water sources autonomous from district supplies, such as groundwater pumping. This decision of selecting the appropriate irrigation technology and/or water source has a long lead time as well and requires year-to-year planning to change. The expense of selecting a different technology is captured in the investment cost of the technology, I, and the cost of pressurizing the irrigation system or for local groundwater pumping, v. However since h = AlE, these costs are actually dependent only on the amount of water applied, a? Thus v and 1become functions of a as well. Perhaps the most important reason for forming any water district is the provision of a reliable water supply. The issues of overall supply and service quality must be addressed collectively because they have clear “common property” traits.
Adding capacity to a reservoir is likely to improve everyone’s supply reliability within the district if the water rights are effectively “correlative” . Defining the property rights to this added capacity would undermine the cooperative nature of the district. The district is then searching for the “optimal” choice for these variables based on a set of rules. These rules begin with deriving the opportunity cost or “shadow value” of the water supply. The choice of the supply capacity, S, directly influences reliability–the greater the storage capacity and transfer capability, the longer the district is able to carry over storage during drought periods. In other words, the probability that full water deliveries will be available, F, increases with the size of storage capacity, S. The average supply availability below full deliveries is the sum of the probabilities of these lesser flows . However to simplify this problem, we can present it as a dichotomous probability case of either full deliveries or drought constrained deliveries without any supply capacity, Sd, which equal approximately the average of the less-than-full delivery conditions. Often the terms “efficient,” “social-welfare maximizing” and “wealth maximizing” are often used interchangeably by economists as though they represent much the same measure. However, attaining the maximum wealth for a group may not be the most efficient outcome because two individuals still might want to trade among themselves. This results from their respective preferences changing at nonlinear rates. Perhaps even more confounding is that the distribution of wealth may also be important in attaining the preferred level of social welfare. Because the classical model often uses monetary measures of well-being, through profits, it reduces the definition of efficiency to maximizing wealth. The problem with defining efficiency solely in terms of net monetary benefits is that the “cooperative” has key difference from the “firm” in the neoclassical sense—cooperative members maximize over their individual preferences which may include non-monetary outcomes, while a firm’s shareholders only derive monetary returns. For comparative purposes though, we define our efficiency measure in this reduced simplistic form, which in turn may be somewhat misleading in a political-economic analysis. If an agricultural water district was managed as a wealth-maximizing cooperative,plastic pot manufacturers it would choose the mixture of investment in water-supply capacity and agricultural production that would generate the greatest net benefits for its members. Water would be priced at its marginal cost internally to signal the most efficient uses to members, and any net profits or losses from water- supply operations would be returned to district members in a fashion which would not distort water-use decisions. In fact, this model is institutionally quite different from the way public enterprise district operate. Existing districts have several characteristics distinct from this model. The most important is the so-called “non-profit” requirement, i.e., that expenditures and revenues must be in approximate balance. Revenues are often limited to sources directly linked to water-use, e.g., prices, charges or property taxes, and thus pricing must approximate average, not marginal, costs. The net benefits from the district also may be allocated in any number of ways, some of which distort water-use choices by farmers.
Finally, water district board members tend to choose policies which allow them to continue to hold office. This means pleasing enough constituents to gain a majority of votes. Policies that increase total district wealth may benefit only a few district members and not generate sufficient political support. Even though the “efficient cooperative” model may not be appropriate institutionally, it is useful as a benchmark to measure performance by other institutional forms. One can assess how a district’s manager might choose to maximize total wealth if the manager could control all internal resource management decisions either through directives or complete internal pricing mechanisms. Thus, this is more appropriately called the “wealth-maximizing” model. This model assumes that farmers see the full and direct costs for the water resources that they use and receive back the net profits from the operations of the district. The institutions that manage and price such water resources are “transparent” in this case. The district does not face a non-profit constraint, nor must it decide how to return any excess profits to district members. Distribution of total benefits is not addressed in this model. However the model provides a useful measure for comparing the different institutional arrangements that water districts use in California. In the efficient cooperative model, we assume that an “omniscient central planner” allocates all resources to produce the highest level of total district net wealth. Of course, in reality these functions are institutionally segregated between an elected or appointed governing board and the individual farmers. In the latter case, the issue becomes coordinating the actions between the farmers and board members through “signaling” such as pricing and voting. This is .confounded by the effects on these signals of distribution of that wealth among district members-the “political economy” of the district. District board members try to stay in office by pleasing a sufficient number of constituents through their policy choices. They attempt to win a majority of votes by addressing the issues that most affect district members. This is the basis of the median-voter model . This idea can be extended to incorporate the “interest group” concept by assessing how voters grouped by key characteristics might respond to different policy choices, and determining whether board members can assemble a majority vote by appealing to these various groups . The existence of different voting rules in California’s agricultural water districts allows us to test this hypothesis. Several different methods are specified in state law to identify qualified electors and how to weigh votes for electing governing boards. The two dominant methods are the property qualification, assessed value-weighted method and the universal-franchise, popular vote method.9 The former allows only those who own property to vote, and each owner is given a vote in proportion to the assessed value of their land. This method might be interpreted as allocating votes in proportion to the value of net output from an agricultural district. The latter method enfranchises any registered voter and simply tallies one person/one-vote. This is also the most common method for electing officials in other governmental jurisdictions. While a board cannot guarantee that a particular voter will vote for them, they can affect the likelihood that they will receive a positive vote. The board has five variables to consider: who the eligible voters are in the district, the well-being of the district’s individual voters, the cost of the district’s water supply, the variability and reliability of the district’s supply, and the mode of collecting the district’s required revenues. We focus on the district board’s objective function which is to maximize the number of voters subject to meeting a non-profit budget constraint. The function y specifies the relationship of individual net benefits for district voters and the likelihood of those voters voting for the incumbent board. y can be interpreted as a single utility function in which the output is a “yea” or “nay” vote on the current district management. For purposes here, we need not specify the exact function, but only note that y increases as net benefits increase for members within each interest group.The objective function for a tenant farmer differs from an owner/operator in two ways from that of an owner. First, tenant farmers are more likely to incorporate a risk premium, p, on fixed irrigation technology investment due to the nature of tenancy versus ownership . Tenants risk not being able to fully recover investment costs since they do not control land use and cannot regain fixed investment in the land value. In other words, their risk of sunk costs in investment stand to be substantially higher. This effectively increases the apparent cost of upgrading irrigation efficiency if we assume improvements require higher fixed investment . Second, a property tax has only a secondary effect through the rent on land costs to tenants. A portion of the property tax incidence is on landlords. Thus tenants do not fully realize the brunt or benefit from changes in this type of tax. Models for two different types of water districts are evaluated in the next two sections. Each model is constructed in parallel to the constrained efficient cooperative to allow direct comparison.