AIC Issues Brief
No. 7 September 1998
WATER TRANSFERS AND GROUNDWATER MANAGEMENT: AN ECONOMIC ANALYSIS
Agricultural, industrial and municipal water use in California clearly depend on the state’s groundwater resources. What is less known is how water transfers and water markets-voluntary transfers of water between willing buyers and sellers-will influence the relationship between overall water use and groundwater.
Both drought and cuts in surface water use to protect riverine habitat have reduced the reliability of agricultural water supplies. In addition, cities are increasingly vocal in their demands for increased and more secure water supplies. Federal and state law changes designed to facilitate water transfers indicate that some water currently being used by farmers may well be sold to cities or to other farmers. Thus, there is need to better understand the two-way relationship between the groundwater resource and surface water shifts to other uses.
This AIC Issues Brief summarizes a case study of the potential effects on groundwater and on agricultural production of reduced surface water use resulting from water market transfers with and without groundwater management. Although inter-basin water transfers associated with water markets remain rare in California, two such projects-the Kern Water Bank and an Arvin-Edison/MWD exchange-already exist or are being planned for Kern County. Its proximity to the Los Angeles region, its location south of the Delta, and its access to both surface and groundwater supplies position Kern County for participation in such projects.
Before discussing the results, certain inputs to the model need to be specified. Annual average surface water supplies in Kern County are about 2 million acre-feet (MAF) from three major sources: the California State Water Project, the federal Central Valley Project, and the Kern River. Our model assumes that 70% of these supplies are available for irrigation and the remainder goes to the aquifer as conveyance losses. Deep percolation flows from farmland irrigation are assumed to be 20% of the amount applied and natural recharge to the aquifer is 52,000 acre-feet per year. Energy costs for groundwater pumping are estimated as $0.132 per acre-foot per foot of lift. The land surface is at an elevation of 385 feet above mean sea-level, while the bottom of the aquifer is 233 feet below sea-level.
All data are reported in 1992 dollars and on a per-acre basis. We use an interest rate of 5% where applicable.
Our model leaves out many real-world features of both the agricultural economy and the hydrologic system. However, it allows one to consider sustainability over a long horizon and provides reasonable order-of-magnitude effects of water transfers on groundwater usage and management, and on crop income in the basin.
Common property usage
This scenario shows how transfers might affect groundwater stocks and usage in a non-regulated setting. Under common property usage, decisions are constrained only by available surface and groundwater supplies. Individual growers make water-use decisions that are strictly in their own best interest, ignoring effects on others. Because each farm uses such a small share of the entire groundwater stock, decisions are made in each period to maximize profits in that period without regard to the future groundwater level.
Results of the baseline scenario without regulation or water marketing suggest that, over 50 years, groundwater levels in the study basin would drop more than 200 feet. Consequently, groundwater becomes more expensive and withdrawals are progressively reduced. Farmers’ annual net profits from producing crops decline by more than $65/acre during this period, due to increased pumping costs and reduced withdrawals that induce more expensive irrigation systems, changes in cropping patterns and possibly reduced yields.
Next, our model considers the impact of water marketing-specifically, a transfer away from the region of either 10% or 20% of the original average yearly supply of about 2 MAF. These amounts bracket the estimated yield of the proposed transfer and storage projects, allowing for offsetting increases of imported surface water in wet years. Under either the 10% or 20% transfer, growers rely more heavily on groundwater, thus further lowering the water table below the level it would otherwise be. For example, compared to the baseline, the 20% transfer lowers the water table by 37 additional feet after 20 years and by 70 additional feet after 50 years, as shown in Figure 1.a. (Results of the 10% transfer are half of the 20% figures. For that reason, only the 20% figure is given here in most cases.)
The effect of water transfers on withdrawals, shown in Figure 1.b, is more complicated. With water transfers, withdrawals increase in the early years, but at some point the extra pumping lift makes groundwater so much more expensive that eventually, in possibly 35 to 40 years, annual withdrawals become less than they would have been without the water transfer.
Net profits from crop production also fall with the transfer (Figure 1.c). At the 20% transfer level, annual net profits are reduced by 14.4 million dollars or $16 per acre per year after 20 years, and $26 per acre per year after 50 years. These figures do not include revenues from the water sales, which at a minimum would be sufficient to offset these losses.
An overall conclusion about the effects of water transfers on unregulated, common-property groundwater usage depends on the perspective. Even with a 20% transfer of surface supplies, short run impacts on net revenues from crop production and on the aquifer are modest. Over a longer term, the effects could be substantial.
The results also suggest that both the groundwater effects and the time horizon are critical in determining the break-even price for water transfers. At the 20% transfer level, a $16 per acre annual reduction in net profits at the 20 year time horizon implies that farmers would break-even at a price of $41 per acre-foot per year. Over a 50 year time horizon, the same deal-a permanent transfer of 20% of their water-would require a break-even price for transferred water of $66 per acre-foot per year.
Both prices are in the range of those discussed for water transfers from the region, but the difference in the implied price between 20-year and 50-year time horizons is significant. This difference arises because the price implied by the shorter time horizon does not account for the longer-term impacts on the quantity of groundwater available, or the costs of using it.
|Figure 1: Effects of a 20% water transfer and groundwater management on the water table, groundwater extractions and annual profits from crop production|
Economically efficient groundwater management
To be economically efficient, groundwater extractions over time must maximize the present value of total net benefits of all users in the basin. This implies a need for some management, recognizing that one user’s behavior affects water available to others. Economic efficiency also incorporates the notion that “investments” in the groundwater stock-the cost of managed reductions in withdrawals-should earn a rate of return equal to that elsewhere in the economy. (Again, our figures do not include revenue from water sales.)
As with common-property usage, the economically efficient usage scenario causes the water table to decline over the entire time span even without transfers. However, the process of resource loss-falling water tables and increasing pumping lift-is slowed. Because this scenario accounts for the effect of current withdrawals on future pumping costs, optimal annual withdrawals are lower. In particular, withdrawals from the basin at the beginning of the period are some 805,000 acre-feet, or 0.89 acre-feet per acre per year less than those under common property usage. However, this difference declines over time. In our model, in fact, withdrawals eventually are identical under the two regimes although the water level is lower under common property usage.
The main tradeoff inherent in optimal groundwater management is illustrated by differences in net profits to crop production over time. During the first few years, annual basin-wide net profits are less under an economically efficient groundwater use system than under common property usage, due to reduced extractions and use of more expensive irrigation systems. Within the first decade, however, the results switch and economic efficiency results in higher net profits, due to lower pumping costs.
We measure benefits from groundwater management as the difference in the present value of net profits from future crop production under the two regimes-common property and economic efficiency. On average, total benefits to groundwater management for the basin are $5.64 per acre per year over the 50 year horizon. This result does not take into consideration management costs such as reaching initial agreement, monitoring and enforcement, and so on.
How is this potential payoff to groundwater management affected by the proposed water transfers? Reducing surface water use by agriculture in the basin makes the groundwater resource more valuable and so the gains from groundwater management increase. A 20% water transfer results in average annual management benefits of $6.18 per acre. Thus, the surface water transfers imply increases in the incentive to manage the resource. Overall, however, benefits from groundwater management appear to be rather modest.
Groundwater withdrawals by individual users involve a range of impacts over time and on other users that are not considered in a common property situation. This neglect leads to less efficient use of the resource. Groundwater management can therefore potentially improve economic efficiency and hence increase net returns to all users of the groundwater basin. Even without market transfers of water, the model shows basin-wide benefits from groundwater management of $5.64 per acre per year. Furthermore, water transfers increase both stress on the aquifer and benefits from management. We find additional increases in management benefits of 5% and almost 10% for the two levels of water transfer.
This analysis is only a first step in evaluating the economics of water transfers on groundwater systems. Since our model considers only one dimension of the problem (quantity effects on the aquifer), the analysis leaves out much of the complexity and variability of both the economic and hydrologic systems. Not addressed, for example, are the additional complications brought on by random fluctuations in annual surface flows, water quality and subsidence issues, regulatory costs, and effects of less-than-perfect regulation and regulatory inefficiency. Further, we hypothesize either a 10% or a 20% water transfer rather than modeling the water market explicitly. Finally, the study only considers one area (Kern County). While the results are perhaps relevant to other heavily-overdrafted agricultural basins in the southern San Joaquin Valley, conditions elsewhere in California differ widely and different results might be obtained there.