This appendix provides detailed information on the data sources, estimation methods, and results of the statistical analysis of the effect of export restrictions on county water sales and exports presented in Chapter 5.
Data Sources
Annual County Water Sales and Annual County Exports These series are developed from the water transfer database presented in Chapter 2 and Appendix A. Annual county sales are defined as the sum of all short- and long-term transfers by county water users in a given year. For sales by cross-jurisdictional water districts, the approximate share of the district in each county has been attributed to that county.
Annual county exports are the sum of transfers not destined for other water users within the county. Environmental water sales were
considered as exports from the county. Although this water is most often used for habitat or instream purposes within the region, it rarely is under control of users in the county of origin. For water districts with multiple jurisdictions, we considered the transaction to be “in-county” if the purchaser was in any of the district’s counties. As such, the exports category unambiguously includes only those transfers going from a user within the county to a user somewhere else in the region or state.
Because the coverage of transfers within the CVP’s Friant group was not consistently available in all years, we have excluded internal Friant transactions from the sales data. This concerns five San Joaquin Valley counties: Merced, Madera, Fresno, Tulare, and Kern. Since members of the Friant group are effectively exempted in the counties with
ordinances, this should not pose a problem for interpretation of the
results. Trades between Friant members and other water users are included.
County Export Restrictions
We consider that counties have an export ordinance in operation beginning in the year of adoption, as indicated in Table C.1. There are two exceptions. Because Kern County’s ordinance applies only to the relatively unimportant desert and foothill region in the southeast, but not to the San Joaquin valley portion of the county where population, agriculture, and surface water entitlements are all concentrated, we have considered Kern to have no ordinance for the purposes of this exercise.
The second exception is made for Glenn County, whose export ordinance was effectively removed in 2000, when the new basin management ordinance was adopted.1 The 2001 season is the first during which Glenn water users worked under the new system.
State and Federal Policy Environment
The general effects of an improved trading environment arising from state and federal policies to facilitate transfers are captured by a time trend.
Agricultural Water Demand
The model uses three measures of agricultural water demand:
average county-level prices for annual crops (defined as all field and horticultural crops), the acreage under annual crops, and the share of perennial crops in total nonrange acreage. All three series are constructed using county agricultural statistics from the California Agricultural Statistics Service databases.2
____________
1Although the 1990 ordinance remains on the books, the numerous persons interviewed in Glenn, including two county supervisors, considered that the new ordinance has supplanted it for operational purposes.
2The annual crop price is calculated using the county’s prior year output data, valued at the statewide average price for the current year. This captures the notion that the farmer has an idea of the average market price for the coming season and can calculate what he would earn by farming the same crop mix as in the preceding year. The series is deflated using the western states urban consumer price index, with 1992 as the base year.
Ideally, we would measure the value of crops on a per-acre-foot basis to capture the water
In principle, we would expect water sales to be inversely related to the average level of crop prices, which reflect the value of using water in agriculture. On average, real price levels have been relatively flat over the period, hovering around $160 per ton in 1992 prices. The range across counties is quite large, however, with averages over $300 per ton along the south and central coast and $100 per ton or less in parts of the Central Valley.
As with prices, there has been little movement over time in the average level of annual crop acreages, although the cross-county differences are huge, with at least several hundred thousand acres in Imperial County and most San Joaquin Valley counties (and close to 1 million acres in Fresno), and fewer than 50,000 acres along the coast.3 In part, this range reflects differences in the overall scale of agricultural operations across counties; in part, it reflects a much higher share of perennial crops (fruit trees, nut trees, and vineyards) in some counties.
The values span a high of over 90 percent of all nonrange farmland in Napa and over 50 percent in San Diego, Ventura, and Madera, to only 1 percent of cropland in Imperial. Over time, there has been a mild upward trend in tree crops as a share of the total, moving from 22 to 26 percent on average.
Because farmers can make adjustments in annual crop acreages fairly easily as a function of water availability, we would expect water sales to be positively related a county’s crop acreage. Conversely, because a higher share of tree crops in total acreage introduces less flexibility in water use, we would expect tree crop share to be negatively related to water sales.
________________________________________________________
intensity of the crop mix. This would require making assumptions about the irrigation technology used in each county for each crop, however. If anything, the use of a per-ton measure probably dampens the effect of this variable, since low-value crops also tend to be those with a relatively high level of water use. Annual crop acreage includes all farm acreage except perennials and rangeland (i.e., including irrigated pasture). The share of tree crops is calculated as the share of perennials in total nonrangeland farm acreage.
Both acreage measures are valued at the prior year levels to account for the fact that decisions on water sales are generally made before final planting decisions.
3San Francisco, the only county in the sample with no commercial agriculture (albeit some fine gardens), has no acreage recorded and no positive crop prices.
Residential Water Demand
County population levels are used to account for residential water demands. The source is the annual population series from the California Department of Finance, based on updates from the 2000 Census
(Department of Finance, 2001). Other things equal, we should expect counties with higher populations to be less likely to sell water.
Water Supply Conditions
Annual deliveries of project water from the CVP, the SWP, and the Colorado River Project are captured in two measures: senior rights and junior rights.4 The senior rights category includes those deliveries with a high degree of reliability, by virtue of the seniority of the contractors.
This includes the CVP settlement contractors in the Sacramento Valley and the exchange contractors in the San Joaquin Valley, the SWP’s Feather River contractors in Butte and Sutter Counties, and the Colorado River contractors in Southern California. On average, 8.3 million acre-feet are delivered annually to these contractors. Half of this volume is destined to Imperial and Riverside Counties and a quarter to the four main settlement-contracting counties in the Sacramento Valley (Colusa, Butte, Sutter, and Glenn). The only dips in supply occurred during the early 1990s drought, when CVP and SWP contractors’
deliveries were reduced by 25 to 50 percent in some years.
The junior-rights category includes the ordinary project contractors of the CVP and the SWP. On average, these projects have delivered just over 6 million acre-feet annually over the 12-year period, to a much larger number of water users. Only two counties in the sample, San Francisco and Yuba, do not have project contractors. Project deliveries have generally been much more variable from one year to the next, particularly for contractors south of the Delta.
____________
4For the Colorado River Project and the SWP, actual delivery data were used. For the CVP, we applied the annual allocation rules by type of project contractor (settlement contractors, north of Delta and south of Delta service contractors, and Friant class 1 and 2 contractors). Because California received surplus deliveries of Colorado River water for the entire time period under consideration, the Metropolitan Water District of Southern California is included as a senior rights-holder. Under California’s official allocation of 4.4 million acre-feet, this agency would have its supplies cut back, as the junior rights-holder.
In general, we would expect counties with higher water deliveries to be more active in the water market. By the same token, individual counties should be more likely to sell in years when their deliveries are higher.
Unfortunately, detailed data on other water supplies—from autonomous projects and from groundwater—are not available. We do have a general indicator of the quality of the water year, however, in the form of the state’s most important rainfall measure—the Sacramento Valley 40-30-30 index.5 Since market demands and water prices are likely to be higher in dry years, we would expect this indicator to be negatively related to sales. The period under review contains an equal number of dry and wet years (Figure 2.1).