The original models were developed to analyze a rate structure with two blocks, so they had to be generalized to fit the Alameda County Water District (ACWD) rate structures in effect over the period from 1982 to 1992.
3 See Table 2 below for ACWD's prices over the period 1982-92.
Part of the data collected by ACWD for its demand analysis had to be retrieved from archived paper records (Brown and Caldwell 1992).
10 This information was obtained from ACWD's response to the water agency survey reported in Dixon, Moore, and Pint (1996).
The blocks were also chosen to be evenly distributed across the ACWD service area, including some blocks in the eastern hills, and to reflect the distribution of property values for the entire ACWD area (Brown and Caldwell 1992, 2-11 to 2-12).
12 In addition, I obtained 1980 and 1990 census data regarding household size and composition, median incomes, etc., at the block group level for the census tracts corresponding to the ACWD service area and matched these data with the household data using the geographical information system ARC/INFO.
21 It is not clear whether the discrepancy between the 54,000 observations in the tax assessor data and the 63,000 single-family accounts reported by ACWD is due to multiple accounts at the same household, incorrect categorization of accounts by customer type, or incorrect codes in the tax assessor file.