15. Sales
SCWC and ORA do not agree on forecasts of the annual sales per commercial customer, as summarized in Paragraph 3.02 of the Stipulation. In estimating sales for the commercial class, SCWC used a monthly regression model, while ORA used the Modified Bean method, which incorporates annual data. As discussed below, we will adopt SCWC's estimates for commercial sales per customer.
15.1. Discussion
The monthly regression model used by SCWC produces results which are statistically more significant than that produced by ORA's model, which failed the F-statistic and Durbin - Watson tests. Additionally, the Water Division46 and Commission decisions47 have encouraged the use of a monthly regression model. ORA does not address these points, which are significant in considering a return to the use of the Modified Bean Method. For these reasons, we adopt test year sales per customer for the commercial class, as forecasted by SCWC.
ORA's criticisms of the monthly regression model lack specific evidence and analysis. For instance, in criticizing SCWC's conservation factor, ORA states that the conservation factor used in the utility's model is an unreliable variable in that it is not measuring a phenomenon known to affect consumption. ORA does not provide evidence that shows the factor is unreliable, does not quantify the effect of using the factor, does not analyze the effects of modifying the value of the factor, does not indicate how ORA specifically addressed conservation, and does not provide an alternative method for addressing the effects of conservation. Such information would be helpful in evaluating the merits of ORA's position.
46 See Exhibit 54, which quotes the Commission's Water Division document, "Water Regulatory Policy," prepared in August 1997, where it is noted that sales forecasting is an important part of GRC proceedings and that, "[i]n 1992, the Modified Bean method was replaced with an Econometric Model which expanded the statistical data information and calculations to include additional variables." 47 See D.94-06-033, 55d CPUC 2d at 190, where the Commission noted that the Division of Ratepayer Advocates had introduced an econometric forecasting method in two water company rate cases and was "encouraging all Class A water companies to use this method instead of the more traditional modified Bean forecasts." Also, the Commission stated that "[b]ecause it produces more accurate estimates of future sales than the Bean method, the econometric model can reduce utilities' risk of underestimating (or overestimating) sales revenue."
