April 2006 | Online Edition
DBEDT Analysis Shows Gas Cap
Cost Consumers $54.9 Million
The following is an excerpt of a letter sent by the state Department of Business, Economic Development and Tourism to Rep. Kirk Caldwell regarding the findings of any analysis conducted by DBEDT to estimate the effects of the state’s wholesale gasoline price controls on Hawaii’s consumers. See the entire letter at:
http://www.hawaii.gov/dbedt/info/energy/publications/gascapanalysis06.pdf
Summary Explanation of Analysis & Results: This analysis reveals that the gasoline price cap cost consumers as much as $54.9 million ($54,855,001), between September 1, through January 2006.
The attached summary (Est_Impact_Of_Gasoline_Price_Updated_Feb-06.xls, an MS Excel file), and the additional 3 worksheets are the program files showing the determination of the estimated average retail price for regular unleaded, mid-grade, and premium (one file for each grade) – if there had been no price cap. The program file also contains charts (Charts 1 & 2) to graphically illustrate the comparative differences of actual and estimated retail regular gasoline prices on Oahu. The program file for the regular unleaded includes the pivot table that was used for all three grades.
The method used in determining the estimated impact:
1. Database used: Recorded Average Daily Price for each gasoline grade (regular, mid-grade, and premium) from January 2002 – August 2005, from OPIS/AAA.
2. The average daily price is summarized into monthly average price, and plotted.
3. Linear trend curve was used to estimate the monthly average price had there been no price cap for each gasoline grade.
4. The difference between the estimated monthly average price for September through January 2006, and the recorded monthly average price for the same months were multiplied by the gasoline sales volume from EIA. See footnotes on the summary file which details this procedure.
Peer Review of the Analysis. In an effort to have this methodology validated, DBEDT asked three well-respected professional economists familiar with energy markets, to conduct a “peer review” of DBEDT’s analysis (the reviewed version had data through the end of December 2005). All three reviewers were very responsive, and agreed to share their views publicly. We have attempted to summarize reviewers’ detailed technical comments, which are available.
In essence, the reviewers offered thoughtful comments and useful suggestions, which provided staff with valuable insights. The three reviewers agreed that in any such analysis different methods would yield different results; i.e., the reasonableness of such estimates and methods hinge on the reasonableness of the assumptions used. Two of the reviewers found that DBEDT’s method was sound, and that DBEDT’s estimate was reasonable, with one of them suggesting it actually appeared low.
One reviewer suggested that DBEDT’s estimate was high in comparison to his own. This reviewer did not use the same type of method as DBEDT. Rather, this reviewer determined and compared the differences between U.S. average and Hawaii average prices over the pre-cap period (beginning Feb 2003), and again over the gas cap period through Dec 2005). DBEDT found that this reviewer’s analysis did not incorporate what he acknowledged is the essential aspect of estimating the gas caps’ cost; i.e.: What would gasoline prices have been had the caps not been in place?
Our perspective is that this reviewer found no fatal flaws with respect to accuracy of DBEDT assumptions, data, logarithms; i.e. specific to our approach. This reviewer did question DBEDT’s use of monthly average price data, which he believed could dampen price fluctuations. However, when DBEDT plotted the daily average price data, it mimicked the linear pattern of the monthly average data and no pronounced fluctuations were found.
Finally, we note that the conclusions regarding the price impact determined by this analysis is limited by actual experience with the price cap formula, five months. The longer the period of analysis, the greater certainty of any conclusions. Therefore, within the context of assumptions and limitations stated, DBEDT stands on its analysis, which we believe is sound.
|
