|
|
|||
|
|
|
|
|
|
||||
|
Company History Key People News Newsletter Customer Profiles Contact Us Employment |
Category: Economics/Geology Estimating the Future Price of Petroleum
"When I do either," Capen says, "I always gather data in hopes of learning something interesting." One task for Data Desk is to look at survey data on estimating future oil and gas prices as a function of whether the estimator is a producer or banker or consultant. Does company size matter? Does size of deals worked on make a difference? Are there functions of the input data that we can use to discriminate among the various professions found in the oil and gas evaluation business? For several years he has been doing this kind of analysis for the Society of Petroleum Evaluation Engineers. He primarily performs tests of hypotheses and uses Data Desk's flexible data management tools to assist that process. Capen teaches a short course in petroleum risk assessment with his associate in PetroRisk Partners. For this work he uses classroom exercises to demonstrate how the variance of a set of estimates changes with the acquisition of information and whether individuals can break away from the "anchoring" position of their first estimate. If he sees highly correlated estimates early in the information-gathering and virtually no correlation near the end, he surmises that the information gathering strategies have worked. He uses regression analysis supported by linked scatterplots for this work. Another Data Desk work-in-progress is Capen's investigations of how cognitive techniques affect estimates of future oil prices. He is attempting to show how using the right brain (intuition) and the left brain (analysis) can alter initial and subsequent estimates. He's analyzing data collected on a national lecture tour a few years back. He has about 2,000 responses so he can afford to use some data to ask a lot of questions and then come back and actually test hypotheses on clean unused data. He wants to see if there are effects based on the size of city one works in--big cities tend to have staff people, while smaller towns have field operating people. Does it matter whether one is a geologist or an engineer? Some argue that geologists tend to be right-brained while engineers concentrate on their left brains. He will also be checking on the degree to which their initial estimates of future oil prices are tied to the price at the time people make their estimate. Capen is asking Data Desk to help find good linear predictors using regression. For example, maybe one's estimate of future price is today's price + $5. If so, that would show another common bias identified by psychologists called "availability." People use readily available information (like what appeared in the morning's newspaper) to help them judge the future even though in the past that information has never turned out to be particularly useful. |
Name: Ed Capen Affiliation: PetroRisk Partners
|
|||
|
|||||
|
|
|||||
Having received the American Institute of Mining and Petroleum Engineers'
Mineral Economics Award and the Society of Petroleum Engineers' Economics
and Evaluation Award, Ed Capen has been recognized for his work in
petroleum economics and is still quite engaged in what has been a long
and rewarding career in the petroleum industry. He began as a research
physicist, did some graduate work in statistics, then moved into research
project evaluation, helped develop a practical and highly profitable model
for competitive bidding, and along the way spent a few years in corporate
finance. Since retiring in 1992 from his position as
Distinguished Management Advisor at Atlantic Richfield Co., he continues
a part-time but lively practice as a consultant and teacher. Data
Desk comes in handy for both roles.