Why explore for oil when it is cheaper to buy?

Abstract

This article uses results of independent US oil companies to examine their decisions in a high-risk environment. When these companies seek to replace oil production, the available choices fall into two broad classifications, each with its own distribution of expected costs and returns: explore for oil; or buy proven oil reserves. Firms prove risk-sensitive in their decisions as the balance struck between building reserves by acquisition and by exploration responds to firm characteristics. The crossover from risk embrace (exploration) to risk aversion (acquisition) occurs when the probability of success from the more risky strategy drops below about 15%. This matches the behaviour of decision makers when facing risks as diverse as acquisitions and racetrack betting. Shareholders, however, do not support risk-taking for its own sake, although they bid up the price of successful risk-takers. This reveals a divergence in goals between principals and agents; and an inverse relationship between risk-taking and return as measured by shareholder value.

Acknowledgement

This analysis is based on ideas suggested by my work in the oil industry and has subsequently benefited greatly from comments and suggestions by numerous colleagues.

Notes

A barrel of oil contains 42 US gallons, or about 159 litres. ‘Oil equivalent’ refers to industry practice of including reserves of natural gas in total hydrocarbon reserves by converting the gas to a volume of oil with the same energy content.

This matches the conclusion of Pesaran (1990) that the price of oil required to justify exploration in the UK in the 1980s was several times higher than the value of oil in the ground.

These calculations are all pre-tax as no data were available on the firms' tax paying position.

Assume the decision results in either success or failure with a 0.2 probability of success. Each event will be similar to a binomial distribution with mean 0.2 and standard deviation of

. To have a greater than 50% probability of success, the decision maker needs to be in the top 23% of the distribution.