Risk Analysis
Posted by D Nathan Meehan October 22, 2011
Category: Oilfield economics, Reserves   |  Tags: ,

The application of Monte Carlo simulations is one of the most important tools that reservoir engineers can master in the field of modern risk analysis.

Nearly all business decisions are made under conditions of uncertainty. Decision making under uncertainty implies that adequate information for assuring the right decision is lacking and two or more outcomes are possible as a result of the decision. Petroleum exploration is a classic example of decision making under uncertainty. The following discussion of risk analysis will be phrased in terms of petroleum exploration, although there are applications to manufacturing, marketing and service company decisions that are equally well suited. It is recommended that exploration wells and programs are evaluated using expected value economics which account for the probabilities of realizing various outcomes.

Risk analysis provides a more thorough and comprehensive approach to evaluate and compare the degree of risk and uncertainty in a project than the methods previously discussed. The intended result is to provide the decision maker with more insight into the potential profitability and the likelihood of achieving various levels of profitability than traditional methods of investment analysis.

Conventional methods of analysis usually involve only cash flow and rate of return considerations. The added benefit of risk analysis to the decision maker’s process is the quantitative review of risk and uncertainty and how these factors can be incorporated into the process of developing and implementing investment strategies. Risk and uncertainty cannot be eliminated from the business decision maker by such analysts, or by any other method of investment review. The advantage of decision analysis is its use as a tool to evaluate, quantify and understand risk so that management can devise and implement strategies that will allow the company to minimize its exposure to risk.

Decision analysis is a multidisciplinary science. It involves aspects of many different disciplines, including probability and statistics, economics, engineering, geology, finance, etc. Certain statistical methods of decision analysis provide excellent ways to evaluate the sensitivity of various factors in a risk-based economic analysis.

Several petroleum industry methods for handling risk are briefly described below.

  • Arbitrary decision minimums – In some instances, risk is treated by raising the minimum DCFROI to accept the project. For example, a normal hurdle rate of 15% could be arbitrarily specified at 30% for projects with higher levels of risk or uncertainty. Such a procedure reflects the need to have the return commensurate with the degree of risk.  Although directionally correct, this method does not explicitly consider the varying levels of risk between competing investments. This is not generally recommended.
  • Allowable dry holes – Some express relative degree of risk in exploratory drilling projects by a parameter defined as the allowable dry holes or dry hole capacity. In this approach the analyst computes an estimated NPV that would result from a prospect if successful. This NPV is then divided by the cost of an exploratory dry hole. The result is a multiple of how many times the present value cash flow from a discovery exceeds the dry hole costs. This approach does not yield any information about the probability of discovery but gives management an insight as to how many exploratory wells it could afford to drill based on the value created by one discovery. As such, it provides a relative indicator of the affordable risk of competing exploration areas. This approach can be useful in certain cases but it has several limitations. It does not explicitly account for estimated probability of discovery nor does it provide a specific ”go/no go” decision criterion. Further, it does not tell us how much greater than some specified number the allowable dry holes multiple or success capacity must be an order to achieve an objective level of profitability.
  • Simulation Techniques – The concept of simulation allows the analyst the option of describing risks or uncertainty in the forms of distributions of possible values of the uncertain parameters. These distributions are combined by a computer (Monte Carlo simulation) to yield the distribution of the possible levels of profitability that could be expected from an investment opportunity. The application of Monte Carlo simulations is clearly one of the most important tools that reservoir engineers can master in the field of modern risk analysis. Additional discussions of Monte Carlo simulations will be provided in my November 19 blog post.
  • Expected Value Economics – This cornerstone of decision analysis is the expected value concept, a method of combining probability estimates with quantitative estimates that results in a risk-adjusted decision basis. The concept is not meant to be a substitute for manager’s judgment but rather a tool to allow evaluations and comparison of the possible outcomes of different investment alternatives.

The decision to drill an exploration well can result in a dry hole, discovery of a giant field, or something in between. Each outcome has some likelihood of occurring, yet no outcome is certain to occur.  Many take the view because of the inherent subjectivity involved in assigning probability estimates, expected value analysis has little to offer. There is no doubt that assigning probabilities to the possible outcomes of a drilling prospect is difficult. Sometimes is not even possible to define all the possible outcomes. The benefit of expected value analysis, however, is confirmed by its application to repeated trials. If a firm consistently strives to maximize the expected value of many projects over the long run, it can be shown that the firm will do better utilizing risk-weighted value economics.

1 response | Add Yours


Rodolfo Galecio says:

Dear Mr. Meehan is great to see your blog active again! I would to discuss how to get the probability distributions needed to built a simulation model, indeed I think this is one of the most challenging issues, it depends on our knowledge on the parameter being modeled but also on subjective judgments about uncertainty. Let me remark one point, applying the expected value criteria in the long run requires some conditions, for example what Paul Newendorp calls “an independent attitude toward money value”, there’s food for thought on this topic!

You must be logged in to post a comment.

We’re here to help
D Nathan Meehan