How to Select the Best Portfolio of Oil and Gas Projects

Abstract

Abstract The traditional analytical tool for selecting portfolios of financial assets is Markowitz's mean-variance model. The final product of this model is the efficient frontier. Markowitz's approach comes up with an infinite set of optimal portfolios in terms of risk and return (called the efficient frontier), but the mean-variance model does not recommend the best portfolio. This paper proposes an extension of the mean-variance model by the inclusion of a three-step approach:an estimation of the risk and return of each project;the estimation of a correlation between returns of each pair of selected prospects; and,the inclusion of corporate goals beyond simple economic return; that is, budget and operational constraints. We show that the selection of the optimal portfolio depends on the diversification level of the investor and the costs of financial distress faced by the oil company. Introduction The oil industry has currently been facing high volatility in price, plus technological and geological uncertainties that reduce cost predictability. Since these uncertain factors are all present in most oil and gas projects, they represent a clear motivation for the development of new tools for improving the rationality of decisions in the process of capital allocation. Some of these tools are for the selection of the best portfolio of upstream projects when the main difference between any two portfolios is risk and return. At this point, we wish to make it clear that the screening or selection of oil projects is a previous step and will not be considered in this paper. We assume that the problem consists of finding the best portfolio of already screened oil and gas projects. Not long ago, the selection of portfolios was based on indicators such as ranking and cut, internal rate of return and payback period. When we use any or all of these indicators, the conflicting, but necessary, risk-return trade-off is not considered, so that the optimal decision could have a very high return. On the other hand, however, its risk level may be unsupportable by management. In the past, in the job of finding and producing hydrocarbons, risk was always present, but there was not a systematic pressure policy towards its understanding, modelling and quantification. Presently, in most oil and gas companies, management is not only under pressure to deliver good indicators for the shareholders, but also to estimate the associated risk level. In the context of this paper we consider risk analysis as an integration of three steps. First is the identification of uncertain variables. Second is the probabilistic modelling of these uncertain variables. Third is the use of Monte Carlo simulation in order to get the frequency distribution of response variable, such as net present value (NPV). One reason for the widespread demand for risk analysis in the upstream industry is that trends of resource availability have pointed out that the largest remaining reserves of hydrocarbons are primarily situated in ultra-deep environments where the potential for return is higher but the risk is also huge.