Most future input prices (unit costs) in the upstreampetroleum industry are not known with certainty. This canhave important effects on asset value and management. Onecriticism of most past petroleum industry applications of realoptions analysis (ROA) is that they have neglected to considerthe effects of this type of uncertainty. This paper explores afew of the issues that arise in discounted cash-flow (DCF) andROA asset valuations, if a specific type of input priceuncertainty is considered.In particular, we look at the effects of the correlation ofunit costs with output petroleum prices. Unit drilling costs,for example, tend to be high/low if petroleum prices arehigh/low or have gone up/down unexpectedly. The first "rent"effect occurs if the market for drilling services is notcompletely competitive. The second "quasi-rent" effect occursif it takes time for the suppliers of these services to adjust theamounts they supply in the face of unexpected changesin demand.We first examine, using both ROA and DCF methods ofanalysis, some relatively simple assets where the asset cashflowdependence on input and output prices, and of the inputprices on the output prices, is linear. Two points are made.First, for a set of given input price expectations, DCFestimates of the value of these assets will be independent ofthe level of input price uncertainty, unless the discount rate isadjusted to reflect the change in risk. ROA estimates of valuewill automatically pick up the effects of different levels ofrisk. If the unit costs are generally correlated with the macroeconomy(as will be the case with a rent effect and most quasirenteffects, if the output prices are so correlated), for a set ofgiven input price expectations, a greater level of input priceuncertainty would decrease the estimate of the value of thecosts in the asset cash-flows, because of the greater riskdiscounting that it induces. This would increase the ROAestimate of the value of the asset as a whole.Second, it has been conjectured that a greater similaritybetween revenue and cost uncertainty, by making revenue andcost discounting more similar in ROA analyses, will tend tomitigate some of the differences between the structure of ROAand DCF value estimates. This would decrease the benefits ofa shift from the use of DCF methods of analysis to the useof ROA.We analyse variations of two previously publishedexamples of analyses of assets with linear cash-flows toexplore these matters.We then examine an asset where there are non-linearitiesin the dependence of asset cash-flows on prices. These maybe due to non-linear price models, non-linear taxes orflexibility in project management. In the specific asset weanalyse, which is also based on a previously publishedexample, the non-linearities arise from flexibility. In this case,we find some similarities and some differences in structure ofthe effects of input price uncertainty on asset value.The example is of an oil field in the final year of itsdevelopment lease, where there is an option to sanctiondevelopment immediately, or to appraise or wait for anotheryear and decide then between development or abandonment.At sanction, there is a production capacity choice to be made,and after sanction there is an annual abandonment option.Again, two points are made.First, if the input price expectations are roughly the same,input price uncertainty of the simplest kind (stemming from alinear "rent" effect) again increases ROA estimates of assetvalue and leaves the DCF value estimates roughly the same.Second, rather than mitigating the differences betweenDCF methods of analysis and ROA, cost uncertainty in thisexample accentuates the losses from using DCF methods ifROA should be used. Unit cost uncertainty interacts with thedifferences in the methods of value estimation to make themanagement policies suggested by DCF and ROA moredifferent in the presence of this uncertainty.All of this suggests that the industry should keep track ofthe uncertainty in its input prices, and their correlation withoutput petroleum prices and with each other. This papershows how this would be useful for project analysis. It wouldalso be useful in the development of appropriate riskmanagement policies.
展开▼
