Evaluating Open Oil’s Financial Modeling of Guyana’s 2016 PSA – 7

Today’s column continues with my critiques of Open Oil’s financial modeling of Guyana’s 2016 Production Sharing Agreement, PSA. To sum up the discussion at this stage, the first critique under discussion urges that a basic obligation of financial modeling studies is to amplify their reporting on the limited predictive power of such modeling simulations and their results. I do not want to involve non-modelers and/or my more general readership into the esoteric realms of financial modeling. The fact is however that the model used in the Open Oil exercise is one based on spreadsheet modeling, utilizing Microsoft Excel. 

This approach has encountered widely recognized problems among analysts. This circumstance indicates itself in the title of Fairhirst’s (2009) study: “Six reasons your spreadsheet is NOT a financial model”. Such criticisms point out the “unrealistic implicit assumptions and internal inconsistencies” of several spreadsheet based models. Such observations also support the view expressed in my previous column that, all too often, modelers fail to state explicitly the full assumptions and limitations of the data, which they are inputting into their models.

Despite such observations though, I firmly believe that “honest modeling” adds much to the debates on the role of petroleum, and perhaps more broadly – the extractive sector, in Guyana-type economies. It certainly, however, does not close the debate or indeed yield finality and settlement of disputes on the role of the petroleum sector in development.

Second Critique: Risk & Uncertainty

My second critique is that, I do not find the Open Oil model has either adequately handled, or sufficiently acknowledged several known unknowns, which are likely to emerge as Guyana’s petroleum project unfolds over the next 22 years. These known unknowns are both endogenous and exogenous to the model satisfactorily acknowledged. I shall illustrate this observation, with two examples, one exogenous and the other endogenous to the model.

Close observers of Guyana’s current situation and recent political economy would concede, I believe, Venezuela’s geo-strategic border threat constitutes, by any measure, the greatest existentialist threat to its coming time of oil and gas production and export. As such, that threat drives the dynamics of Owner (Government of Guyana) and Contractor (Exxon and parties) negotiated relations going forward. Indeed, the Government of Guyana has admitted to this circumstance. National security considerations have guided Guyana’s choice of the Contractor and its relentless determination to sweeten the pot in order to ensure this group continues to lead the development of its petroleum sector.

On the endogenous side, the model does not address the consideration that both the Owner and Contractor will be “learning by doing”. That is, as the project advances, the Owner (Government) will be “learning” the business of “doing” (operating) under the PSA. This will improve Government’s capacity to monitor costs, improve tax collection, and incentivize new investments, whether in the form of local context, complementary linkages, scale effects, or skills enhancement.

“Learning by doing” is not only likely to lead to positive outcomes for the sector. Based on the historical record of petroleum sectors in poor countries, much of this “learning” will be located in the practice of corruption and deception! Contractors will grow in their experiences of dealing with regulators. Historically, all too often a fertile environment for corrupt practices has flourished to the disadvantage of both partners to the contract. One recent positive occurrence has occurred in Guyana to confront these challenges. That is, the Government has indicated it will establish a national oil company (NOC), which could have a significant impact on the dynamics of its PSA.

All the issues raised in this and previous columns about Open Oil’s financial modeling are integral to Guyana’s development, based on its recent petroleum discoveries. It is difficult therefore, to envisage a useful model of this sector that does not explicitly incorporate these elemental factors.

Risk & Uncertainty

The above discussion is symptomatic of the model’s weak treatment of risk and uncertainty. Although addressed in the model, particularly in terms of market price and field size, their treatment remains too limited. In earlier columns of this series, I have often repeated that, the petroleum industry distinguishes itself from other businesses by the level of risks it carries. This feature must therefore affect modeling of the industry.

To recap these risks are quite wide ranging. They include geological, geo-strategic, political (expectations) technical and technological (highly location dependent, for examples, in Guyana “deep offshore”); economic/financial (prices, cost, revenue flows, supply and demand) fiscal revenue patterns, high capital-intensity, physical properties of the, oil “find”, as well as environmental. Readers should note this listing includes all of A. Beattie’s “5 biggest risks faced by oil and gas companies”, 2018.

Third critique: Simulation or Evaluation

My third critique relates to the distinction, which is central to project modeling. That is, studies are either projection-based (“before the event, prospective-based”). These take place before the project is completed and its effects felt. This type of study is termed ex ante. It simulates the projected behaviour of the project through time. Studies are also ex post. These take place as evaluations of the project or impact assessments. The truth is however, that such exercises conducted on a variety of projects, which I have come across support the conclusion: “Most ex ante models had a tendency to overestimate the benefits and underestimate the costs”. This has therefore, led to the view that ex ante studies have “weak credibility” and the advice is “Policy makers should thus treat [their] results with … appropriate skepticism”.

While I agree with this and support caution, ex post studies also face severe limitations. The most striking of these, is the position where (due to privacy and confidentiality consideration) access to the requisite statistics and information are not forthcoming.

Conclusion

Next week I shall round off my evaluation of Open Oil’s modeling of Guyana’s 2016 PSA. This task has taken much longer than I had originally intended. I shall in conclusion therefore focus on three topics, namely 1) the information gaps identified by the Author in model preparation 2) the “treatment” of Government take, given my earlier discussion of this topic, and 3) national versus project-based financial/economic modeling in Guyana-type economies.