April 2nd 2010

We have to commend Government’s recent efforts to loosen up a bit over the now contentious issue of Ghana’s predicted oil revenues. The question is whether it’s opening up enough.

A couple of days ago, the Ministry of Finance’s Resource Tax Advisor, Dr. Amoako Tuffuor, granted an interview to Joy FM’s Kojo Oppong Nkrumah on the above-mentioned subject, in which he appeared to have outlined, albeit only in brief detail, Government’s approach to projecting estimates of national earnings from present (and future?) oil finds.

Dr. Amoako’s efforts, far from settling the issue, have only raised deeper questions about the seriousness with which government of Ghana is investing official time and commitment towards designing a fiscal framework that responds effectively to the needs and threats of an oil economy. The following, judging by Dr. Tuffuor’s responses to Kojo Oppong Nkrumah’s questions, appear to be the elements that together constitute official government thinking on the subject of prudently anticipating revenue outcomes from oil in order to more soundly plan how to spend any such proceeds.

1. The calculations are based on the entire anticipated lifetime of the Jubilee field.

2. The calculations ar e based on estimates of the total “recoverable” petroleum resources in this field.

3. The key determinant of proceeds from above calculations is the “international price of oil”.

4. Wide ranges of earning estimates are acceptable for fiscal planning purposes.

We consider each of these principles less rigorous than should be the case and within certain policy contexts actually dangerous.

But firstly, we need to establish clearly why it is important for our country’s financial managers to make rigorous forecasts of oil revenue prior to and throughout actual production.

1. If our intention is to put some of the oil proceeds in interest bearing instruments or other commercial investment vehicles we can only be successful if in our search for opportunities and negotiations with counterparties we have a good idea of what levels of investment on our side are at stake.

2. Many truly effective social investment programs, such as for instance building technology transfer mechanisms between research institutes and industry, have a lifecycle longer than the annual budget round. If oil revenue is likely to be a key source for such investments, then it would be important to know in advance, to a reasonable degree of accuracy, how much money would be available over the multi-year cycle of such programs. We must at least be able to predict with some accuracy what the MINIMUM inflow shall be in order to budget appropriately for the scale and scope of such programs.

3. In the same vein, government borrowing and lending may be predicated on future flows of oil revenue. For instance, government may decide to rack up debts for two years running in anticipation that budget surpluses may be possible in year 3 and 4, thus neutralising the real impact of such debts. Such fiscal techniques can only be prudent when government is properly advised about, at least, the minimum expected inflows from the oil sector. There are of course several reasons why rigorous and evidence-based forecasting is critical, but the above-listed should suffice for the moment in underlying the absolute importance of not taking this exercise lightly. We will now return to Dr. Tuffuor’s implied principles and state our objections thereto.

*24-Year Lifetime*

1. If our hearing was sound, then Dr. Tuffuor and his team believe that Jubilee shall be active for 24 years, produce 500 million barrels over the period, and that the average price of oil shall be $100 per barrel. Once again, as we have had occasion to complain time without number, no detailed analysis of these assumptions have subsequently followed these statements. It is not clear whether some projections have been made regarding the prospectivity (likely output) of other finds beside Jubilee that shall come on stream over the period and how such finds shall impact these figures.

2. The main point of course is that the combined wisdom of all the Winners of the Nobel Memorial Prize in Economics shall fall short in any attempt to rigorously project proceeds from a resource like oil over 24 years. The “black swan” (catastrophic) events that are possible over such a long stretch of modern time are simple uncountable. War in the Middle East. Calamity in the Persian Gulf. Extension of Delta Militancy over the Gulf of Guinea. Invention of a portable thermonuclear automobile engine. Global depression. A destabilisation of the Ghanaian continental shelf. 24 years simply give rise to too many scenarios.

3. For the reason above, it is highly doubtful that a use of historical data shall be sufficient to allow one to accurately predict the average price of petroleum over a 24-year period through any defensible technique of interpolation or extrapolation.

And at any rate the minimal and maximal figures, based upon which that average price is determined, are likely to vary so widely as to make the average annual income so derived meaningless. This is simply to say that there will be many years in which the actual annual incomes deviates so wildly from the estimated mean (average) income that the model ceases to be helpful to fiscal managers. This method clearly violates a widely accepted norm in statistics.

4. Perhaps most importantly Ghana has in the past few years developed a key innovation for fiscal management in the form of the “medium-term expenditure framework” (MTEF). The MTEF methodology is to employ 4 (in rare cases, 5) years as a planning guide for financial budgeting purposes. If, as most of us agree, oil proceeds shall constitute a significant source of national revenue in the coming years and ought therefore to be accommodated, then what we should be aiming to do is to integrate such revenue forecasts into the MTEF. That means we ought to be breaking down our projections into blocks of 5 years, each with its own explicit assumptions. It seems to us far more prudent to attempt to define assumptions and expected prevailing conditions 5 years ahead into the future and thus to make price and volume forecasts about our valuable oil resource over such a limited range of variability than to try to do the same over a period of 24 years.

5. Luckily for all concerned, the development of the Jubilee field itself has been demarcated into phases. The first phase is expected to last between 4 to 5 years (2010 to 2015 or 2011 to 2015). It seems to us more prudent therefore to restrict initial forecasting to this phase and to develop models for other phases in the clear knowledge that such models are bound to be revised closer to their time of use.

After all, besides the black swan events mentioned above and their potential effect on oil price volatility, there are other more certain developments that shall strongly impact on how much money Ghana makes from oil. Taxation and investment regime policy, the evolution of deepwater technology, changes to the profile of Ghana’s petroleum geology, changes in the value chain linking upstream and downstream, and structural changes to the oil industry itself (in the equipment market for instance) are all prime candidates. *Recoverable Resources*

1. The use of the “500 million recoverable barrels of oil” estimate over the 24-year period compounds the problems associated with the 24-year model described above. Apart from the complicated issue of predicting which other finds may come onstream, there is of course the fact that “reserves forecasting” is in itself a very complex, contentious, process all over the world. The methodologies for estimating “proven reserves” are constantly in turmoil. In our specific case, we simply do not know enough of the geology of our offshore petroleum-bearing rocks to bother making such definite estimates.

2. Luckily for us, when we restrict ourselves to the “phase of development” approach. Most of these problems disappear. That is because rather than using the “total recoverable resources” - two kilometers under the sea - method as our guiding light we shall be using the “installed capacity” approach. The latter approach is based simply on what investments have been planned, costed, and financed, to produce what amount of oil.

This is a far more rigorous approach because such investments embody market risk analysis, which is always superior to bureaucratic calculation. Simply put, Kosmos, GNPC, Anadarko and Tullow are no fools. If they are putting in, say, $4 billion to produce 120,000 barrels of oil per day (at peak level) for the next 5 years we can reasonably estimate that some figure close to this volume shall be produced over the period. They are borrowing hard cash to do this; they are thus incentivised to be more rigorous in their forecasting.

3. Also, as discussed in the previous section, the prevailing conditions, including the legal and policy regime, are likely to be more stable over the 1st phase of development. We can thus safely assume that using the likely production volume over 5 years beginning with first oil is a more prudent approach. As also is using an average price that accords more closely with the historical data encompassing those 5 years.

Thus it is perfectly reasonable to use the $60 figure previously mentioned by the Ministry of Energy and a computed volume of 144 million barrels of oil (80,000 barrels per day average instead of 120,000 barrels in order to accommodate various fluctuations, due to weather, geological variability, industrial action etc.) for phase 1 of Jubilee, and to assume that no other find (Teak, Odum, Dzata etc.) would have been sufficiently developed to begin production. Given current challenges in raising financing for the Jubilee gas projects and uncertainties about the value chain of Ghana’s offshore gas in the medium-term, it is also prudent to ignore the impact of wet (as well as “stranded”) gas in these calculations.

4. The average annual revenue over the five years thus amounts to $1.72 billion in the model adopted above. As argued in the preceding paragraphs, the deviations from this mean figure for actual annual inflows are likely to be much, much, more suppressed because of all the precautions we have taken. 5. Note however that this is gross economic output. We need to find the net national income. Doing so will require addressing the other two principles enunciated by Dr. Amoako Tuffuor.

*Oil Price as Key Determinant*

1. While there is indeed no doubt that the price of oil is an important determinant of net national income, it is by no means the only critical one. Equally as important is the unit cost of production (i.e. how much it costs to bring one barrel of oil from the seabed into the FPSO – Floating, Production, Storage & Offloading vessel). Dr. Tuffuor’s complete silence on this matter during the interview and subsequently was worrying.

2. When one looks at the historical price data for the past two decades, one is actually pleasantly surprised that over stretches of 5 years or less the graph appears quite smooth actually. Major price spikes last a few weeks after which levels then revert to the historical trend. In that sense, we can be reasonably assured that computing an average price for oil over 5 years is far from reckless, and provided the right calculations are performed on the historical data the emergent figure is likely to prove useful. $60 is thus a sound estimate in acknowledgment of the factors identified in the preceding.

3. In similar vein, while it is true that cost of production figures also vary, historical trends are useful in gauging their likely levels over time. The secondary information we have from the program of development of the Jubilee field approved by the Ministry of Energy, on GNPC’s advice, makes no sense.

4. As a matter of urgency, the relevant authorities should publish the full assumptions underlying their production cost model. Until they do, observers like us are entitled to use for our purpose West African typical rates for FPSO costs, labour costs, insurance/short-term finance costs, transport costs, and safety/environmental costs. A triangulation of Nigerian, Ivorian, and Equatorial Guinean cost trends (bolstered with data from the United States Energy Information Administration) should suffice. We would maintain in those circumstances that $34 per barrel is a reasonable estimate of unit production cost. We have seen figures from industry insiders for the relevant geographies that go as high as $50.

5. Then also, there is a need for the accounting template being used for all the income analysis to be published. The accounting template determines what investment types count as an exploration, development or production cost, all of which receive different kinds of treatment when determining what the net proceeds to Ghana shall be.

For instance, the latest indication is that not only shall the builders of the FPSO operate the platform, they shall also be leasing it to the Jubilee consortium. In accounting terms, does this imply a transfer of costs from development to production category, or what? Is Government carried interest to be treated as a dividend, and therefore levied on profit, or a direct share of output, and therefore levied on gross production prior to EBITDA, and if the latter what are the tax deductibility implications, precisely?

6. In the absence, as we have long complained, of any serious information from the relevant authorities, save for media soundbites, we shall proceed to use reasonable, elementary, accounting and analysis, and shall be fully justified for doing so.

7. There is no doubt that government’s 5% royalty is applied directly to gross output. Thus government will be entitled to a little over $86 million from the $1.72 billion mentioned above before the serious math even begins. That leaves $1.64 billion.

There is now a question as to how carried interest is to be treated. Let us assume that we can apply this also directly. That gives us another $164 million to be apportioned to government share. We have $1.47 billion remaining. We are on safe ground if we assert that government from this point on can only receive income derived as a percentage of profit, which means we must going forward accommodate development and production costs.

If we use $34 as the production cost per barrel, then we are confronted with a whopping $980 million of cost per annum. Taxable income then declines to $490 million. Now, we must account for development costs which, under basic accounting rules, should be amortised over 5 years. As total development costs have been pegged at $3.15 billion, the annual sum slated for capital depreciation equals $650 million.

8. It should be obvious to even the most casual reader that in the scenario we have portrayed above the Netherlands-registered Jubilee Field Consortium shall declare a loss of $160 million for year 1 at least, and depending how they work out “retained losses”, not only shall they not pay any taxes for that year but they may actually pile up these losses over the first five years to the tune of $800 million. Ghana shall thus receive its statutory $250 million per year over this period, but shall see this sum barely rise in phase two of development, when production is expected to ramp up to 250,000 barrels per day, due to accumulated losses.

9. What we have sought to demonstrate above is that “cost of production”, as well as the threat of tax evasion due to “internal mispricing” (wherein business costs are exaggerated) are as important a determinant of government earnings as the price of oil.

Now it may well be that in the absence of an official accounting template we have committed the basic error of double counting, wherein some costs already accounted for under one category are assigned to the other, but it is for this reason that government should publish its production costs assumptions and the accounting template it has adopted. Bear in mind, dear reader, that in the scenario described above government of Ghana shall be responsible for 3.75% of the losses declared above, a not insignificant $6 million per annum, and as Government increases its participating interest as it is seeking to do the greater its share of any such losses shall be.

*Range of Estimates*

1. Without putting too fine a point on the matter, there is hardly any doubt that the position taken by the Dr. Tuffuor-led team that we accept the validity of wide ranges of estimates ($244 million as demonstrated above up to $1 billion based on the MOFEP model) is a dangerous one. It amounts to saying that we should not bother to plan at all, since the statistical variation of the figures being produced makes nonsense of any notion of serious financial forecasting.

2. Or, we could endeavour to use the lowest figure in any “rigorously produced” range of estimates for planning purposes as this would clearly be the prudent thing to do. An overflow account can easily accommodate any extra revenue, but certainly not shortfalls. Thus government should states its position on how it treats ranges of estimates. It cannot tell us that its fiscal models treat all figures on the sensitivity analysis graph equally. Which brings us to the point on rigour.

3. The role of variables in forecasting is to develop probabilities. The role of the government’s fiscal manager is turn those probabilities into a model. In a system such as the one we are discussing we are concerned about “prudency”. The simple argument is that it is fine to underestimate but dangerous to overestimate. Thus forecasting for anything other than academic purposes, and in particular forecasting for fiscal purposes, should be based on the rigorously produced “minimum figure”. Furthermore, variables cannot be admitted on the whim of government analysts. Variables ought to be based on rigourous and explicit assumptions, hence our constant requests to the relevant authorities to PUBLISH THE ASSUMPTIONS

*CONCLUSION*

As far as transparency and accountability in the oil sector is concerned, Government of Ghana has begun on a not too good note. There has been a tendency to opt for bureaucratic substitutes to truly open discussion of the issues. So called “public forums” mean little unless government managers are prepared to engage in a full and open conversation about ALL ASPECTS of oil resource and revenue management.

We see a subtler but more dangerous trend with respect to the above concern. By going through the motions of subscribing to such mechanisms as the EITI (Extractive Industries Transparency Initiative) for the oil and gas sector, government would be able to argue that it is abiding by the tenets of good natural resource governance, when the truth is that it is complicating the very process of accountability through bureaucractic posturing.

We note that the Ghana EITI process (or GHEITI) has been in place, on candidacy basis, for the general mining regime since 2007. We are still at least 3 months away from an aggregated, *independently vetted*, publication of revenue payments and other crucial data related to the non-petroleum mining industry in Ghana, as part of the country’s validation report. We note that the EITI process as currently constituted boast of civil society participation, and we have nothing but respect for the individual members of the National Steering Committee of the EITI, we are however not sure on what basis the process is considered representative of the diversity of civil society focuses and interests in this matter. We note further that the GHEITI validation exercise undertaken by Mr. Martin Amidu and others is not in the public domain. Indeed, even the draft processes supposedly guiding Ghana’s extension of EITI to the oil and gas sector have not been adequately publicised. This falls short of the transparency the entire process has been designed to foster. In short, we are not convinced that government of Ghana shall be able to fulfil its obligations with respect to accountability and transparency by simply complying with the mechanics of the GHEITI, especially as presently designed. As far as we are concerned, and we know we speak for a great many individual Ghanaians and civil society organisations, transparency and accountability can only come about through a full and unrestrained discussion with the GENERAL PUBLIC about ALL ASPECTS of the oil sector management process. Until that happens, the course for Ghana shall be into the oil cesspool of shattered hopes and disillusion. *Courtesy of IMANI Center for Policy Education & AfricanLiberty.org*