2. Money in and out: project cash flows
The economics of extractives projects are complex, and have evolved around structures which are a combination of market economics, international best practise, and national law over a long period of time. But they can be broken down into three successive stages: cash flow, or the underlying “raw” economics of how much investment is needed to produce a resource selling for how much; “fiscal”, or what rules a host government uses to levy royalties, taxes and other revenue streams from the cash flows; and project finance, or how fiscalised revenue flows are affected by such features as incorporation structures, debt financing and intra-business group arrangements of finance and transfer pricing.
This section deals with cash flows, which itself breaks down into three categories: production, price, and costs.
Once the quantity of the commodity has been estimated (see the previous section), a company needs to decide its production profile. Even at the highest level of proven reserves, this cannot be arbitrary. That is to say, a company will rarely be able to treat the resource as a liquid asset, capable of any configuration of production. There are two guiding factors which will determine how much of the resource is produced when: geology, and market conditions.
Many oil fields follow a pattern in which the first year or two of production are “ramping up”, a “production plateau” is then reached which could last several years, and then there is often a long, and slow decline in production. And this is a pattern which is mainly determined by geology. When the oil field “comes online”, pressure within the reservoir is at its greatest. It may take a year or two to complete the infrastructure of wells and pipelines which allow the company to take advantage of this pressure, hence the ramp up period. Once it has, the field will move to plateau production. Generally, though, the natural pressure in the reservoir starts to weaken quite quickly – as oil or gas gets produced out, the pressure is released and the remaining oil and gas flows through the reservoir more slowly, resulting in lower production levels.
Classic production profile figure
It is important to emphasise, though, that such a production profile is purely schematic. Each field’s production history and future profile is, in the end, unique. But also, different classes of asset are likely to have different generic profiles. A field producing oil or gas through fracking techniques, for example, is likely to have a shorter lifespan than its conventional counterpart, with a shorter plateau and depletion rates, or the decline in the tail, which is steeper.
The company has options to influence this process, within limits. It can choose to produce all, or some part of the oil rising through natural pressure. It could decide to increase the pressure by various enhanced recovery techniques, such as injecting the field with gas, water or dioxide, but these all require extra capital investment. Such processes of reservoir management have a number of aims: to maximise production of the field over the long term, or to increase production when the market is booming and there is more profit to be made, or to find a sweet spot of maximum production against a given level of capital investment.
There are similar considerations in mining. A resource estimate, as we have seen, will generally include different quantities of reserves at different levels of confidence, and a big factor in determining the degrees of confidence is market price. It is common, therefore, for an operation to plan production in a way which seeks a path to the highest grade ores, but that will also need to be gauged against the specific geology of that mine. A gold mine, for example, will have several different future production scenarios, each mapped against a particular “cut-off grade” of the ore. More ambitious scenarios, involving lower grade ores, will only be brought into play if market prices can support them. But also, just as with oil and gas, a mine may have incremental development options: to progress from an open cast mine to an underground shaft, for example, or open up a new operation nearby, or to process tailings which may contain low levels of resource that have been left until now.
The big picture, then, is that it is important to see levels of production within the life of a project as a dynamic variable, capable of being influenced by many factors on an ongoing basis.
Price and valuation
Much of the material in this sub-section is dealt with in the separate paper on Oil and Mineral Trading. What is offered here is a potted summary.
The single biggest concept to grasp in terms of commodity values is that it is not as simple as price, for two reasons. First, relatively few commodities are sold on a like-for-like basis, like potatoes, say, or wheat, but instead valued indirectly against “benchmark” grades. Second, significant proportions of commodities are never sold on an open market at all. They are traded between business groups in related party transactions, or as part of “term contracts” lasting many years. As an indicator of the scale of this non-market trading, the OECD has estimated that over half all international trade that takes place every day is between related parties, not “arms length”. And the proportion might be on the rise. The US government has estimated that in 2014 some 42% of oil imports were between related parties, compared to just 23% in 2002.
This means that in terms of economics, we are really talking about “valuation”, not just price.
Benchmarks, spot markets and liquidity
Many commodities that are sold in open market are sold against benchmark prices, rather than directly. Of the thousands of grades of crude oil that exist, for example, most are sold against one of less than a dozen benchmark grades, such as Brent, West Texas Intermediate (WTI), Dubai Light, Urals Heavy Blend, and so on. This means they will not be quoted as a direct price themselves, but instead as “Brent plus or minus”, so to speak. One grade could be sold at Brent plus 3 dollars, another one at Brent minus 2 dollars. As long as this differential holds, the price of the crude will go up and down in sync with Brent itself.
Benchmarks evolved in different markets to serve as reference prices against which other commodities can be priced. Each is likely to be specific to the markets on which it is used as a benchmark. So in the case of crude oil, Brent functions as a benchmark in the UK and Europe, WTI in the United States, Dubai Light in the Middle East, and so on. With iron ore, the London Metal Exchange offers one benchmark price, but because Australia and Brazil are major producers, they offer additional benchmark prices.
Differentials exist because of differences in the underlying quality of the commodity, and market related factors, like how close it is physically to the market, and what the specific market demand is at any given moment for that grade of commodity. Crude oil, for example, is not a useable end product. It has to be refined to make petrol, kerosene, tar, petrochemicals and a host of other products. The range of end products producible from the crude oil going in is called “the slate”, and differs widely from one grade of crude to the next. A barrel of light and sweet crude may generate a lot more high value end products like petrol and diesel, than another barrel of crude which is heavy and sour (containing high amounts of sulphur).
And different commodities also vary widely by how much of a spot market exists.
Oil is a “thick” market – there are thousands of buyers and sellers, and benchmark prices are relatively well established. So is gold, where the end product of one Troy ounce of gold is directly comparable around the world. But natural gas, by comparison, has no single global market price because it is so difficult to transport. At any one time, prices for a million British Thermal Units (BTU), the main unit of sale, could vary by as much as 600% between North America, Europe and east Asia. And there are other commodities, such as uranium or mineral sands, where there are only a small number of producers, such that even if quantities are sold directly in port markets, they are relatively few and far between, and mean a spot price cannot be reliably established.
It is commonly assumed that open, competitive markets are the norm for international trade. But not only is this not entirely true today, as we have seen, because of the prevalence of intra-group transactions, spot markets are a relatively recent invention altogether. The oil industry was historically dominated by the “Seven Sisters”, the British, French and American supermajors who were vertically integrated in industry terms. As long as the controlled upstream production, most oil was being transferred from their own upstream arms to refining, petrochemicals and retail distribution arms within the same groups, and relatively little was being traded on an open market. It was only when economic nationalism created much more active state involvement among newly independent producing countries, in the 1960s and 1970s, which included direct ownership of a portion of the oil, that spot markets developed, and trading houses such as Phibro, Glencore and Vitol became actors with significant impact on global markets.
Contractual terms relating to price and valuation
Since the price fetched is such a critical question, extractives contracts often contain previsions about how valuation is to be handled. There are several critical features:
Point of valuation: Transporting a commodity from where it is produced to an end market is often expensive. The location at which a commodity is sold along the value chain therefore influences the price. The further from market, the lower the price.
Commodity basket: A commodity can be value relative to a single benchmark, or a “basket” of several. In such cases, the basket is likely to be weighted, with each component of it accounting for a percentage of the basis of comparison.
Currency and date: Valuation formulae will also specify a given date of sale of any benchmark and currency, to guard against day to day fluctuations in the commodities and foreign exchange markets. Any of the transactions in the value chain which relate to revenue flows that will eventually be taxed in local currencies are likely to specify which rates of conversion were used.
The 21st century BCE has already seen massive volatility in market prices. From 2000 up until 2014, oil and minerals saw a boom in both prices which led some analysts to talk in terms of a supercycle, and others to suggest that commodities were now under such pressure from rising demand, itself driven by population growth and a rise in prosperity, that the higher price levels now achieved represented new equilibria – a long-term step change in market conditions. And then, beginning in 2012-13 with gold but accelerating in 2014 with oil, prices crashed. This is simply the latest round of volatility, which is a constant feature of commodity markets as long as they have existed.
One important thing to understand is that the relationship of price to demand is non-linear – that is to say, prices rise and fall by percentages which are much greater than the rise and fall in underlying demand and supply. The first decade of the 21st century, for example, saw the emergence of China as a “swing consumer” of many commodities – a large economy with such high economic growth that it could drive pricing in world markets like iron ore, aluminium, tungsten, and coal. In the case of oil and gas, the emergence of a new segment of the industry in the United States based on fracking was also considered a key factor in the crash of 2014, and a response driven particularly by Saudi Arabia to try and drive the new shale producers out by flooding the market.
But in both cases shifts of demand and supply were more modest than the price swings. Shale production in the US started in 2005 and focused on gas. In terms of oil, total production added may have been, cumulatively, four to five million barrels a day. Certainly significant for the United States, but since global consumption of oil has been running at about 90 million barrels a day through this period, it represents a rise in supply of perhaps 6%. The other factor often cited was in the slowing down of Asian economic growth, which led to a drop in demand. But this drop, in terms of global demand, might have represented perhaps 1-2%. The cumulative fluctuation of both demand and supply, then, was less than ten percent, and yet prices were halved, from about $100 per barrel at the start of 2014 to below $50 per barrel by the end of 2015.
Financialisation and speculation
One feature of markets in the last couple of decades is their increasing financialisation – the ratio of investment in financial instruments that are tied to underlying commodity prices in some way, but are not sales of physical oil, or gold, or iron ore. Significant use of such instruments first evolved in the1980s, when major consumers of physical commodities such as airlines needing large amounts of fuel, or manufacturers needing significant quantities of, say, steel, began to buy futures and options in commodities, to allow financial planning and protect themselves against market volatility. If an airline, for example, knew that it would need a million gallons of jet fuel in six months time, it could pay for an option to buy it at a fixed price. If the market price stayed below that then they would not exercise the option and it would expire. But if prices soared, they would be protected against those rises. Producers too might want to insure themselves against price drops, in what became known as “hedging”.
By the early 1990s there was more “paper oil” being traded than physical oil. But the relation between the two was altered dramatically in the next 20 years, until by 2010 trade in “paper oil” was an estimated 20 to 30 times higher than the physical commodity itself. Even within physical markets a barrel of oil might be traded ten times between being loaded on a ship in, say, Nigeria, and offloaded at Rotterdam. Markets also created commodity-tracking financial instruments which rose or fell directly in response to the price of benchmarks, or a basket of different commodities, without necessarily being directly invested in an underlying consignment of the commodity.
Such financialisation has led to a debate about the extent to which speculation has fuelled volatility on global markets. Experts are divided. Some say the speculation has played an extreme and negative role, while others maintain that speculation, on balance, is positive, and that markets continue to be governed by “fundamentals”, the supply and demand curves for the commodities themselves.
Costs in extractive industries tend to fall into two categories: capital expenditure (or “capex”) which is required to develop and maintain an industrial project, and operating expenditure (“opex”) to run it day to day. These are significant because they have different treatment in the tax codes of most countries – in other words the same amount of overall investment in two different projects could lead to different rates of profitability, depending on the mix of categories of spending in the total.
The large up front expenditures usually needed to develop a mine or an oilfield, one of the distinguishing characteristics of the industries mentioned in the introduction, are capex – building the mine itself, sinking wells, building access roads, pipelines, on-site processing facilities and so on. These can often run into billions of dollars. The Gorgon liquefied natural gas project in Australia, for example, was completed in 2015 at a cost of over $40 billion. Another project under consideration, the Libra oil field off the Brazilian coast, could involve up to $90 billion of capex.
Mining projects are usually not quite as large. Nevertheless, in 2016 Rio Tinto went ahead with an expansion to the Oyu Tolgoi copper and gold mine in Mongolia scheduled at $6 billion, and the Simandou iron ore mine in Guinea, whose estimated capex could exceed $20 billion, since it involves building a new deepwater port.
Capex includes spending on exploration, initial development, and all major infrastructural spend. It also includes abandonment costs, although special provision is often made for these since a major part of these are incurred after the mine has stopped producing, and cannot therefore be paid for out of current revenues.
The critical thing about capital expenditure is that once it is done it cannot be undone. This sometimes gives rise to what economists have called the “time inconsistency” problem. Companies are keen to sign agreements that last for the lifetime of a project before risking large amounts of capital in developing a project. At that time, with no production or revenues in sight, governments are also keen to make the investment happen.
But once the project goes ahead, and in particular the capex has been sunk, that dynamic can change. The government may see considerable production, and revenues, materialising but without major tax benefits – because the project is in a “cost recovery” stage where most revenues are going towards repaying the large initial capex. This can lead to a government changing its view about what is reasonable, and demanding changes in the contract, knowing that the investor cannot now walk away from the project without losing all its capex.
Although cos recovery happens early in the life of a project, capex is depreciated in accounting terms. That is to say, its impact on cost recovery or taxation is staggered over a few years, sliced into tranches that go into the mix every year, rather than being “expensed”, or immediately deductible.
Operating expenditures include salaries, day to day running costs of the mine or oilfield and its associated infrastructure such as access roads or pipelines, some categories of transportation, and overheads such as management and administration. By definition, operating expenditure must come later in the life cycle of a project since they can only begin once production starts, which in megaprojects is often five or more years after the start of the project. If capex is often the leading cost factor in determining whether a decision will be made to invest, opex is often key in deciding whether a mine or oilfield continues in operation or not under changing market conditions. And once there is production, operating costs are immediately expensable from turnover.
Differences between oil and mining cost structures
Although a typical project has a lifecycle of the kind described, with high and early capex followed by opex, there are some structural differences in these schematic views between the oil and mining industries. While “conventional” oil loads all capex up front to the start of a project, mining has a category of capex called “sustaining capital” which remains a factor throughout the life of a project. Sustaining capital is used for the maintenance of vehicle fleets, and for replacing any part of the wear and tear caused on equipment caused by mining on an industrial scale. Significant sustaining capital costs will changes the cost and risk profile of a whole project, since capex is now needed continuously.
In hydrocarbons, unconventional projects such as fracking or tar sands have a different cost profile too. Tar sands is essentially a form of mining and so entails what are effectively sustaining capital costs. In fracking there is a need for continuous capex, not so much for maintenance purposes as due to the fact that each well produces less oil or than a conventional field and has a much shorter life. A shale project therefore has more wells, which are drilled continuously over most of the life of a project, in contrast to conventional oil where most drilling work is done at the start.
The rise of costs, and the rise of the mega-project
Industrial scale extractives projects have always been relatively large. But they have got bigger and bigger in recent decades. In 2015 Ernst and Young published a review of some 365 “megaprojects” – those requiring over a billion dollars of investment – and concluded that this scale of project had become a new norm in the oil industry as the end of “easy oil” forced companies to look for new resources in more and more challenging places with more and more sophisticated technology: the Arctic, the deep sea, coal bed methane, liquefied natural gas (LNG).
They also found that as projects got bigger, so did the likelihood that they would overrun on costs, and time to production. The oil industry spent about $100 billion a year on development in 2000, but this had risen to over $600 billion a year by 2010. The Ernst and Young report suggested this was likely to rise again to an average of over $900 billion a year in the coming years. To some extent development costs are driven by prices. When there is a boom market, producers have more incentive to develop new resources, and so costs naturally rise in those areas where there is limited supply, such as renting drill rigs and personnel. But the extent to which costs have risen seems to go beyond simple reaction to changes in demand in the market.
The study estimated that these megaprojects were more than 50% likely to entail significant cost overruns, more than 70% likely to incur time overruns, and that the cost overruns would average 57%.
It is the same story in mining. Another Ernst and Young study in the same year showed mining megaprojects likely to average cost overruns of 62%.
The phenomenon is not limited to the extractives sectors. There is a growing body of literature among economists about overruns in sectors such as public infrastructure, utilities, airports and Chinese construction. So far the common factor seems to be that since megaprojects take a long time, there is more chance unaccounted for externalities can impact execution, such as higher than anticipated general inflation rate, for example.
But the fact that the megaproject is accounting for more and more development and production in extractives needs to be seriously factored in by governments. All estimates of revenue flows depend on forward looking estimates, with the company often the sole source. If those estimates, influenced by the rise of the megaproject, are systemically optimistic, then estimates of how much money when need to be revised down.
At the heart of extractives projects is the economic concept of “rent”. Literally hundreds of volumes have been written about economic rents in a literature which goes back to David Ricardo, one of the earliest economists, writing in the early 19th century.
At its simplest, though, rent could perhaps be described as the compound effect of three core principles: first, that producers in a given market have to sell at the same price but can have radically different cost bases, which are “blocky”, or discontinuous, rather than “smooth”, or continuous; second, that a large portion of those costs may have weak correlation to price, meaning that superprofits – or superlosses – could be created by changing market conditions without any significant change in the operation of the project itself; third, that such changes in the market can be non-linear – a relatively small change in demand or supply can drive a large change in price, and a still larger change in profitability.
In the chart on the right, the cost bases of different projects producing a mythical commodity are represented on the horizontal axis, left to right. The amount it costs each project to produce is how high each green block extends up the vertical axis. So project A, on the left, has costs only half of project F, on the right hand side. The line P running horizontally is price, and the blue area in between the top of each block therefore represents profits – which are therefore different for each project against the same market price. This kind of chart as a whole is known as a cost curve.
If this seems abstract, a real world example can clarify. Imagine that the cost curve is for oil, project A on the left is Saudi Arabia, and project F on the right is drilling in the Arctic, or the Canadian tar sands. Even if the details are a closely guarded secret, we know that Saudi production costs remain the lowest in the world – with operating costs of perhaps somewhere between $5 and $10 per barrel. While a barrel of oil from the Canadian tar sands might cost $70 per barrel to produce. This means that at the price point P, Saudi Arabia can still make large profits out of its production whereas tar sand producers must either take losses – because their cost base (project F, in orange) is now higher than the market price – or stop production. If project F were not tar sands but shale production in the USA, the chart would reflect what many analysts have said was Saudi thinking in continuing large scale production (“flooding the market”) while shale production grew, in the hope of forcing prices (the line P) down the chart to a point where the Saudi industry could still survive, while US shale producers could not.
Real cost curves of actual commodity prices look messier than this neat example, but the same principle is in play. Take this cost curve of the iron ore industry.
These real projects show estimated cost curves for major companies such as BHP, Rio Tinto and Vale, in 2014. Again, the projects on the left show low cost projects which are earning higher profits than the projects on the right of the chart, against the benchmark prices represented by the three horizontal lines across the top – the then current spot price, and “netbacks”, or prices set at the export point, in two of the major producers Australia and Brazil.
Interplay between local and global cost curves
A company will always be looking for a project with a low cost base – on the left of the curve. There are even cases in commodities where a single big project could affect the whole of a global supply curve – the amount of a commodity available at a given price – as in the example below from BHP Billiton, where the company states to its investors that its new project, WAIO in Western Australia, will “flatten” the cost curve of iron ore globally. In other words, this one project will bring on enough new production of iron ore, at a low enough cost, that it will bring down prices in world markets. That might be bad news for BHP Billiton’s competitors. But since the company’s own production will have low costs, it will continue to make profits, even with lower prices, and its position will be strengthened vis a vis other producers.
The minimum necessary return on capital
Part of the theory of rent which is not readily intuitive to non-economists is that the cost base for each project includes what is defined as a minimum necessary return on capital. Going back to the theoretical cost curve above, then, this explains why project E would continue to produce. Its cost curve is right up against the price line, so that if it did not include any return on capital, it would be making no profit at all, and there would be no incentive to continue production.
This is intrinsic to the concept of economic rent. The blue area above the cost bases of each project are therefore not just profits, but profits above and beyond the minimum return to capital – in some cases they could be described as superprofits.
While the theory is relatively straightforward, its application is complex. First comes the question of how such a minimum return would be defined. Ricardo first posited this theory in terms of agricultural production in 19th century Scotland, when, it might be reasonable to suppose, land and food prices might be relatively stable. But fast forward to the 21st century, and these global industries, operating in a liberalised trade environment in financial markets that are digitised, and it can readily be seen that capital is so fluid across borders and economic sectors that such a “minimum return” is much harder to define even in theory. In practice, many financial models of extractives apply a simple rate of 8 to 10 percent per year, thereby defining “rent” (the blue area) as lying in profits above this.
The second practical difficulty with applying the theory of rent is that it assumes cost bases are easily knowable. But companies do not routinely publish costs of individual projects. These data are subject very much to the vagaries of compliance reporting and investor confidence seen in the section above on reserves and resources. Some companies publish some of the time. In both of the real cost curves included already, the cost bases are estimated, or imputed, by investment analysts or financial publishers. Such estimates can be informed and skilful, but there is no getting away from the fact that they are in most cases estimates, not facts. In the case of companies, such as BHP Billiton, they at least have firm knowledge of cost bases in their own projects to serve as a basis for estimating the cost bases of their competitors. Governments, however, may not have any actual operations and so rely totally on external databases, which are themselves estimates more often than not.
Who captures the rent?
The theory of rent sets the ground for a specific view of what taxation policy should be in the extractive industries, quite apart from other economic sectors. Since rent is defined as profits above and beyond what is needed for an investor to invest, in theory very high taxation of such rents does not deter investment. Since also, sub-soil resources in the vast majority of the world are publicly owned – that is to say, by the state, or at least managed by the state on behalf of the people – it follows that one objective of fiscal policy around extractives is to capture as much of the rent as possible.
It is important to understand that the specifics of extractives lead to these conclusions apart from any political or economic doctrine about the role of the state in the economy in general. The International Monetary Fund, for example, generally seen as an advocate of free market economics and global trade, consistently advocates that the state should try to capture as much of the economic rent of the oil and gas industries as possible.
The question, given the methodological issues raised above, is how. A separate paper in this series deals with the individual mechanisms of different kinds of royalty, taxes and so on which governments can deploy. There are a wide variety, each with strengths and limitations, acting on their own or as a suite, often called “the fiscal regime”. It is worth noting in passing, though, that some of these mechanisms are explicitly targeted at capturing rent, such as a tax specific to the petroleum sector (in Nigeria, Norway, the United Kingdom and elsewhere), or a Resource Rent Tax (first introduced in Australia in the 1980s). Others are implemented in a way which implicitly targets rents, such as royalties which are staggered according to market price in the Ivory Coast. To relate it back to the cost curve, higher prices means the horizontal line of price goes higher, and the blue area increases. So a higher rate of royalty or tax corresponds to greater rents being earned.