4. Transfer pricing – special extractive industry issues
- Michael Keen, and Victor Thuronyi
- Published Date:
- September 2016
The previous chapter (Shay) presented an overview of transfer pricing. It set out the policy arguments and objectives for valuing controlled transactions on the basis of the arm’s length principle for tax purposes, discussed methods of determining arm’s length prices and the challenges of applying them in practice, particularly to extractive industries, and outlined general strategies available to governments to respond to those challenges.
This chapter as far as possible avoids repeating the general discussion of the previous chapter and instead focuses, in more detail than was possible there, on transfer pricing issues special to extractive industries, particularly in developing countries.
It starts by discussing extractive industry transfer pricing risks and explains in more detail how the special nature of extractive industry tax regimes affects those risks and how, particularly through ring-fencing, it extends them into domestic as well as cross-border transactions. It then considers transfer pricing methods but looks in more detail than the previous chapter at the use of special methods for valuing commodity1 transactions. It then discusses separately oil, gas and minerals, bringing out the similarities and differences in the challenges they present for the use of such methods. In the course of this discussion it looks at transfer pricing issues relevant to royalties, not considered in detail in the previous chapter. Next it looks at transfer pricing of costs, again considering in more detail the use of special rules for extractive industries, for example the no-profit rule for payments to associates common in petroleum agreements, and special rules for limiting finance costs. It concludes with a discussion of practical administrative issues, looking in more detail at the administrative consequences of special valuation rules, for example benchmarking and “physical audit” procedures.
A general theme is the scope for reducing transfer pricing uncertainty and risk by developing rules that are specific, objective and predictable while still broadly consistent with the arm’s length principle.
2 Extractive industry transfer pricing risks
Risks of tax loss from abusive transfer pricing are not peculiar to extractive industries, but a number of special features affect the nature of those risks and of the typical responses to them.
Governments often impose special taxes2 on “upstream”3 extractive industries. They can, of course, tax anything they like, and these special taxes are not necessarily based on profits or on the value of commodities sold.4 Where they are not, transfer pricing is of limited relevance.
The main taxes imposed by governments are, however, usually profit based. These include income tax, but often, particularly with petroleum, this is charged at a special rate and with other special modifications, such as ring-fencing, explained more fully in what follows. Sometimes (again more commonly for petroleum than mining) governments also impose other profit-based taxes such as resource rent tax (RRT), excess profits tax or production sharing. These profit-based taxes are clearly vulnerable to abusive transfer pricing. Particularly in developing countries, governments often take equity participations in extractive industries, and profits from these are vulnerable too.
Governments generally (with limited exceptions, mainly in developed countries) also impose royalties. In a few countries these are profit based, but in developing countries they are normally output based. Royalties can be based simply on the weight or volume of output, in which case transfer prices are irrelevant. Volume-based royalties are rare, however, applying mainly to low-value bulk commodities like sand and gravel. Value-based royalties are a much more common and important fiscal instrument, but there is considerable variation in how output is valued, and the royalty value may not be the same as the sales value. There is no international convention requiring valuation at arm’s length prices as there is for income tax. Where royalties are not based on sales value, transfer pricing may again be of limited relevance. Many governments do, however, as a policy choice, base royalties on sales value, in which case they are equally vulnerable to transfer pricing abuse.
The result of these higher or special taxes is that the total tax take from extractive industry profits can be very high, sometimes in excess of 70% (though some high tax rates may have been reduced by recent falls in commodity prices). This provides exceptional incentives for transfer pricing abuse.5
There are also exceptional opportunities. In developing countries, large-scale extraction operations are carried out mainly by foreign-owned multi-national enterprises (MNEs); most production is exported; operations are financed with foreign capital; and the highest-value goods and services used are imported. The MNEs are often “vertically integrated”, carrying out the full range of operations from exploration through to sale of final products, so that sales to downstream associates are common. Even without vertical integration, sales are often channeled through an associated marketing or trading company. Goods and services are commonly provided to locally based upstream companies by a foreign-based group management and services company. Associates are often based in tax havens, maximizing potential tax savings from non–arm’s length pricing.
Transfer pricing risks are normally associated with cross-border transactions, but, where special extractive industry taxes apply, they also arise in transactions with domestic associates not subject to those taxes. (Structures involving domestic downstream associates are a common feature of complex extractive projects.) The implications of this are discussed more fully later in this section.
It is often said that withholding taxes (WHTs) are the first line of defense against transfer pricing abuse. There is undoubtedly some truth in this, but WHTs generally apply at standard income tax rates and provide limited protection where extractive industry profits are taxed at higher rates.6
Some factors may be considered to reduce extractive industry transfer pricing risks compared with other industries. Extractive industries are physical operations. They are tied to their location – they cannot be transferred to a favorable tax jurisdiction overnight like an Internet trading business. Outputs are standard commodities, not branded products. They can be physically weighed and measured. Variations in their type and quality can likewise be physically defined and measured. Standard commercial measurements are used. The most common commodities have prices quoted on international exchanges such as the London Metal Exchange (LME), or published by price-reporting agencies. These features increase price visibility. A barrel of oil is easier to value than a designer handbag, and although its value on any particular date may be subject to dispute, this will usually be within a narrow range. Costs generally relate to physical goods and operations, not nebulous intellectual property; and, in the case of petroleum, the prevalence of unincorporated joint ventures may limit transfer pricing risks, as discussed in more detail in section 5.
None of this means that the risks are negligible. Comparable transactions can be hard to find; for some commodities there are no quoted prices; even when there are, they may need significant, complex and uncertain adjustments to establish a comparable price. MNEs may furthermore be able to exploit the high day-to-day volatility of prices to shift profits, if they can use hindsight in selecting pricing dates for controlled transactions. As the OECD points out,7 if a taxpayer in an extraction industry sells all its local-country output to associates, small pricing discrepancies in each individual sale can add up to large reductions in the local tax base. Costs paid to associates often relate to services and provision of know-how, where transfer pricing can again be problematic, as discussed in Chapter 3 (Shay, this volume).
MNE representatives often maintain that extractive industry transfer pricing risks are wildly exaggerated. Apart from the risk-reducing factors listed earlier, they point out that transactions with associates are relatively minor in many cases; that such associates are not always tax haven based; that their trading subsidiaries are incentivized to treat associates and non-associates equally; that reputable MNEs make efforts to comply with laws requiring arm’s length pricing; that antagonizing governments by aggressive tax planning is not in their interest given the political vulnerability of their massive up-front investments; and that the scope for abuse is often limited by imposition of special valuation rules (discussed in section 3). Despite this, tax authorities should not be complacent about these risks. Chapter 3 (Shay, this volume) presented general evidence of such abuse, and specific examples in the extractive industries can be found.8 But authorities should understand that abuse is just a risk, not a certainty, and may not materialize in every case in practice.
2.1 Upstream, downstream and ring-fencing
The prevalence of special taxes on extractive industries means that transactions with domestic associates may be vulnerable to transfer pricing abuse, in which case governments need to ensure they are appropriately valued.
Additional complications arise where a company or its associate is involved in domestic refining or processing. Special extractive industry taxes are generally intended to apply only to upstream operations. This may be because downstream operations do not have the same capacity as upstream operations to generate rents (excess profits), which special taxes are intended to capture; or because, unlike upstream operations, processing can often be carried out abroad, and special taxes might encourage this, reducing rather than increasing tax collection.
Governments in developing countries often see domestic processing of the nation’s natural resources as particularly desirable, because they see it (sometimes with scant justification) as a higher value-adding activity than extraction and/or as a stimulus for wider industrial development. Far from imposing special taxes, they often provide special tax incentives for it. This further increases domestic transfer pricing risks. Purchases or sales may be routed through downstream domestic associates to exploit their favorable tax status. It is worth noting that governments sometimes inflict transfer pricing losses on themselves by requiring highly taxed upstream companies to supply production for domestic processing and consumption at below market prices.9
If special taxes are limited to upstream operations, they must be clearly defined and “ring-fenced” from other operations. Production must be valued fairly at the point where it passes from upstream to downstream (vertical ring-fencing). It can sometimes be difficult, particularly for mining, to identify the appropriate point of valuation. There is considerable variation in the approaches governments take to this issue in practice, as will be discussed in more detail in section 4. If upstream and downstream activities are carried on within the same company, there will not be an actual sale at that point, making it necessary to establish and use an internal transfer price for tax purposes.
Where special upstream taxes are profit based (more common for petroleum), downstream costs must be defined and excluded from the calculation (horizontal ring-fencing) to ensure that costs of lower-taxed activities are not deducted from revenues of higher-taxed activities. Where an MNE carries on both upstream and downstream operations, costs have to be fairly allocated between them, in effect requiring use of internal transfer prices.
Some governments impose different levels of profit taxes on different license areas or even different developments within a license area.10 To prevent companies setting losses from one area against profits from another or allocating costs disproportionately to more highly taxed areas, these areas are individually ring-fenced. Where a company operates in several ring-fenced areas, as is common, its costs must then be fairly apportioned among them, again requiring use of internal transfer prices. Area-based ring-fencing thus hugely increases the volume of controlled transactions. Sometimes MNEs are required to create separate subsidiaries for each project, which formalizes the need for separate cost accounting but does not reduce the incidence of controlled transactions within the group.11
The commercial pricing basis used for sales is an important issue for both vertical and horizontal ring-fencing. There are four main commercial pricing bases:
Free on Board (FOB) – property and risk pass to the buyer at the point of loading, and the buyer is liable for any costs of further transport and insurance.
Costs, Insurance, Freight (CIF) – although property passes to the buyer, the seller remains responsible for costs and risks till the cargo is unloaded.
Costs & Freight (CF) – similar to CIF except that the seller is not liable for cargo insurance.
Delivery – property and risk pass on delivery at the final destination point, and the seller is liable for all costs to that point.
Clear and consistent rules are needed on the basis to be used at the point at which sales are valued for the purposes of upstream taxes. Most commonly this is the FOB price. In that case any costs of the seller beyond that point must be disallowed as downstream costs, since they are already reflected in the FOB price. Where sales are made on terms such as CIF or CF, prices are normally adjusted to FOB for tax purposes. If there is no other way of establishing the FOB price this can be done by netting back the seller’s costs beyond the point of loading from the CIF or CF price, but if these are paid to an associate the costs must be charged at arm’s length prices. Standard international freight charges could be required to be used for transport by associated companies – for example, awards under the London Tanker Brokers Panel, widely used in the extractive industries to fix shipping rates.
A controversial issue is whether gains and losses from hedging against movements in commodity prices (for example by use of forward contracts) should be recognized in calculating upstream taxes. There is considerable variation on this among countries in practice. This chapter does not aim to set out all the arguments for and against,12 but it is worth mentioning that hedging contracts sometimes involve associated parties, directly or indirectly, and transfer pricing issues will have to be addressed if such gains and losses are recognized. Sometimes the transactions are so clearly disadvantageous that they can be challenged on simple commonsense grounds, but often the position is less clear-cut and requires some understanding of commercial pricing of hedging instruments.
3 Extractive industry transfer pricing methods
Extractive industry transfer pricing rules vary considerably from country to country. Some allow more transparent and effective administration than others.
3.1 General transfer pricing rules
Income tax legislation normally contains a general rule requiring use of arm’s length transfer prices, sometimes backed up by regulations and/or reference to OECD guidelines.13 Sometimes there are no such rules in royalty legislation – a serious weakness if royalties are based on sales values.
In some countries general transfer pricing rules are deficient:
Sometimes the definition of non–arm’s length transactions is too narrow. Association may be poorly defined.14 Or the definition may not capture transactions with a non-associate that form part of a wider agreement involving an associate.15 Chapter 3 (Shay, this volume) explains the definition of controlled transactions in more detail.
Sometimes the rules do not oblige taxpayers to report transactions with associates at arm’s length prices for tax purposes. In Anglophone developing countries, transfer pricing rules are sometimes still based on the UK’s pre-1998 legislation, which merely permitted the tax authority to substitute arm’s length prices if transactions between associates were priced in a way that reduced tax. This leaves taxpayers free to misprice such transactions with impunity and puts the onus on the tax authority to detect such mispricing and determine the arm’s length prices to be substituted. This is an inadequate foundation for ensuring appropriate transfer pricing and is incompatible with self-assessment principles.16
Transfer pricing rules may not apply to domestic transactions (a problem if extractive industries are taxed differently from other domestic businesses).
It is essential to eliminate these weaknesses if tax authorities are to have any hope of using general transfer pricing rules effectively. Merely strengthening administrative capacity and training will not suffice. The rules should be amended, or supplemented by regulations, to define non–arm’s length transactions comprehensively and to require taxpayers to price them on arm’s length terms for tax reporting purposes and maintain records showing how they established arm’s length prices. The onus should be clearly on taxpayers to demonstrate that prices equate to arm’s length terms. Penalties should apply for non-compliance with these procedural requirements, in addition to penalties for understatements in returns. (Penalties are discussed more fully in Chapter 3; Shay, this volume.) In many developing countries extractive industries benefit from tax stabilization agreements, but in most cases these would not prevent governments from changing procedural rules in this way.
Some transfer pricing methods can only be verified with data from the foreign associate. Tax authorities in developing countries are sometimes fobbed off by claims that this data is unobtainable or confidential. It is often recommended that they enter exchange-of-information or mutual-assistance agreements with other countries. These are fine in principle but can be a slow, cumbersome and ineffective way of resolving audit enquiries. A more effective approach may be to make deductibility of payments to associates conditional on tax authority access to the associate’s accounts and records, where reasonably required to verify the pricing basis used. This is particularly important where tax havens are involved.
Even if general transfer pricing rules are strengthened, the question remains how exactly extractive industry transfer prices should be calculated in practice. Chapter 3 (Shay, this volume) explained five standard pricing methods recommended by the OECD and UN, and those explanations are not repeated here. For extractive industry costs, any of the standard methods may be appropriate. Comparable uncontrolled price (CUP) may be difficult to establish, since costs may not involve widely traded goods and services, but alternative methods can have varying and unpredictable results. For sales, CUP is arguably the most suitable method, with resale minus a possible alternative. (For commodities, “netback” pricing is the term generally used to describe the resale minus method, though if the costs netted back include a profit element, it becomes more like a profit-split method.) The case for CUP is based not just on the standard physical properties of commodities but on the fact that other methods might produce values with no relation to normal market prices. For example, since commodity market prices bear no fixed relation to production costs, the cost plus method would be unlikely to produce normal market values. Likewise profit-split methods would be unlikely to produce normal market values, since there is no consistent relationship between the profitability of upstream and downstream operations. Factors producing high rents in the former do not apply to the latter (indeed, overcapacity has often resulted in refining industries struggling to generate any profit at all). To price transactions between upstream and downstream operations on the strength of a comparison of their functions, assets and risks might, therefore, particularly at times of high commodity prices, produce values lower than normal market prices. For these reasons commodity-producing countries normally prefer the CUP method wherever a CUP is available. If differences of timing, quality, location and contract terms make it difficult to find a true CUP, they normally prefer to adjust the price of a known transaction to reflect those differences, modifying the normal CUP method, rather than fall back on alternative pricing methods that are complex and uncertain and may not reflect the economic dynamics of commodities markets. But companies may not follow this approach, and even if they do it may be unclear what type and level of adjustments are appropriate.
General transfer pricing rules may therefore leave companies with considerable latitude to manipulate transfer prices. The risk identified by the OECD that systematic minor discrepancies could have a major impact on the local tax base is clearly present.
Most developed countries struggle to apply general transfer pricing rules effectively. Highly trained specialists are normally employed, but even then transfer pricing audits can require long and complex investigations and negotiations, often with conclusions disputed, outcomes uncertain and limited evidence of any wider impact in stemming profit shifting. Tax authorities in developing countries may have neither the resources nor the capacity to undertake these. They are often recommended to resolve uncertainties by negotiating specific transfer pricing methods with individual taxpayers under advance pricing agreements (APAs) but may lack both information and negotiating skills for this.17
Developing countries may in any case prefer, where possible, to base transfer pricing on standard, published, objective methods, for reasons of transparency as well as simplicity. The importance of transparency of extractive industry taxation is widely emphasized, particularly for countries where general standards of governance are poor. Tax law that is clear and predictable and not a matter for departmental discretion or negotiation is an important element of transparency. Individually negotiated agreements on transfer prices are arguably inconsistent with this.
.2 Special rules for commodity sales
General transfer pricing rules are therefore often supplemented or overridden by special rules, in legislation or contractual agreements, setting out exactly how extractive industry transfer prices are to be calculated. This is particularly common (not just in developing countries) for commodity sales. An example of a special rule might be that non–arm’s length sales of a commodity must be valued for tax purposes on the basis of a published international benchmark price. Benchmark prices (discussed in more detail in section 4) can produce realistic arm’s length values if the benchmark price relates broadly to the same commodity and appropriate adjustments are made for any differences affecting relative values. The government may specify the benchmark to be used and the adjustments to be made to it and may supplement this with further rules prescribing the pricing date or period to be used in applying it. Latin American countries often call this approach the “sixth method”, though there are variations in its detailed application. Under the BEPS project, the OECD has issued a discussion draft on commodity transfer pricing18 proposing additional guidelines, first to clarify that CUP would generally be the appropriate method for pricing commodity transactions between related parties and that quoted prices can be used under this method – but only if appropriately adjusted; second to allow imposition of a deemed pricing date – but only if there is no reliable evidence of the actual pricing date agreed by the parties; and third, in due course, to provide additional guidance on comparability adjustments to quoted prices. In practice the qualifications included in these proposals are not always observed by countries using the “sixth method”.19
Other types of special rule also depart to some extent from the normal arm’s length principle. A fairly common rule is for commodity sales to associates to be valued at the average (or weighted average) arm’s length sales price for all sales of that commodity over the period – say the month or quarter – in which the sale occurred. Since there may not be enough arm’s length sales in a period to provide reliable data, there is often provision for average benchmark prices to be used as well or instead. An advantage of the use of average prices is that it increases the pool of arm’s length sales from which prices can be established, allowing use of a modified version of the CUP method. Use of average prices also simplifies administration – once the average has been established it can be published and applied uniformly to all non–arm’s length sales in the period. It makes it harder for companies to benefit from manipulating pricing dates with the benefit of hindsight. It is not strictly in accordance with the arm’s length standard, since commodity prices are highly volatile (with particularly wild swings in recent years), so that the average price may differ significantly from the market price on the contractual pricing date. This could lead to double (or non-) taxation, since it is unlikely that the associate’s purchase will be valued on the basis of average prices in its home tax jurisdiction. MNEs might therefore in theory be able to seek adjustments under mutual agreement procedures, but often developing countries will not have the necessary double taxation treaties, and MNEs may in any case be prepared to accept an element of rough justice if they assume that these differences will cancel out in the long run.
A risk for governments is that companies may be able to exploit the use of average prices. For example, at the end of the period they may choose whether to sell to an associate or a non-associate depending on whether prices at that stage are higher or lower than the likely average (see Figure 4.1). They may be able to use hindsight to decide which deliveries to treat as sales made to non-associates. The longer the period over which prices are averaged, the greater are the risks of this kind of manipulation. It is necessary to examine the rules in detail to establish how far it is possible (they are often more complex than the description suggests). But often it is, and if it is possible in theory, it will almost certainly happen in practice. In 2006 the UK revised its petroleum valuation rules to curb substantial tax losses resulting from this kind of manipulation.20 There is often little awareness of this risk.
Figure 4.1Exploiting the use of average prices
Although general transfer pricing principles should be set out in primary legislation, it may be sensible to include special transfer pricing rules in secondary regulations to allow flexibility, for example to respond to the emergence of more appropriate benchmarks.
An alternative and more radical approach sometimes adopted is to apply a general valuation rule to all sales of a particular commodity, whether or not an associated party is involved. An example might be that all sales of crude oil in a period must be valued for tax purposes at the average of the arm’s length sales prices achieved in the period for sales of that crude rather than on the basis of the actual sale price achieved in any particular transaction. (Again possibly benchmark prices or a combination of actual prices and benchmark prices might be used instead.) Another example might be that (say for royalty purposes) all sales of a mineral ore must be valued at the closing LME price for its mineral content on the date of export. Since all sales are valued on the same basis, there is no need for transfer pricing rules (though the main rationale may be to prevent transfer pricing abuse). This approach is sometimes described as reference or norm pricing or, where benchmark prices are used, as benchmark pricing (though there is a risk of confusion since benchmark prices are also a common feature of special transfer pricing rules of the kind described earlier).
The intention of a general valuation rule is normally to produce a reasonable approximation of arm’s length sales values where it applies for income tax purposes; but this may not be the intention with royalties, since these are often not based on sales values (as in the second example, where the value of an ore’s mineral content would exceed the sales value of the unrefined ore).
General valuation rules may reduce the importance of transfer pricing rules, but they usually remain of some importance, because a general valuation rule may not be practicable for all commodity sales and will usually be impracticable for valuing costs.
Valuing all sales on the basis of a common rule has advantages and disadvantages compared with a special transfer pricing rule.
It further simplifies administration, since sales can be valued by simply measuring physical production and applying the stipulated price. (As with specific transfer pricing rules, average prices are often used, so that all production in the period concerned can be valued by this simple calculation.) In theory this makes it possible to verify taxable revenues entirely on the basis of “physical audit” (though this may not be advisable in practice). Physical audit is discussed more fully in section 6.
If average prices are used the “rough justice” problem discussed earlier is worse with a general valuation rule, because the average applies to all sales, not just non–arm’s length sales. This could again raise double taxation issues, as discussed earlier. The problem is reduced, though not eliminated, if the period over which any average is calculated is kept short.
The scope for companies to exploit average prices by choosing whether to sell to associates or non-associates is eliminated. Where general valuation rules are not based on average prices, there may be scope for avoidance if MNEs can set pricing dates with the benefit of hindsight.
Where a general valuation rule requires use of benchmark prices, a significant advantage is that it removes the need for tax authorities to identify non–arm’s length transactions. These can be hard to identify. Tax authorities may be able to spot suspicious transactions, but while it is one thing to spot them, it is quite another to prove they were non–arm’s length in order to impose a transfer pricing adjustment.21 (Where a general valuation rule requires use of the average price of arm’s length sales in a period, it still remains necessary to identify and exclude non–arm’s length transactions, since their inclusion might depress the average.)
General valuation rules based on benchmark prices may also be perceived as having advantages for transparency. These prices can be published and verified, and the public may have more confidence in the government’s ability to ensure they are applied than in its ability to identify non–arm’s length transactions and ensure they are valued appropriately.
There are many variations both between and within countries in how these different types of rules are used. Special transfer pricing rules may provide greater simplicity, clarity and predictability than general transfer pricing rules, but it is hard to design them to value non–arm’s length transactions realistically (with some types of commodity presenting greater difficulties than others). They are generally unpopular with investors because of the administrative burden of substituting government-imposed transfer prices and the risk that the imposed prices will overvalue sales in practice, whether because of averaging or because they use an inappropriate benchmark or pricing date or because they are applied asymmetrically. (The “heads I win, tails you lose” application of special valuation rules is not uncommon – for example, Argentina and Nigeria apply them only where they produce a higher value than the recorded price.) A general valuation rule may have additional advantages for governments but will be even more unpopular with investors, since it applies to all sales and not just controlled transactions.
4 Commodity valuation
Transfer pricing of different extractive industry products involves common broad issues, but there are significant differences of detail among oil, gas and minerals.
Crude oil is normally valued, for the purpose of both royalty and profit-based taxes, at the point of sale or delivery, that is, the point at which ownership passes to a buyer in a sale, and at which production is measured for sale purposes. The point of valuation tends to be less of an issue for oil than it is for gas and minerals. Upstream and downstream operations are fairly distinct. Extraction and refining are usually carried out by separate companies. Operations between the point of extraction and point of delivery normally consist of limited initial treatment (removal of water, salt and other impurities), transport and storage, and marketing. There is therefore usually no major difference in value at those two points. The costs, for example pipeline fees and demurrage charges, are normally limited and relatively easy to quantify. In practice they are normally deductible for the purpose of extractive industry profit taxes, subject to transfer pricing issues if paid to an associate. They may or may not be deductible for royalty purposes but in most cases are unlikely to be a major source of dispute either way, so long as the basic rules are made clear. For developing countries exporting most of their production, the point of delivery is most commonly a terminal where oil is loaded onto a tanker. For offshore production the terminal is often at an offshore platform where the oil is brought to the surface, sometimes from a number of different undersea wells. Some production might be delivered to a domestic refinery, in which case the delivery point would be a terminal at the refinery.
There may, however, be exceptional cases in which costs between the point of extraction (the wellhead) and the point of delivery are significant, and in those cases the alternative approach of valuing oil at a point nearer the wellhead may be adopted. One example is where production in a land-locked country is transported by pipeline or less commonly by road or rail transport to a tanker-loading point at a port in another country. The transport costs to the tanker-loading point could be substantial. If the point of valuation is the point of loading into the pipeline, then for purposes of profit taxes the value (or internal transfer price) at that point may be calculated by netting back the pipeline fees from the sale price at the tanker-loading point (assuming this is an arm’s length price). If the tanker-loading point is the point of valuation, the sales price at that point would be used, but the pipeline or other transport costs would normally be deductible in calculating profit taxes, effectively producing the same result. (Whether pipeline and transport costs were deductible for the purpose of royalties would depend on whether the government intended them to be based on arm’s length sales values.) If deductible pipeline fees are paid to an associate, establishing an arm’s length price for them can be a significant and difficult issue, particularly if the pipeline owner has a monopoly, but in these circumstances fees are often government regulated in practice.
Another exceptional case in which substantial costs can arise between wellhead and point of delivery is found in Nigeria, where costs of onshore pipeline transportation are greatly increased by “bunkering” (illegal extraction of oil from pipelines). Oil in Nigeria is valued for tax, royalty and production sharing purposes on the basis of reference pricing, and a controversial issue has been whether, as the authorities have argued, this valuation should be applied to production measured at the wellhead rather than at the point of delivery (where the volume might be substantially lower).
Oil is generally valued for tax and royalty purposes either on the basis of sales prices subject to a specific transfer pricing rule or on the basis of general norm or reference pricing. The latter is less common but is adopted by some major producers such as Norway, Angola and Nigeria. For oil, arm’s length sales from the same reservoir can reasonably be regarded as providing a CUP for the purpose of valuing non–arm’s length transactions, since the quality is generally fairly consistent, at least over the short to medium term. Oil valuation rules are generally based on this assumption. They commonly allow for use of monthly or quarterly average prices to increase the pool of CUP.
Since there may be insufficient arm’s length sales from the reservoir to provide a reasonable range of CUP in the period, some countries require use of benchmark prices. These may be used on their own or in conjunction with actual arm’s length sales of the crude to be valued, and combinations of different benchmarks may be used. There are differences in the quality of different crudes – light or intermediate crudes contain a higher proportion of the lighter fractions, such as gasoline, most in demand, and sweet (low-sulfur) crudes can be refined more cheaply than sour crudes – but crude from one reservoir is often physically comparable in quality with crudes from other locations and priced similarly in commercial transactions. FOB prices for a range of widely traded crudes22 are quoted by price-reporting agencies such as Platts, Argos or ICIS, providing the data required for benchmark or reference pricing. Generally, once a similar-quality benchmark crude has been identified, standard formulae can be applied to adjust its price to reflect measured differences in its physical quality. It may also be necessary to make an adjustment for transport cost differences, in which case, as discussed earlier, standard international freight charges could be used.23 Benchmark pricing is consistent with commercial practice, since sales between independent parties are often priced on the basis of a benchmark crude with a premium or discount. If arm’s length sales of local crude are consistently priced in this way, it may become possible just to use the standard market premium or discount rather than calculate comparability adjustments independently.24
Where special valuation rules are not based on average prices, they usually impose a specific pricing date, normally at or around the date of delivery, broadly consistent with normal commercial practice for oil sales (other than forward contracts).
Many developing countries impose royalties and production sharing, with the oil valuation rules for those taxes set out in petroleum law or the production sharing agreement (PSA) itself. These are usually special transfer pricing rules,25 whereas income tax legislation often contains only a general transfer pricing rule. Such special rules, for example the requirement to use a particular benchmark, could not be inferred from a general transfer pricing rule and might even be inconsistent with it if there is a requirement to use average prices that differ from arm’s length prices at the contractual pricing date.
There are clear administrative advantages in valuing oil on the same basis for income tax and production sharing (and also royalties if the policy intention is to base these on sales values).26 It would be unfortunate, for example, if disputes about oil transfer prices had to be resolved under different rules for different taxes for no good reason. PSA valuation rules might be applied for income tax purposes as a matter of practice, but as a matter of law PSAs cannot normally override general tax legislation. Special legislation is therefore normally required if income tax rules are to be aligned with PSA/Petroleum Act rules.
Gas, like oil, is generally subject to a special tax regime, usually either the same as that for oil or a modified version of it. In developing countries some gas may be sold into the domestic market, but usually most is exported. In a few cases it is exported by pipeline, but more often it is exported by tanker as liquefied natural gas (LNG). In that case, unlike with oil, significant local processing costs are necessary, since LNG plants are expensive to develop and need to recover their costs. There are broadly two approaches to the taxation of LNG exports:
The aggregated approach. Here upstream gas production and downstream (or midstream) operations including LNG processing are taxed as a single project. With this approach, there is a single point of valuation for the purpose of any special taxes on gas extraction, namely the point of sale from the LNG plant. Transfer pricing is relevant only to non–arm’s length sales made at that point.
The (more common) segmented approach. Here upstream production and LNG processing are taxed separately, with a different tax regime for each. Usually the LNG plant is just subject to normal business taxes, though sometimes it enjoys a preferential regime. For the purpose of any special taxes on upstream production, gas has to be valued at the point it passes to the LNG plant. If there is common ownership of the LNG plant and gas fields that supply it, or if the owners of each are associated, a transfer price needs to be established at that point. For various reasons, however, governments in developing countries normally regulate the pricing of gas sales to LNG plants. The regulated price might be designed to give the LNG plant a fixed markup on costs or a prescribed after-tax rate of return, and all sales to the LNG plant are calculated on that basis whether they involved associated parties or not. In such cases the government should agree with LNG plant operators what records they must keep and supply on request to show that prices have been charged on the prescribed basis in practice. Auditing these records will require some technical expertise, but the transfer pricing method used should not itself be a source of uncertainty or dispute.27
It is more difficult to design specific transfer pricing rules or apply norm pricing for gas than for oil. There are some variations in the quality of gas at the point of extraction, depending on the extent of liquids and impurities, but once processed it is a fairly consistent product. There is nevertheless no standard international spot price that can be used as a proxy for arm’s length sales values. Spot prices are quoted in regions like the U.S. and Europe with efficient gas distribution networks and extensive and highly developed infrastructure for domestic gas consumption, but these differ from one region to another. They are of limited relevance to domestic markets in developing countries, whose circumstances are quite different. (Domestic gas prices are, however, often subject to government regulation in developing countries, so that – just as in the case of sales to LNG plants – transfer pricing is not a factor, all sales being priced on the same regulated basis whether or not made to an associate.)
International quoted prices are potentially of more relevance to exports, but, while the huge growth in LNG international trade has led to more standardized spot pricing, for the present there remains considerable regional variation in gas prices based on local supply and demand. It is important that any benchmarks chosen should reflect this. Gas is, furthermore, frequently sold under long-term contracts rather than at spot prices. Prices payable under these contracts are often based on the price at the date of supply of some non-gas comparator in the buyer’s location, for example oil or alternative fuels. Sometimes obligations to take a certain amount of supply (“take or pay”) are built into the contract. Transport costs may be netted back. Where a gas producer sells gas to an associate, for example a related marketing company, it will often be under a similar long-term contract. There will not usually be a range of arm’s length spot price gas sales from the same gas field in the month or quarter that can be used to establish a transfer price; spot prices quoted elsewhere may not be relevant; and in any case use of spot prices is inconsistent with how sales are actually priced in normal arm’s length transactions. Although future prices payable under the contract may be based on benchmark prices, all the contract terms and other factors such as the buyer’s location are relevant in determining whether it equates to arm’s length terms. Tax authorities will need to consider whether the comparator used, the length of the contract, any break clauses, take-or-pay obligations, transport cost adjustments, and so on, reflect normal arm’s length terms for sales into the same market, and, if not, ensure that any non-standard variations are reflected in the price. This can be a challenging task, especially when gas is first sold. There may well be no local contemporary long-term gas contracts on arm’s length terms available to provide a basis for comparison. The fact that international gas markets are evolving rapidly with the growth of LNG and unconventional gas adds to the difficulty. Governments often obtain assistance from external industry experts on valuation of natural resources, and it may be particularly useful for gas.
Since long-term contracts determine prices for years to come, there is a good case for governments to require terms to be approved or agreed in advance if an associate is involved. They should also carry out checks later to ensure that those terms are applied in practice (or modified only with agreement).28
4.3.1 Income tax
In many developing countries mining is taxed at the normal income tax rate. Gross revenues are normally based on sales values, usually FOB the export port. In some cases, the vast majority of mining sales are exports to non-associated smelting/refining companies, and transfer pricing issues do not arise.
Even where production is sold to an associate, it is often not because the MNE does its own smelting, but because sales to independent smelters or end users are negotiated by a group marketing company, which collects the proceeds and takes a marketing/administration fee. This company may be based in a tax haven (for example, Singapore for sales to the Asian market). In these cases the resale minus method may be appropriate, with the transfer price based on the sale price achieved by the marketing company. To ensure this method is properly applied the tax authority must be able to verify the resale price from the marketing company’s sales records. An arm’s length price must then be established for any deductible marketing fee. In the absence of a clear CUP this may involve an analysis of the functions, assets and risks of the marketing associate. This, however, is a much-disputed topic, and in practice developing countries may prefer the simpler approach of setting the fee at a maximum percentage of sales by law (or even, at the extreme, disallowing it as a downstream cost altogether,29 though MNEs would argue that marketing companies make a real contribution to the value chain, which should be recognized under arm’s length principles).
Where production is sold to an associate that is not merely a marketing company, pricing of sales is more difficult. Can special methods or reference pricing reduce this difficulty? Sometimes metals are exported in near-finished form. For example, on-site mining processes may produce copper cathode or gold dore. Quoted prices (for example LME) could be used for the former. The London gold fixing price could be (and often is) used for the latter. This overstates the value of production, but the difference between dore and refined gold values is marginal (and possibly a minor standard adjustment could be applied to allow for it).
Metals are, however, usually exported from developing countries as ore or concentrate, with a value significantly less than refined metal. Unrefined mineral production does not have the same consistency of quality as oil production, so it is more difficult to establish CUP from arm’s length sales, even if average prices are used. Where benchmark prices are available only for the refined mineral, much more substantial and varied comparability adjustments are needed to arrive at a price for unrefined product than with crude oil benchmarks. There is some transparency of pricing of more common concentrates such as iron and copper, but it is more difficult to use standard formulae to adjust values for quality differences, as can be done for crude oil. The need for complex and varied adjustments makes it more difficult to design standard, simplified methods that produce a reasonable approximation of arm’s length prices.
Norm pricing is generally not a feature of mining income tax, but some countries have attempted to develop special transfer pricing rules based on benchmark prices, reflecting the fact that benchmark prices do play a role in the normal commercial pricing of many common minerals. Usually the terms of an arm’s length sale to a smelting/refining company are that it will sell the finished mineral on the mining company’s behalf and pay the mining company the price quoted on a recognized exchange such as the LME, less a deduction for treatment and refining costs (TC/RCs) – with credits for valuable mineral by-products and penalties for impurities above a permitted level.
This price – the “net smelter return” (NSR) – is often the basis of valuation used for income tax. It is necessary to ensure that the pricing basis used (FOB, CIF or CF) for the NSR is clear and consistent with rules for deduction of transport costs, as discussed earlier. In controlled transactions it is also necessary to ensure that charges for TC/RCs and other adjustments used in calculating NSR are accurate and reasonable. Full access to the associate’s records may be needed for this purpose. There is, however, reasonable pricing transparency on TC/RCs. Market rates for TC/RCs in particular countries are often determined for up to a year in advance, and some commercial organizations publish data on pricing of TC/RCs for certain minerals. It might be possible in those cases to impose a special transfer pricing rule based on benchmark TC/RC rates and include it in regulations (or, if that is not possible, use it as the basis for individual APAs). Some countries instead allow simple percentage deductions for processing costs. Strictly this is inconsistent with the arm’s length principle, but where such costs are relatively minor and the percentage used broadly in line with past experience, it may be acceptable in practice and have advantages of simplicity and predictability – though there will inevitably be cases in which either companies or governments lose out from such standardized methods.
For rarer minerals, markets may not be deep enough to allow quoted prices to be developed and used in calculating NSR. The onus must be on companies to demonstrate that sales to associates are priced on an arm’s length basis, but it will not be possible to prescribe use of a particular benchmark.
The smelter may pay the miner a provisional amount based on quoted prices at the date of sale or shipment, but it will be noted in the earlier discussion that the final NSR payment reflects the quoted price of the finished mineral when it is sold. If transfer pricing rules required the use of the quoted price at the date of shipment this would differ from the commercial pricing date. There is a very wide range of pricing date conventions for commercial commodity transactions (probably wider for minerals than oil). Many respondents to the aforementioned OECD discussion draft have made this point and argued strongly that imposition of a standard pricing date for tax purposes should be exceptional and limited to cases where reliable documentation of the contractual pricing date is unavailable (though others have pointed out that unscrupulous companies could in that case select with hindsight the transactions in which it benefited them to have reliable documentation). The variety of commercial pricing dates clearly makes it more difficult for a special transfer pricing rule to specify a particular pricing date that will produce results consistent with normal arm’s length prices. One option would be to prescribe the most common commercial pricing date in use for the commodity concerned, but – at the cost of some complexity – allow some flexibility to use a different date if approved in advance by the tax authority.
For gemstones, although it is possible to identify characteristics that determine value, there is no standard benchmark, and larger stones may require individual expert valuation. For non–arm’s length sales there may be no alternative but to require them to be submitted for physical inspection and valuation by the tax authority before sale, with arrangements for arbitration in case of dispute.30
Coal, like oil and gas, is a hydrocarbon. It is also similar in the sense that it generally undergoes less complicated processing than minerals. This may allow the development of similar valuation methods to those described for oil.
4.3.2 Special mining taxes
In many developing countries the main special tax on mining is value-based royalty, though there is an increasing interest in imposing special profit-based taxes such as RRT.31
One option is to use the same sales value for these taxes as for income tax.32 This allows for simplified and coordinated administration. The same transfer pricing issues will arise but may be manageable if controlled transactions are limited or satisfactory transfer pricing methods can be established.
A practical difficulty with using NSR for royalties is that they often have to be calculated and paid around the date of sale, perhaps monthly, when the final payment is not yet known. Either they have to be based on the initial payment, or they have to be paid provisionally and adjusted later. This complicates administration, but the solution may be to design better royalty assessment procedures. Most countries successfully administer income tax as an annual tax paid by in-year installments rather than requiring it to be calculated monthly, and there is no obvious reason royalties cannot be administered under a similar regime.
Mining royalties are, however, often not based on sales value. There is a huge variety in the methods countries use for valuing mining output for royalty purposes.33 Royalty policy and administration are often the responsibility of the mining department, with no attempt made to coordinate them with income tax.
Basing royalties on sales value may be considered to have disadvantages. One is that if mining companies process minerals locally, this will be reflected in the sales price: imposing special taxes on that value might discourage local processing, when governments wish to encourage it. This may not be a problem for RRT, since processing should not produce any “rent” or excess profit to which RRT would apply (if it did it would probably be profit shifted from mining).34 It might not in practice be a problem for royalties either, since MNEs might avoid royalties on value added by processing by moving it to non-mining domestic associates.
Some governments prefer, however, to distinguish mining from processing in applying special taxes. But the distinction between upstream extraction and downstream processing is less clear-cut for mining than petroleum. Minerals are found in low concentrations, and the initial extraction is mainly waste earth and rock, which is then subject to a series of further extractive processes. Some of these (e.g., crushing, grinding, leaching) may be carried out at the mine site, others (e.g., smelting and refining) at a separate site (sometimes operated by the mining company). Extraction and processing are thus conceptually similar, and it may therefore be difficult to distinguish them clearly in practice.
A common practical solution is simply to value production at the mine mouth or mine gate. The value at this point excludes any value added by off-site processing, and physical measurement is relatively easy.35 The main disadvantage is that sales are generally not made at this point, so that all output must be valued at an internal transfer price at a point at which there are no comparable uncontrolled sales and no relevant benchmark prices.
Mine gate market value can in theory be established by taking the value at a later point of sale and netting back costs incurred between the mine gate and that point.36 But this may not be straightforward. Costs of processing beyond the mine gate may vary significantly and involve several different stages. Ore may contain different minerals with different values and different processing costs. Minerals from different sources may be blended during processing. Varying amounts of transportation may be required, again possibly in several stages. Processing and transport may be carried out by associates, whose charges may not be at arm’s length prices. The sale from which costs are netted back may be a sale to an associate, for which a transfer price may in turn have to be established by netting back from a later sale to an independent buyer, with similar difficulties. The farther downstream the point of sale to an independent buyer, the greater the potential complexity of the netback calculation. The netback costs may not be known at the time output is measured at the mine gate, causing difficulty where royalties are payable monthly. Royalties are supposed to be simpler than profit-based taxes, but netback calculation is complex. Ginger Rogers once pointed out that she did all the same dance steps as Fred Astaire, only backwards and in high heels – a netback calculation is rather like doing a profit calculation backwards and in high heels.
Royalties do not, however, have to be based on actual value. Countries can simplify valuation by using benchmark (e.g., LME) prices for finished minerals and restricting netback costs (perhaps to those that are easy to measure). A more extreme option is to disregard all costs and simply value output at the benchmark price for its mineral content.37 This is not entirely straightforward, since it requires expert sampling and assay procedures. But it avoids all the difficulties of quantifying netback costs and identifying the appropriate valuation point (the valuation is the same wherever output is measured and can be done wherever measurement is most convenient). It creates no disincentive for domestic processing; indeed, there is an incentive, since royalties will be a smaller proportion of the value of processed than of unprocessed minerals.
Values calculated by these simplified methods will be higher than the actual value of mining output. Royalty rates can be lowered to compensate for this, but there will be no consistent relation between royalty value and actual value. (Where mineral content is used, 2 tons of 35% concentrate would be valued the same as one ton of 70% concentrate but are worth less because of the higher transport and refining costs required.) Various adjustments are possible to bring the values more closely into line (for example, graduating royalties according to the degree of processing) but often introduce more complication than accuracy.
Does this overvaluation matter? Mining companies see any royalty not related to profit as somewhat arbitrary – does this not just make it arbitrary in a slightly different way? Royalties are often a relatively minor tax, so the fiscal impact may be minor and may be balanced by the advantage of a valuation method that is simple and certain.
Countries are, however, increasingly looking to capture a greater share of mining rents, often by introducing graduated royalty rates. Companies do tend to regard royalties as particularly arbitrary and unfair where they bear no consistent relation to sales value, and this could have negative effects on investment and voluntary compliance if royalties assume greater importance.
Governments are furthermore showing increasing interest in the alternative approach of imposing special profit-based taxes on mining. There is a stronger presumption in favor of using arm’s length values for taxes such as RRT. Their main rationale is to target economic rent more accurately than royalties, an aim that might be frustrated if profits were based on unrealistic notional values.
Royalty valuation methods based on notional values might make royalty administration simpler, considered on its own, but different valuation methods for royalty and profit-based taxes make the administrative regime more complicated and less coherent overall.
Nevertheless, special taxes do potentially increase the incentive for transfer pricing abuse, as discussed in section 2, and governments may prefer simple valuation methods that eliminate or reduce that risk in relation to royalties. Decisions on how far to impose royalties and what valuation method to use may also reflect wider policy considerations than transfer pricing – valuation based on finished mineral prices may simply have greater political appeal.
5 Extractive industry costs
Extractive industry costs are so variable that a general valuation rule substituting standardized for actual costs is unlikely to be practicable or consistent with the arm’s length principle. Developing countries often apply cost recovery limits to profit taxes (limiting costs to a maximum percentage of sales), but generally these merely affect the timing of deductions. Nigeria (which has high costs for many reasons) has considered the idea of benchmarking costs administratively for tax purposes, but there is little clarity about how this would be done in practice. Countries should, of course, seek to compare their costs against costs in other countries and, where they are higher, establish the reasons and, if possible, find ways of reducing them. And cost differences for different projects within a country should feature in audit risk assessment. But such differences are common and often have nothing to do with transfer pricing. They may reflect greater physical and technological challenges, higher costs imposed by regulation38 or greater perceived risks of providing goods and services to particular countries.
There is, however, some scope for applying specific transfer pricing rules to extractive industry costs. Again practices on this vary significantly from one country to another.
For oil and gas, joint ventures (JVs) are common. An operating company incurs costs on behalf of the JV and bills each participator for its share. The other participators have an adverse interest to the operator in respect of shared costs. If an associate of the operator charges excessive transfer prices, it will reduce the other participators’ profits as well as the government’s tax. Joint operating agreements, fairly standard throughout the industry, therefore incorporate specific fact-based transfer pricing rules. These are based on the principle that costs charged by an associate should be charged at original cost to the associate. Participators are given powers to audit compliance. This gives governments significant protection against transfer pricing abuse.39
Governments cannot necessarily leave it entirely to JV partners to enforce this no-profit rule, but they can and often do build it into PSAs (closely modeled on JV agreements) and/or petroleum tax legislation. The no-profit rule might at first sight seem inconsistent with OECD guidelines, but is arguably the CUP for costs charged between non-associated participators in petroleum JVs the world over.40
Where this rule applies, tax authorities need to be able to establish that goods and services were actually provided by associates at cost. General information requirements may be adequate for this purpose, but, if not, specific provisions may be needed. Some PSAs allow the government to require a certificate to this effect from an approved independent auditor of the associate’s records. Alternatively costs charged by associates could be disallowable if for any reason (including claimed confidentiality) the taxpayer could not or would not show they were charged at original cost.
For mining, JVs following the no-profit principle are not common, nor is it standard practice to build this principle into mining legislation and contractual agreements. There may, however, be scope for doing so. Associates often supply mining goods and services that they have in turn brought in from external providers, and it may be reasonable to treat the cost to the group as the relevant CUP for determining the intra-group transfer price. (Again ability to require evidence of the cost to the group would be vital.) An alternative is to allow a small markup under the cost plus basis to reflect any value contributed by the associate, but governments must be aware of the risk that companies will inflate costs by routing them through tax haven-based service companies that add little value in reality.
Alternative specific transfer pricing cost rules are possible. One fairly common approach is to limit management service charges to a maximum percentage of total operating costs or total revenues. The percentages used vary significantly from country to country. This is clearly an inaccurate and somewhat arbitrary method of determining arm’s length prices, and the risk of such “safe harbor” rules is that the ceiling quickly becomes a floor. But it does have the important advantages of clarity and objectivity. If it is built into legislation or contractual agreements at the outset, then governments can take the generosity or otherwise of the limit into account in planning their overall natural-resource fiscal regime, and companies can similarly take it into account in planning whether to invest in the country concerned. Within limits, therefore, such rules are often found to be acceptable and workable in practice.
PSAs often contain standard rules for costing previously used equipment.41 Again these may not strictly meet the arm’s length standard but have advantages of simplicity and predictability.
Mining and drilling costs charged by associated companies can be very large and will present significant risks. If it is not possible to “look through” to the original cost, it may be possible to use standard rates, for example, for hire of drilling rigs, but there will often be special factors that make like-for-like comparison difficult, and data may be difficult for tax authorities to obtain. (Countries with natural resources might well benefit from exchange of such data.) It may be impossible to devise specific transfer pricing rules, in which case it is essential to require companies to justify the prices charged.
Payments to associates for intellectual property (special processing technologies, technical research, etc.) are less common than in other industries but are sometimes made. Ownership is often located in tax havens. Pricing is notoriously difficult. Ring-fencing rules may disallow or at least limit such costs if they do not relate specifically to a project in the country concerned, but if not, claims for such costs may have to be considered on their individual merits (which may be questionable).
Ring-fencing adds further difficulties. It may be unclear how shared costs should be apportioned, for example, where machinery is moved from one site to another, or (even more difficult) where costs have to be apportioned between oil and associated gas. The best course may be to prescribe or agree on formulary rules based on measurable outputs or inputs that, while they may not strictly meet the arm’s length standard, produce a reasonably fair and evenhanded result if applied consistently. Devising, applying and monitoring such rules can, however, be difficult.
Companies may avoid tax by charging excessive finance costs. This puts income tax at risk, because finance costs are generally deductible, whereas costs of equity financing are not.42 Some governments negotiate generous interest deduction rules as a deliberate tax incentive,43 but the following discussion assumes the aim is to apply arm’s length terms.
Finance costs can be excessive in two ways:
They may be excessive relative to the amount borrowed (e.g., interest rates or guarantee or facilitation fees at higher-than-normal market rates); and/ or
They may be excessive because the amount borrowed is itself excessive (generally described as “thin capitalization”).
Thin capitalization is a special type of transfer pricing problem. It describes the situation where a taxpayer that is part of a group of companies borrows more than it could or would if it were an independent entity. Thin capitalization risks are not unique to extractive industry taxation, but there are the same special incentives and opportunities as for other transfer pricing abuse. Special features of the extractive industries, such as the high risk of exploration, and their exceptional importance in some cases to government revenues may be seen as justifying special rules.
There is a range of approaches to countering thin capitalization; some lend themselves to more effective and transparent administration than others. Some countries have little protection other than general transfer pricing rules (which, as discussed, are often poorly designed and are particularly hard to apply to this issue). Others impose specific limits. For example:
A maximum debt-to-equity ratio, with interest disallowed on borrowing exceeding that ratio. There is considerable variation, ranging from 1:1 to as much as 4:1. Legislation sometimes leaves it unclear whether this represents a “safe harbor” (with the disadvantages mentioned earlier) or simply a maximum level subject to further possible restriction under general transfer pricing principles. Often there is no special debt-to-equity ratio for extractive industries, so that a generous ratio may apply even to an exploration company operating in a previously unexplored and/or politically unstable area (which could probably borrow nothing at all on arm’s length terms if operating independently). Sometimes the restriction applies only to borrowing from abroad and/or borrowing from associates. This may allow excessive borrowing by means of “back-to-back” loans (loans from an associate routed through an independent bank) or parent company guarantee (explicit or implicit). The rules often leave scope for argument on the definition of debt and equity.
A limit imposed by reference to the purpose of borrowing. For example, interest may be deductible only on borrowing to fund development costs or a maximum percentage of such costs. A risk with this approach is that companies may accumulate debt no longer required for the original purpose, so an additional ongoing finance requirement test may be needed. An example can be found in Uganda’s model production sharing agreement, which allows interest on loans to finance development operations only to the extent it relates to debt (from any source) not exceeding 50% of the total financing requirement, and disallows interest on loans to finance exploration. A restriction of this kind could possibly be supplemented by regulations or guidance to the effect that the financing requirement was to be taken as the cumulative negative cash flow including tax paid but excluding other disallowable costs.
A limit restricting interest to a maximum percentage of taxable profits (sometimes known as an earnings-stripping rule). Again limits vary from country to country. Because disallowed costs are usually carried forward to future years, the effect may be that excessive deductions are merely deferred rather than disallowed altogether.
Deductibility of interest is normally a condition for foreign tax credit, and there is a theoretical risk that it could be prejudiced by such anti-avoidance rules. But there is little evidence of this in practice.
Some countries take simpler approaches than others to limiting interest rates paid. General transfer pricing rules may be relied on to limit interest to a normal commercial rate, but this can be difficult to define. Cases have been seen in which, with inadequate thin capitalization rules, companies borrow far more than they could do if operating independently and then argue that the interest rate should be high to reflect the exceptional risk of such high borrowing! An alternative approach is to set the maximum rate at a specific percentage above a quoted benchmark rate, for example, 6-month dollar LIBOR + 2% (assuming that LIBOR is considered a sufficiently reliable measure). This is strictly a departure from the arm’s length standard (since companies might be able to negotiate different rates on the open market), but so long as it reflected typical borrowing rates, it might be considered a reasonable approximation and has advantages of clarity and predictability.
Countries sometimes use ministerial or departmental discretion to limit finance costs for large projects. Companies have to seek approval of loans and of the terms applying to them. Sometimes this is the only protection against excessive finance costs; sometimes it is combined with other restrictions. Clearly it is better than nothing, but it is preferable to have published rules that are administratively transparent and effective.
Some countries re-classify disallowed excessive finance costs as dividends. This adds complexity but, depending on the country’s WHT rules, may be necessary to ensure they do not have tax advantages over dividends.
Disguised interest presents a further risk, so restrictions on deductibility need to apply to all finance charges, not just those explicitly described as interest. Companies may disguise interest to circumvent restrictions on deductibility of the kind discussed, to circumvent the (normal) disallowance of interest for taxes such as RRT or production sharing or to avoid withholding tax on interest payments. Finance leases are one instrument that may be used for this purpose. A finance lease is in legal form an asset rental but in substance a loan-financed asset purchase (since under the terms of the lease the risks and benefits of ownership are passed to the lessee). International accounting standards recognize the substance and require a proportion of the lease payments to be characterized as interest, but in some countries tax rules do not.44 Other kinds of payment (e.g., guarantee fees) or financial instruments (e.g., interest rate swaps) may also be used to disguise interest costs.45
6 Administrative procedures
Practical and procedural issues need to be addressed if transfer pricing is to be controlled effectively. An essential starting point, already discussed, is that the onus should be on taxpayers to apply transfer pricing rules within a self-assessment framework. This requires that valuation rules should, as far as possible, be clear, predictable and not open to dispute or manipulation. General transfer pricing rules may need to be supplemented by regulations – any special extractive industry rules will clearly need legal or regulatory support. Guidance should be published explaining in detail how the rules will apply in practice if there is any ambiguity about this. Tax authorities should discuss such guidance with industry representatives and seek their agreement. It should be regularly updated in the light of experience.
It then falls to tax auditors to monitor compliance and enforce it if necessary. The key principle of modern tax administration is risk management, and a coherent risk-based approach is needed for transfer pricing. To facilitate this, taxes on extractive industries that are vulnerable to transfer pricing abuse should ideally be audited by a single department, possibly on the basis of a consolidated tax return. Tax returns should be designed to provide data required to assess transfer pricing risks. Taxpayers should be obliged to disclose significant associated party transactions. Cost analyses should be designed to identify categories particularly at risk of transfer pricing abuse, such as management charges.
The tax authority will need to obtain, maintain and analyze data required for transfer pricing risk assessment, develop risk factors, and continually refine them in the light of experience. The data will include not just returns data but also relevant taxpayer data – for example, group structure (especially links with tax haven–based associates) and past compliance history – some of which may need to be gathered from outside sources, such as consolidated group accounts, company websites and foreign tax authorities. It will also include third-party data on production volumes, arm’s length sale prices, benchmark sale prices and costs. The industry regulatory department will be a key source of this data, and effective exchange of information is essential (assuming the taxing authority is different from the regulatory authority).
Once a decision has been made to audit transfer pricing, it will be necessary to identify high-risk transactions (including those ostensibly with non-associates), obtain explanations of how they were priced, examine contract documentation and seek further explanations as necessary. If mispricing is established, follow-up action – penalties, education, legislation or whatever – should be planned to reduce the risk of recurrence.
This is demanding work and requires competent, well-trained and qualified staff. Staff employed on this work often do not have the pay, status and authority needed to ensure the required quality, even in countries where only a few key staff are required to administer the tax of a few large companies paying the lion’s share of government revenues.
Tax authorities sometimes create specialist transfer pricing teams to assist general audit staff with transfer pricing audits. This may be less appropriate for extractive industries, in which auditors should normally be industry specialists, who will be more familiar with the valuation issues peculiar to the industry and with any special methods applicable. Industry specialists should, however, receive training to increase understanding of transfer pricing risks, issues and procedures.46
Where the law imposes a special transfer pricing rule or a general valuation rule for sales, special administrative procedures are required. Key formal procedures become those that establish quantity and price, and audit of revenues becomes largely a question of checking that the quantity-times-price valuation imposed by law is correctly applied in taxpayers’ returns. (Where there are no special rules, data on quantity and price should be gathered for risk assessment and audit purposes but will not have the same formal legal status.)
There will usually be documentary evidence of quantity and quality (for example, in sales documents), but real-time monitoring by the authorities of physical production and measurement is advisable to some extent – so-called physical audit. This tends to be more straightforward for oil and gas, which pass through a limited number of delivery points where special metering equipment is installed, than for minerals, which may have a wide range of routes to export, may not have accurate measuring facilities such as weighbridges installed and may need to be assayed by subjecting samples to detailed laboratory analysis. This requires specialist expertise and procedures and gives obvious scope for error and dispute. Physical audit is therefore not foolproof and is best regarded as a supplement to, not a substitute for, financial audit.
Physical audit is often the responsibility of a department other than the tax department – for example, the industry department or, for exports, the customs department. There is considerable variation in countries’ approach to physical audit. Some exercise little oversight and rely on the accuracy of company measurements and documentation. At the other extreme, some require direct government measurement of all production.
Companies should be subject to clearly defined obligations to measure and record physical production. The point of measurement for sales valuation should be clearly defined. Companies should be required to implement controls to ensure that all production reaches that point (and to control and measure unsold stock). At the valuation point they should be required to put in place equipment and procedures to accurately measure volume and quality and systems for recording those measurements, which should be reported to the responsible agency under specified rules. The responsible agency should not necessarily have to take all the obligations of measuring volume and quality on itself, particularly in the case of large MNEs – instead it should oversee companies’ performance of those obligations and monitor measured output against other data, such as production plans. Large companies are at constant risk of misappropriation of resources by their employees, particularly in the case of precious minerals, and will normally have sophisticated systems and controls of their own in place to prevent it and to ensure proper valuation of sales. The responsible agency should review the adequacy of those systems and controls and can generally build upon them rather than seek to replace them wholesale with its own. It should have unfettered rights to physically inspect the movement of resources up to the point of measurement and to physically monitor and test company measuring equipment and procedures but should adopt a selective risk-based approach reflecting both the amounts at risk and the probability of error (taking company characteristics and history into account among other factors). It should carry out tests (requiring competence in mineralogical analysis and sampling techniques) at unpredictable intervals. It should have a clearly formulated strategy and plan for physical audit and should keep records of its physical audit checks and their outcomes. Where inaccurate measurements or recordings are established, procedures should exist to correct them, including reasonable extrapolation to other periods. There should be penalties for such failures and rights of arbitration where there is disagreement over volume and quality measurement. Procedures should be established for production data to be passed to the department responsible for tax audit.
Where production has to be valued on the basis of specified benchmark prices (or on the basis of average prices of actual transactions), there should be clear procedures in place for establishing and publishing those prices. In some countries the tax department is responsible for determining prices; in others it is the industry department or national resource company (but this may create conflict of interest if the company operates commercially); or different agencies may have joint responsibility. In some countries the responsible department takes the lead in identifying suitable benchmarks and calculating the prices to be used; in others resource companies put forward proposals along with supporting evidence, which the department can accept or amend. The latter approach may foster better relations with the industry, enhancing voluntary compliance. Either way, the responsible department needs to have procedures (and funding) in place for identifying, accessing and recording relevant price data.
Normally companies will have rights of appeal against assessments, but where values to be used are prescribed in advance, there is often separate provision for arbitration of valuation disputes, usually involving international experts rather than local courts. In other cases, prices to be used are simply imposed by government. If pricing rules are clear, objective and predictable, rights to arbitration may not be necessary, but often the choice of benchmark, the adjustments made to it, the pricing of actual transactions (where these feature in the calculation), the weighting of different elements of the calculation (where combinations of methods and benchmarks are used) and the calculation of weighted averages are open to dispute, and an arbitration mechanism is appropriate.
In theory, if rules prescribing prices to be used are clear and simple, companies can be left to apply them; in practice, it is often best if they are published by the responsible agency. Pricing can be quite complex for the reasons explained. The data required to calculate prices may not be readily available to all who have a need to report or monitor them. Even if the calculation is simple (e.g., a simple unweighted average mineral price quoted on a particular exchange), it is still best, in the interests of transparency and certainty, if the actual figures to be used are clearly published, for example on an official website. When annual tax returns are received, the correct application of the prescribed prices should be checked as part of the audit or audit risk assessment process. Publication of the figures should leave no room for dispute between companies and tax auditors.
MNEs generally favor the arm’s length standard for pricing controlled transactions but prefer to be left to make their own choices as to the most appropriate method, including whether to make use of quoted prices, which ones to use if they do, how to adjust them for comparability differences, what contractual terms to use, what pricing date conventions to apply and so on. They see this flexibility as essential to respond to the complexities of commercial commodity markets. They accept the right of tax authorities to challenge transfer prices but want them to do so with a full understanding of all the complex facts and circumstances that may be relevant. They see prescriptive standardized rules as bureaucratic and inconsistent with arm’s length principles, since they inevitably fail to reflect the complexity and variety of commercial pricing.
Governments of developing countries, on the other hand, are uncomfortable with some transfer pricing methods that they feel are less favorable to commodity producers than CUP. They are aware too of the risk of sophisticated transfer pricing strategies being used to shift profits. Many have introduced improved transfer pricing legislation consistent with international standards and are seeking help with its application and enforcement. But often their tax authorities lack resources and capacity to establish all the relevant facts and circumstances, let alone identify and challenge abuse effectively; and their standards of governance are ill suited to complex negotiations with a wide range of possible outcomes. “Must try harder” is the advice they often receive, but they may feel that the improvements in capacity and governance needed to counter abuse effectively are, realistically, unlikely to happen any time soon (considering that even developed countries struggle to do so). The arm’s length standard may be fine in principle, but if they embody it in law but cannot enforce it, the practical result is likely to be not arm’s length pricing but base erosion and profit shifting – which, once their non-renewable resources are gone, they will not have another chance to put right.
It is perhaps unsurprising, therefore, that developing countries have sought to develop more simple, objective and predictable transfer pricing methods for commodities, often based on quoted prices. These can be criticized as not fully consistent with arm’s length standards (and sometimes they are themselves vulnerable to abuse). But for oil such methods are now well established, and companies have generally learned to live with them if not to like them. They may not produce entirely realistic arm’s length prices, but may be acceptable if they are applied evenhandedly, have some basis in normal commercial practice and achieve a reasonable approximation, taking one transaction with another. For other commodities, the challenges of developing standard, simplified methods that meet those criteria are greater, there is probably more variety in the methods countries have developed in practice and in many cases investors consider them seriously flawed and unacceptable. It must be hoped that the debate initiated by the OECD on this topic and the further research proposed will identify methods that developing countries find appropriate for commodities and within their capacity while producing results acceptable to investors. But there is still a long way to go.
CalderJack. (2010) “Resource Tax Administration: The Implications of Alternative Policy Choices” in PhilipDanielMichaelKeen and CharlesMcPherson (eds) The Taxation of Petroleum and Minerals – Principles Problems and Practice. (London, New York: Routledge) pp. 331–334.
HMRC (2014) Oil Tax Manual Available at http://www.hmrc.gov.uk/manuals/otmanual
KellasGraham. (2010) “Natural Gas: Experience and Issues” in PhilipDanielMichaelKeen and CharlesMcPherson (eds) The Taxation of Petroleum and Minerals - Principles Problems and Practice. (London, New York: Routledge) pp. 163–183.
OECD. (2013) Public Consultation: Draft Handbook on Transfer Pricing Risk Assessment.
OECD. (2014) BEPS Action 10; Discussion Draft on the Transfer Pricing Aspects of Cross-Border Commodity Transactions.
In this chapter “commodities” refers to hydrocarbons and minerals. Special rules often apply to other commodities too, but these are not discussed.
In this chapter “tax” is used to mean any payment imposed by government, including, for example, royalties and production sharing.
“Upstream” describes exploration and extraction; “downstream” transportation beyond an initial delivery point, refining/processing and distribution, marketing and sales of finished products.
Examples of common taxes not based on revenues or profits are signature and production bonuses, volume-based royalties, surface rents, withholding taxes, payroll taxes, dividend taxes, import duties, free and carried equity and infrastructure building obligations. Some “nuisance” taxes for example, education and regional levies, may be based on revenues and vulnerable to transfer pricing abuse, but are not discussed further here.
Governments can apply a special low-tax regime to extractive industries, and some have done this at times, particularly with mining. In that case transfer pricing can reduce tax by overvaluing revenues and undervaluing costs, thus shifting profits into the mining tax regime. When commodity prices are high it is more common for extractive industries to be subject to a high-tax regime.
For example, say company A paid associated service contractor B $100k when the arm’s length rate was $90k, subject to a typical 10% WHT (if one third of a contractor’s receipts represents profit, this equates to a 30% tax on that profit). If tax was charged on A’s profits at, say, 60%, the excess $10k payment would reduce A’s tax by $5k ($10k at 60% less $10k at 10%). B might have to pay foreign tax on the $10k but almost certainly much less than $5k.
OECD Global Forum on Transfer Pricing, Draft Handbook on Transfer Pricing Risk Assessment ¶38 (OECD 2013).
For example, unpublished research by UK tax authorities identified significant tax loss from manipulation of North Sea oil transfer prices (HMRC 2014; http://www.hmrc.gov.uk/manuals/otmanual/ot05012.htm).
There are less sophisticated ways of exploiting this than tax planning – for example, smuggling abroad production designated for domestic processing or consumption.
The granting of tax holidays for mining operations is fairly common in developing countries. The obvious transfer pricing risks this presents are sometimes exacerbated by lax record keeping and reporting requirements during the tax holiday period.
It is beyond the scope of this chapter to consider the policy arguments for and against area-based ring-fencing, but the administrative burdens it imposes and the capacity of companies and tax authorities to meet them are clearly factors to take into account.
Appendix 1 to Chapter 11 in Taxation of Petroleum and Minerals – Principles, Problems and Practice (Calder 2010) discusses the taxation of hedging instruments in more detail. It outlines the arguments for recognizing hedging gains and losses (basically, they affect companies’ profits) and against (basically, they present risks of tax avoidance, and governments should anyway decide for themselves how far to hedge natural resource revenues and not leave it to the vagaries of whether MNEs choose to do so).
Lao PDR’s is a rare example of a country without transfer pricing rules in its primary tax legislation, though rules are included in mining contractual agreements.
Cases are sometimes seen where a parent company and subsidiary fall within the definition but not sister companies under common control.
For example, A could sell production for $X to non-associate B, who as part of a wider agreement undertook to sell it for X + y% to a refinery owned by A’s associate. B has no interest in X but only in y, so X is a non–arm’s length price. An example of rules to counter this can be found in the UK’s legislation; see http://www.hmrc.gov.uk/manuals/otmanual/OT05025.htm.
Similarly, section 482 of the U.S. Tax Code ostensibly leaves the onus on the tax authority to identify and adjust non-market prices, but this is supplemented by extensive regulations that in effect require companies to use arm’s length prices for tax return purposes and impose penalties for failure to do so.
Chapter 3 (Shay, this volume) discussed APAs more fully, and expressed reservations about the limited wider benefits from the resources devoted to them and their lack of transparency. Those reservations are even more relevant in developing countries. Discussions with extractive industry MNEs suggest that they are virtually never used in developing countries in practice. They may nevertheless be an option to consider if other means of producing certainty cannot be found. There may also be cases in which tax auditors and MNEs can reach working agreements without the need for a formal APA.
OECD, BEPS Action 10; Discussion Draft on the Transfer Pricing Aspects of Cross-Border Commodity Transactions December 2014 (http://www.oecd.org/ctp/transfer-pricing/discussion-draft-action-10-commodity-transactions.pdf).
For example, business organizations claim that Argentina’s application of the sixth method is unfair because it defines controlled transactions too widely, does not adjust quoted prices appropriately and imposes shipping date price even where other pricing dates are required by normal market conventions, thus – because prices are only adjusted upwards – essentially penalizing companies for market volatility.
Mining and petroleum agreements in some countries apply a specific transfer pricing rule to non–arm’s length sales but also have a general rule requiring all sales to be made at international prices or forbidding discount sales, potentially giving the tax authority scope to re-value a sale even if ostensibly negotiated at arm’s length.
Brent and West Texas Intermediate are among the most commonly quoted.
A minor adjustment may also be needed if the volume of the transaction differs significantly from that on which the benchmark price is based.
Where oil is not freely tradable on international markets, domestic oil prices may be significantly affected by local supply and demand – for example, where there are legal export barriers (as in the U.S.) or limited export pipeline capacity (as sometimes experienced in Russia). In such circumstances international quoted prices may be of limited relevance, and it can be particularly difficult to establish arm’s length prices if the domestic market is dominated by vertically integrated businesses.
PSA valuation rules vary but might typically require arm’s length sales to be valued at actual sale prices and non–arm’s length sales at the weighted average of arm’s length sales in the month or quarter but with a provision that if less than half of the oil sold in the period is sold at arm’s length, they should instead be valued at the average price of one or more designated benchmark crudes, adjusted as necessary for quality and transport cost differentials.
Policy objections to aligning income tax and PSA valuation rules are not obvious, but one might be that meeting international investors’ expectations that income tax will be fully consistent with OECD international standards is more important than the administrative advantages of consistency.
See Graham Kellas’s chapter in Taxation of Petroleum and Minerals – Principles, Problems and Practice (2010) for a comprehensive discussion of LNG pricing.
Long-term contracts are particularly common for gas but are sometimes used for other commodities, and similar issues will arise.
In the UK, for example, marketing fees are deductible for petroleum revenue tax purposes only if the sale is to a non-associate.
Many countries require this for all sales, not just those to associates. This can be expensive (in Sierra Leone in recent years, valuation fees have amounted to around a third of royalties and export duties on diamond sales). Where the intention is to base taxes on actual values, it is not clear why government valuations should be substituted for arm’s length sale prices, particularly where it is clear that the buyer is not an associate.
There are often many additional “nuisance taxes”, not discussed here.
This approach is broadly followed by Zambia for copper, though there is debate about its implications for local smelting.
This is comprehensively discussed in Otto et al.: Mining Royalties (Washington, DC, World Bank)
The greater risk would be that processing was unprofitable, so that its inclusion reduced RRT.
It includes value added by on-site processing, however. The policy rationale for taxing this but not off-site processing is not clear, and it may encourage companies to reduce royalties by moving processing off site.
In broad terms this is the approach adopted in Australia for the purposes of its federal resource rent tax. (Royalties are imposed at state level, and varied methods are used.)
This, broadly, is the approach adopted to royalty valuation in Mongolia, for example.
Local content obligations, for example, can have an impact on costs and are common in developing countries.
It may for this reason be prudent for governments to award petroleum licenses to JVs rather than to single companies.
One effect of limiting cost to original cost to the group may be to limit deductions for insurance payments where an MNE self-insures through a captive insurance company that does not re-insure outside the group.
For example, used equipment of a defined age and state of repair will be valued at x% of original cost.
The application of WHT to dividends may counter some of the tax advantages of excessive borrowing.
For example, a London-based mining company investing in a Sierra Leone subsidiary in the aftermath of the civil war persuaded the government that the subsidiary should get interest deductions on loans from its parent company because no bank was prepared to lend to it!
Another advantage of finance leases for companies is that they may circumvent PSA rules giving the government ownership of assets acquired for petroleum operations if the PSA terms do not treat a finance lease as an acquisition.
Interest-hedging instruments may provide scope for other kinds of tax avoidance, especially if there is no requirement to “mark to market”, and offset finance costs against finance gains.
Auditors in developing countries often see themselves as responsible for controlling industry costs and revenues generally, particularly where they belong to the industry department. This can result in a lack of focus on transfer pricing risks. The author has in several countries met senior audit staff who expressed concerns that companies were undervaluing sales and overvaluing costs but had no real idea of the extent to which they transacted with associates.