Chapter

8. Fiscal issues for cross-border natural resource projects

Editor(s):
Michael Keen, and Victor Thuronyi
Published Date:
September 2016
Share
  • ShareShare
Show Summary Details

1 Introduction

Natural resource projects located in one state requiring transport across another to access world markets present special challenges for effective fiscal regimes, particularly for low-income countries. For example in the case of large mining projects, typically the investor will have responsibility for the development of the required rail and port infrastructure as part of the project. This in turn requires the governments in the transit and mine jurisdictions to develop models for the determination of taxes, royalties, and other government charges in their respective jurisdictions.

Although the project’s total expected return can be calculated, there are no economic principles commanding a particular allocation of revenues and profits between the parts of the project – the mine and the transportation infra-structure – and hence the two governing jurisdictions. There would be no investment in the rail and port absent the mine; similarly without the rail and port, the mine would not be developed. In the case of one integrated project with the same owner, there is only a single rent or return to be taxed by the two sovereigns. This issue is, of course, not specific to mining. It is also applicable to oil or gas produced in one country and piped through another or for that matter to any integrated project straddling national borders.

Compared to traditional single-country projects, cross-border natural resource projects impose additional burdens on both governments and investors, making already complex transactions even harder to bring to fruition. For investors there are the added costs of negotiating and coordinating with multiple governments and additional political risk. For governments there is the need to coordinate infrastructure and operations as well as to manage differing development interests, which may affect project costs and benefits. Finally, there is the need to agree explicitly or implicitly on a division of project returns and associated tax revenues between the two countries.

Figure 8.1 provides a simple illustration of the revenue flows from an integrated project including a mine and rail and port facilities. In one case, all are located in a single country. In the other case the mine and part of the rail are located in Country A (the “mine jurisdiction”) and the port and other part of the rail are located in Country B (the “transit jurisdiction”).1 The real economics of the project absent the border in the two cases are the same, but the existence of the border and two sovereigns adds considerable complexity and may add real costs.

Figure 8.1Illustration of a single-country versus cross-border resource project

In this analysis, we examine a number of approaches to allocating taxable income or tax revenues between the two countries including the arm’s length principle endorsed by the OECD as well as other alternatives such as the use of a formulary basis or a special fiscal regime with the states “bargaining” over their respective shares. In some cases the tax issues may be resolved through structuring solutions in which transportation is provided by a separate operating entity with a different ownership structure. Each presents its own set of problems, and in all cases care must be taken to preserve for the mine jurisdiction a real return for its depletable resource while also fairly compensating the transit jurisdiction for its value added to the project economics.

It is also important to recognize the complexities associated with the threeparty nature of the negotiations. The investor is concerned with the total burden of taxes, royalties, and other government charges on the project. In the process of negotiations the investor will also be sensitive to tax concessions in one jurisdiction that can be picked up and mirrored in the second (or elsewhere). In turn each of the governments is concerned with getting a “fair” share of the revenues consistent with its own tax and social and economic policies. Depending on the degree of coordination of the two tax regimes, the gain of one may come at the cost of the other. In some cases, a coordinated strategy may increase the total gain for both. Our hope is that making the complexities more transparent will lead to better-structured agreements that maximize economic benefits and allocate them fairly to the relevant parties.

2 Arm’s length pricing

The situation in Figure 8.1 is hardly a unique tax situation. It is inherent in any multinational enterprise operating across taxing jurisdictions or more locally for enterprises operating in multiple sub-national jurisdictions as in the United States. In theory, the government and the investor could negotiate and determine the transfer price (or the principles for determining the transfer price) in the course of an audit or advance pricing agreement after the investment has been made. But given the investor’s interest in determining its potential tax exposures when making the decision to invest and the existence of multiple jurisdictions, the investor is likely to seek some assurances from the taxing authorities up front regarding the applicable principles of taxation and for the determination of taxable income in each country. This need not be an agreement regarding actual charges; rather it is an agreement of how they would be calculated on the basis of expenditures actually incurred.2

The classic OECD approach for income tax purposes (and note that income tax is only one element) is to act as if the two national operations are separate entities and then to determine taxable income for each based on assumed transfer prices determined according to the arm’s length principle. In this context that means determining the charges that an independent transportation company (“Transco”) would impose on the mining company (“Mineco”) for the rail and port services provided in the transit jurisdiction.

This starting point is in fact an artificial division of the project. In a large min-ing development the mine and the infrastructure that transport the resource form an integrated whole. Typically management and operation of the transportation network and its careful integration with the mining operation are critical to the success of the investment and are one of the reasons separation of the logistical network from the mine is rare in large-scale investments. In fact, while the main driver of the investment is the presence of the natural resource, frequently the rail and port facilities constitute the larger part of the capital investment and an even larger part of operating costs than the mine. The division of the project among political jurisdictions is even more artificial since only a part of the delivery infrastructure is located in the transit jurisdiction, the other part being in the mine jurisdiction.

Once a project is thus apportioned among political jurisdictions, the OECD guidelines for determining arm’s length prices use a hierarchy of methods beginning with the comparable uncontrolled price (CUP). Applying the CUP requires identification of what independent companies would charge for the services in comparable circumstances, that is, what an independent rail and port operation would charge for the services being provided in the transit jurisdiction. For the reasons set out, however, most projects are developed on an integrated basis. In fact there are no comparable fully independent rail and port operations serving mining operations, especially within Africa, to provide comparable prices.3

The lack of comparable transactions leads one to the second traditional OECD method, the cost-plus method. The OECD guidelines (2010) for the cost-plus approach read:

  • [The cost-plus method is] a transfer pricing method using the costs incurred by the supplier of property (or services) in a controlled transaction. An appropriate cost plus mark up is added to this cost, to make an appropriate profit in light of the functions performed (taking into account assets used and risks assumed) and the market conditions. What is arrived at after adding the cost plus mark up to the above costs may be regarded as an arm’s length price of the original controlled transaction.4

While the formula may appear straightforward, determining the costs incurred and, more importantly, correctly accounting for the “risks assumed” and the local “market conditions” requires judgment calls in practice. Even minor differences in the assumptions that are made can significantly affect the pro forma financial results of the project and, consequently, the estimates of the corporate income taxes due to different fiscal authorities. The OECD acknowledges that

“because transfer pricing is not an exact science, there will also be many occasions when the application of the most appropriate method or methods produces a range of figures all of which are relatively equally reliable.”5

Unfortunately, as our analysis will show, the range of figures that the method can produce is so wide as to undermine the credibility of the method itself for these purposes.

2.1 Cost-plus method overview

To model the cost-plus approach one can think about the charges an independent transportation company (“Transco”) would charge an independent mining company (“Mineco”). In various regulated industries and project finance structures these are usually rendered as two-part charges: one part to recover the cost of capital with appropriate return on the investment and one part to recover the operating expenses. The first charge is frequently referred to as an “access fee” – a fixed annual charge that reimburses Transco for its capital investment. If the parties were unrelated this access fee almost certainly would be accompanied by some form of a take-or-pay provision under which Mineco would be required to pay the access fee whether or not it shipped the product.6 The effect of this is to transfer to Mineco the project risks of Transco.

The second charge, a “variable fee,” is a volume-based charge based on the operating costs incurred in providing the shipping services elevated by a markup that provides some return to the operating activities. The analogy to rate setting in the context of regulated industries is obvious, but there are also some differences. Regulatory rate setting is usually conducted in the context of historical data, often for a group of highly comparable entities, and importantly is usually accompanied by a highly developed set of regulatory institutions to make the necessary determinations.7 More fundamentally there is usually no identity of carrier and shipper.

While the overarching principle of the two-part fee appears to be a reasonable way to generate a transfer price between the related entities, each of these fees is composed of a number of elements for which various choices are plausible and which can significantly affect the estimate of taxable income.

2.2 Elements of the access fee

Broadly stated, determining the access fee requires the parties to make an estimate of all capital expenditures (capex) that will be incurred in the construction and ongoing maintenance of the rail, port, and rolling stock to calculate the net present value (NPV) of the capital expenditures and then to allocate the recovery of those costs over the project’s lifetime to arrive at the annual charge that would allow the investor to recover the cost of the investment (including the assumed required return on capital).8 Formally, the NPV of the capital expenditures would be calculated as:

where t is the time period in which a capital expenditure is incurred, N is the total number of periods (usually years) in the project’s lifetime, Capext is the capital expenditure incurred at time t, and r is a discount rate (which is discussed at significant length in what follows).9 Next one must calculate the required annual access fee that would ensure that all capital expenditures are recovered, including a minimum return on capital. The access fee must satisfy the following equation:

where the Access Feet is the pre-tax annual access fee. The formula makes it clear that the access fee is set so as to recover the present value of capital expenditures for a given discount rate and that the only difference between the two is in the timing of the inflows and outflows. While the capex tends to be lumpy and front-loaded, the annual access fee is a fixed per period charge.

The challenge of the exercise is that all of these variables have to be estimated by the parties up front, while each element of the calculation is highly uncertain. For example, the numerator in the NPVCapex formula, Capext, will depend on the unique environmental, geological, and engineering features of the project. While companies with significant expertise in construction and operation of large-scale infrastructure projects use their best judgment to assess the likely costs, each environment carries unique risks that may lead to cost overruns, delays, and other unexpected expenses. In the world of project finance, a 10 to 15 percent cost-overrun “cushion” would typically be added to the numerator to account for this likely volatility in actual project costs.

To counter this uncertainty, the tax authorities could agree on a methodology for determining the access fee and ultimately calculate the access fee according to actual capital costs incurred prior to operation.10 However, even then the formula would need to take into account the subsequent capital expenditures that will be incurred during the lifetime of the project. Alternatively and preferably, the access fee will have to be adjusted periodically to account for the additional capital expenditures.

An even more important source of variation in estimates of the access fee will come from the choice of the discount rate r that appears in the denominator of both of the formulas shown. The higher this discount rate, the higher will need to be the dollar amount of the access fee. The rate most commonly used to discount expected cash flows in project NPV calculations is the weighted average cost of capital (WACC). The WACC is the estimate of the return that debt and equity investors would require to invest in a given project, taking into account the specific risks and market conditions of the country, the project, the industry, and so on. In the case in which a company is financed by debt and equity, the WACC or r is calculated as:

where E denotes the value of equity, D the value of debt, V = D + E is the total value of the firm, Ke is the cost of equity, Kd the cost of debt, and Tmc is the marginal corporate tax rate.

Like the access fee itself, the computation of WACC requires estimating a number of ambiguous parameters: E, D, Ke, Kd, and Tmc. For each of these parameters a range of seemingly plausible proxies are available. Indeed, according to the Association for Financial Professionals “providing the ‘right’ answer about how to estimate each of these variables is a difficult, if not impossible, task.”11 In practice, for most financial transactions that firms undertake, they consider outcomes under various plausible WACC values that often include a range of about 10 percentage points. Such a range of WACC values would result in considerable variation in the computed access fee, which in turn leads to a highly magnified variation in taxable income.

We now consider each assumption that is used in calculating the access fee.

2.2.1 Capital structure

The WACC will be sensitive to the assumed ratio of debt and equity that is funding the project. Since debt is typically less costly than equity and because of the beneficial tax effects of using debt (deductibility of interest payments), a higher debt/equity ratio will ordinarily lower the WACC and hence the computed access fee. For the purpose of considering an “arm’s length” price that would be charged between unrelated parties, one would want to consider what the typical capital structure would be for an independent company providing transport services without the benefit of a throughput contract or other third- party credit support.

To do so, we can look for data on the capital structure of other local transportation companies; however, in many settings few or no comparable companies will exist. Moreover, in such environments, debt markets will likely be relatively underdeveloped, and a typical company will largely be funded by equity, resulting in low debt/equity ratios. For example, according to a widely used data source the average debt-to-capital ratio for railroad companies in emerging markets is 19.8 percent.12 In reality Transco may be largely funded through an intercompany loan from the parent or a third-party loan guaranteed by the parent. This would permit a higher debt/equity ratio (but should also result in the use of a lower interest rate).13 Thus, the range of justifiable value for the debt/equity ratio could be as low as 20/80 when relying on comparables for truly independent projects or as high as 60/40 (when relying on a guarantee from a credit-worthy parent).

2.2.1.1 COST OF EQUITY

The cost of equity is arguably the parameter with the widest range of plausible estimates because of the number of parameters that go into calculating it. While several alternative methods for calculating Ke exist, the most common is the Capital Asset Pricing Model (CAPM).14 It calculates the cost of equity as:

where Rf is the return on a risk-free security, CRP is a country risk premium, Rm is the return on the market portfolio (the difference, Rm − Rf, is known as the equity market risk premium) and beta measures the volatility of the particular issuer relative to the market.

Risk-free rate: While a number of choices are possible, the most commonly used estimate of Rf is the rate of return on a U.S. Treasury bond. However, there are differences in views as to what the correct underlying Treasury security should be (e.g., a short-term bill, a 5- or 10-year note, or a 30-year bond). Given that the historically the difference in returns between 90-day Treasury bills and 30-year Treasury bonds is approximately 3 percentage points, the choice for the risk-free rate will have real effects on WACC. Moreover, once the reference security is agreed, it is far from clear what the time period is from which the rate should be quoted. On the one hand, one can argue that the rate should be the one quoted on the date that the decision to invest in the project is made. However, given that the project would unlikely borrow at a 20- or 30-year horizon and would face refinancing risk, another reasonable argument is to use a smoothed rate, such as a 5- or 10-year average. These estimation differences can impact the bottom line, especially if the spot rate differs significantly from the long-run average.

Country risk premium: If the U.S. Treasury is used as reference for Rf, then a CRP should be added to it in order to compensate investors for the risk associated with operating in countries with less developed markets and more volatile conditions than those prevailing in the U.S. The CRP can usually be gauged from sovereign credit ratings published by ratings agencies or from actual spreads on sovereign bonds. However, many emerging-market countries (including the majority of Sub-Saharan African countries) are not yet included in such standardized ratings and may not have rated sovereign debt outstanding. For such countries, estimates of the appropriate CRP could vary widely, from a range of 8 percent (based on credit ratings of low-grade but rated countries) to upward of 14 percent (based on actual sovereign spreads of certain high-risk countries).

Equity market risk premium (EMRP): The EMRP is the additional return that investors require in order to invest in equity as opposed to a risk-free security. The EMRP is usually calculated by averaging the historical differences in the returns on equity investments and risk-free investment across global markets. But the historical premia that are calculated by academics and practitioners around the world are acknowledged to be extremely imprecise with large standard errors.15 Moreover, these premia may be time varying, and there is no consensus on what the correct rate in a given time period is or whether geometric or arithmetic historical averages should be used for future forecasts. The EMRP remains a highly contentious issue not only in the finance literature – in practice, firms use EMRP estimates ranging from below 3 percent to upward of 7 percent (AFP, 2011). These differences in estimates can lead to large differences in WACC and affect decisions regarding investment projects.

Asset Beta: In mature equity markets, the asset beta is typically calculated as the historical volatility of the price of a company’s stock relative to the market in which it operates. Even for companies with available volatility data, however, different estimates can arise from the use of different time series. Although frequently a 3- to 5-year historical times series is used, there is no agreement in the finance literature or in practice on what the “correct” approach is (AFP, 2011; Fama and French, 1997). For unlisted companies or greenfield projects, the exercise is even more complicated, as there is no existing volatility data. Here, it is common to use the beta of comparable companies operating in the same industry. Again, the lack of readily available comparables in many emerging markets will complicate the analysis, requiring reference to comparables from other markets/regions or from somewhat different industries. Each of the choices can affect the final result. For example, standard calculations of “railroad” industry betas may lump shipping companies together with urban mass-transportation systems, railroad equipment manufactures, and other businesses with different risk profiles, leading to beta estimates ranging from 0.5 to 1. Moreover, in an integrated project in which the primary offtaker is a mine, it may be more relevant to use the asset beta of mining companies, which have historically been higher than those of transportation companies.16

2.2.2 Equity tax adjustment

The return on equity in the WACC formula is intended to be a net equity return, that is, the distribution to equity holders after payment by the project entity of all corporate income taxes and other taxes and fees. Thus in applying the formula to determine the access fee it is necessary to take account of the corporate income tax that would be applied to Transco’s pre-tax income by the transit jurisdiction.17 If Ke is 14 percent but the average corporate tax rate on Transco’s earnings in the transit jurisdiction is 30 percent, the required pre-tax equity return for calculating the access fee would be 20 percent (0.14/(1.00–0.30)). Substituting this into the earlier equations gives the following for the determination of the tax-adjusted WACC or r to be used for the computation of the access fee where Tac equals the average effective corporate tax rate:

Alternative Estimates of Ke. Depending upon the assumptions made, in our example for Transco the CAPM yields a range of possible estimates for Ke even before adjustment for corporate tax (see Table 8.2). While the most optimistic estimate of 12 percent appears unrealistic for many emerging market projects, the highest estimate is above 20 percent, a level frequently cited (but also disputed) as the minimum rate of return an investor would expect for a mining investment in a low-income, low-capacity country.

Table 8.1Ranges of parameters in determining the access fee
ParameterJustifable rangeSources of variation
NPVcapex+ /− 10–15%Cost estimates, environmental, engineering, geological, political risk factors affecting construction
Weighted average cost of capital (WACC):
Capital structure20–60% debtReference to actual versus hypothetical capital structure, choice of comparables for hypothetical arms-length lender
(i) Cost of Equity:
Risk-free rate2–5%Underlying risk-free security, time period over which the risk-free rate is quoted
Country risk premium8–14%Determining reasonable comparables in absence of credit ratings and sovereign spreads
Equity market risk premium3–7%Choice of historical estimate, choice of baseline risk-free security, averaging methodology (geometric vs. arithmetic), timeframe of underlying data
Asset beta0.5–1.0Determining reasonable comparables in absence of historical returns data (which industries and regions are most comparable)
(ii) Cost of Debt:
Risk-free rate+/−1%LIBOR or U.S. Treasury rate. Reference time period can vary.
Risk adjustment+/−4%Depends on availability of sovereign spreads, cost estimates of comparable companies, existence of third-party guarantees
Tax rate25–30%Choice of expected versus target tax rate, actual versus marginal
Table 8.2Key assumptions and financial results under three scenarios
Values in US$ millions unless specifiedLow CaseMid CaseHigh Case
ASSUMPTIONS
Up front capex1,0001,0001,000
Replacement capex (per year)505050
Weighted average cost of capital (WACC)9%14%23%
Capital structure (% debt/equity)60%40%20%
Cost of equity, Ke:12%18%26%
Risk-free rate, Kf2%4%5%
Country risk premium (CRP)8%11%14%
Equity market risk premium (EMRP)3%5%7%
Beta0.50.751
Effective (after tax) cost of debt, Kd:7%8%10%
Cost of debt (nominal)12%14%16%
Inflation adjustment2%2%2%
Tax rate30%30%30%
Operating expenses100100100
Operating cost markup3%7%10%
FINANCIAL RESULTS
Access fee146196278
Variable fee105108110
Operating expenses100100100
Depreciation707070
EBIT81133218
Interest expense443622
Taxable Income3797196
Income Tax112959

2.2.3 Cost of debt

The cost of debt should reflect current market conditions and the perceived riskiness of the particular borrower. Kd is typically determined by reference to a risk-free rate such as the London Interbank Offered Rate (LIBOR)18 or the rate on U.S. Treasury securities plus a risk premium that is based on the credit rating of the borrower (usually a company credit rating available from credit ratings agencies). In data-poor environments where many companies are unrated and in the case of greenfield projects, assessing the cost of debt is more difficult. If one were looking at debt from an independent third party with no affiliate support, one can consider the rate at which the local government borrows and adjust upward/downward based on the perceived credit risk of the project relative to the government. Another method is to look to data of the average borrowing rate of local companies and adjust in reference to that average.

There is an interaction between the assumed debt–equity ratio and the assumed cost of debt. The higher rates calculated on country-specific experience should only be applied in conjunction with the lower debt–equity ratio that would typically apply in such cases. If the debt is guaranteed or supported by contractual commitments or directly provided by an affiliate, then a higher debt–equity ratio may be appropriate, but the rate used for the debt should be related to the actual cost of debt to the affiliated entity or the entity providing the contractual guarantee, for example, the parent company.

2.2.4 Debt tax adjustment

Finally, the pre-tax cost of debt needs to be adjusted to account for the tax benefits of debt in the cost of capital calculation. In principle it should be project and country specific, as it should represent the likely tax rate that will apply to the project for which the access fee is being calculated.

To sum up this part of the discussion, Table 8.1 lists all the parameters that would enter the access fee calculation and the plausible range of variation for estimates of each of the parameters for a hypothetical greenfield port and rail project in an emerging-markets country. We will later show the implications of these ranges for any effort to compute the access fee and in turn taxable income.

2.3 Elements of the variable fee

In contrast to the access fee, which if agreed in advance may be based on a number of estimated factors, the variable fee should be based on the actual costs that are incurred in operations. To permit budgeting and planning, the variable fee in any year could be based on estimated or historical costs subject to a “true up,” that is, reconciliation of estimate with actual after the close of the period. This is a procedure used in many regulated industries or project structures in which costs are passed through.

Still, at least two elements need to be decided: first, what costs are to be allowed. Second what is the appropriate “markup,” if any, by which actual expenses will be elevated in order to arrive at the variable fee and a reasonable profit.

Allowable costs should include all direct costs of operations including materials, labor, insurance, third-party services, and other expenditures that are not of a capital nature. For clarity and to minimize disputes, a schedule of allowable costs should be developed. Where acquisition is from an affiliated entity, accounting may be more difficult. One approach would be to allow affiliated purchases only at cost, relying on the markup to account for any profit that the affiliate entity might have otherwise earned.19 Note any transit fee (see discussion that follows)20 should be an allowable cost for purposes of determining the variable fee.

In the spirit of the OECD transfer pricing guidelines, a cost markup is to be applied to costs in order to provide an “appropriate profit” to the service provider for its activities. Although not so characterized in the OECD transfer price guidelines, the markup might also be viewed as administrative overhead or indirect costs to recognize management services provided by the home office if those services are not recovered otherwise.21 Unfortunately, neither in theory nor in practice is there a consensus on what constitutes an appropriate profit markup for service providers.

There are two other aspects of the variable fee worth noting. The costs used to build up the fee could be the same costs allowed as deductions in determining taxable income in the transit jurisdiction, but that does not necessarily have to be the case. There would be some administrative advantage in preserving this identity, as that would facilitate auditing and would assure the transit jurisdiction that the access fee would preserve some taxable income in all periods.22 Second, to the extent that the variable fee is recognized as a cost to Mineco deductible in the mine jurisdiction, the mine jurisdiction has a strong interest in the elements permitted as recoverable costs and the markup. Of course, as noted throughout, the mine jurisdiction need not necessarily recognize the transfer price negotiated by the transit jurisdiction. In particular the access fee may incorporate elements of a nature or magnitude, for example, a high management fee or markup that is not consistent with the mine jurisdiction’s tax laws.

2.4 Gross income to taxable income

Determining transfer prices or formulas for the access and variable fee is only the first step in determining taxable income. The agreed prices together with actual shipments will generate gross revenues and costs. Then it is necessary to compute taxable income utilizing all of the rules generally applicable to determining taxable income. A couple of points are worth noting. The computation of the access fee, if agreed up front, will embody an assumed debt/equity ratio, but the actual capitalization utilized by the taxpayer may differ from the assumed ratio. Further, the revenue code itself may set limitations on the debt/ equity ratio and amount of interest expense that may be deducted pre-tax, giving rise to additional differences between earnings before interest and taxes (EBIT) and taxable income.

Similarly, some expenses used in the computation of the variable fee may be subject to limitations or may not be deductible for purposes of income tax. Conversely, there may be other expenses, for example, a management fee that the taxpayer might claim even though they may not have been used in the buildup of the variable fee. Finally, expenses may differ due to capitalization practices. Capital costs used to calculate the access fee (if negotiated in advance) will differ from actual capital costs incurred, and depending on the particular capitalization and depreciation rules used, the return to capital to may be spread through time quite differently than the timing of returns implicit in the financial model used to build up the access fee. In addition to these complexities, many of the deductible operating costs or those costs embodied in capitalized assets will be incurred in transactions, with affiliated entities raising transfer price issues of their own.23 In short, net income implied by the transfer fee and taxable income may differ significantly.

Therefore even if one were to apply the cost-plus approach to derive the project’s revenues, there still remains much uncertainty and administrative complexity in arriving at taxable income and tax liabilities, as the revenues are only the first element of the calculation.

2.5 Illustrating the impact on tax revenue

To further illustrate the issues set out in the prior section, we have put together a simple model to show the effect that the choices of the various parameters can have on the computed transfer prices, the access and variable fee, and in turn on the income tax revenue of the downstream government. The example assumes a US$1 billion hypothetical rail and port project with a lifetime of 30 years located in a high-risk emerging-markets country where few or no alternative transport modes are available. The income tax computations are simplified but sufficient to illustrate the fundamental points.24

Table 8.2 summarizes how different but plausible ranges of the various input parameters discussed yield a range of values for the WACC and hence the imputed access and variable fees and, ultimately, the taxable income. The main financial results are shown in Figure 8.2.

Figure 8.2Summary of key results

The “low case” in which each parameter is chosen to be least favorable to the project EBIT yields a WACC of 9 percent and the lowest taxable income and tax revenue. Meanwhile, the “high case” in which parameters are chosen to be most favorable to the project EBIT yields a WACC of 23 percent and the highest revenue to the government. The “mid case” is an arithmetical average to the two extreme cases. What is remarkable is the wide range of outcomes that results from the different choices of parameters. Across the scenarios, average taxable income ranges from $37 million annually in the low case to $196 million in the high case, yielding tax revenues that range between $11 million and $59 million – a range of more than 500 percent!25 These results and in particular their wide variance should give no one comfort.

3 Derived or negotiated tax liability

Although the OECD has chosen to try to allocate income on the basis of the arm’s length principle, it is not the only possible approach. We now lay out a number of alternative approaches that may prove more appropriate than the cost-plus approach in some cases.

The first is what we term a negotiated tax liability. It would be possible using the model to directly negotiate tax revenues as a function of capital investment and operating costs, simplifying administration and reducing uncertainty about future revenue flows.

The first step would be to use a simple economic model (such as that used in the computations of Table 8.2) to calculate the annual access fee and the variable fee. The second step would be to calculate the tax liability in any year as the sum of three elements: (1) the tax on the imputed return on equity assumed in the calculation of the access fee, (2) the tax rate times the markup embodied in the variable fee, and (3) the tax rate times the net miscellaneous income from any activity not included in the core infrastructure functions subject to the access or variable fees.

For the first element, one could calculate a levelized access fee without adjusting for tax on equity capital, that is, using Ke without adjustment for tax. The difference between the fee so calculated and the higher annual access fee calculated using Ke/(1 − Tac) would be the amount of the annual tax payment.26 This still requires the parties to have agreement on all the variables determining WACC, including Ke, the permitted debt/equity ratio, and the cost of debt. The second element, the variable component, would be simply equal to the tax rate times the agreed mark up on allowable costs. As before, agreement would have to be reached on what costs would be used to determine the base for the markup. Finally, it would be necessary to take account of any incidental income of the project or income earned from third-party use of the infrastructure.27 Operating cost deductions against the latter should only be allowed for costs not included in the deductions subject to the variable fee markup and in any case would have to be limited to costs that would not have been incurred except for such incidental activity. Losses from the third component should not in principle be allowable against the income tax liability attributable with respect to the fixed and variable streams.28

While this suggested approach suffers from the same computational uncertainties inherent in determining the access fee and the variable fee under the cost-plus approach, it avoids the additional uncertainties related to the difference between gross income and taxable income and would provide much more certainty in administration.

Alternatively, in light of the wide range in estimated tax liability coming out of the model, authorities might consider simply negotiating a fixed annual amount payable to the government in lieu of income tax. The payment might also include any transit fee that the government imposes. This may seem primitive and inconsistent with the notion of an income tax, but it may be more honest to the situation and could greatly simplify tax administration for the transit jurisdiction. Negotiating the fee would require some considerable sophistication on the government’s part.

A negotiated income tax or income tax plus transit fee could also present the mine jurisdiction with some interesting tax alternatives. Rather than calculate a transfer price between Mineco and Transco, the mine jurisdiction could (if agreed) treat the project as a unitary entity, calculating taxable income for the project as a whole, allowing deduction of expenses incurred in both jurisdictions. The tax payment made to the transit jurisdiction could be treated either as a deduction in calculating the tax owed to the mine jurisdiction or perhaps more properly as a credit against the tax due to the mine jurisdiction to the extent that the negotiated tax reasonably approximates the income tax that would have otherwise been due.29 The tax-and-credit approach is obviously analogous to the system employed in which a government taxes residents on world-wide income but allows full or partial credit for income tax paid on activity outside of the taxing jurisdiction.

Such a system would have interesting risk-sharing properties between the mine and transit jurisdictions. The transit jurisdiction would have a relatively certain and secure tax stream, while that of the mine jurisdiction would remain more volatile, reflecting both market volatility and cost risks. On the other hand, the mine jurisdiction would capture the full upside of more profitable projects and in particular would not be subject to any indirect sharing of rents or resource payments with the transit jurisdiction through the tax system. Despite these intriguing properties, the authors know of no examples of such schemes being applied in practice.

4 Formulary system

A second alternative for allocating income for income tax purposes is to use a formulary system to allocate the income of the combined enterprise among tax jurisdictions (or to the jurisdiction utilizing the system) on the basis of various items such as the percentage of property, payroll, or sales of the total enterprise within the tax jurisdiction. The formulary system is used within the United States30 to allocate enterprise income to particular states and is being advocated by the European Commission for use in Europe.31 Even the OECD has recognized the formulary approach in the use of the split profit for highly integrated activities, although it still purports to do so within the context of the arm’s length principle (discussed earlier).

To apply a formulary system, one would treat the entire enterprise (mine, rail, and port) as a single entity, just as if it were in a single jurisdiction. One would then compute the taxable income for the entity using either common rules agreed to by the two governments and the investor or using the particular rules of the tax jurisdiction applying the system. Once taxable income was so calculated, one would apply the formula to allocate it among the two jurisdictions.

In theory this could be done on a unilateral basis by either the transit or mine jurisdiction computing taxable income before allocation on the basis of its own tax system and using its own system for allocation, just as different states in the United States may apply different allocation formulas and calculate enterprise income differently. Alternatively, the computation could be done in a threeparty agreement among the investor and both jurisdictions. The latter would have the advantage of appearing to tax all but not more than all of the income, avoiding no taxation or double taxation. But it also makes matters considerably more complicated since it is necessary to have agreement among the two jurisdictions concerning both the rules for calculation of taxable income and the formula for allocating income.

Among factors usually considered for allocation – assets, labor (payroll), sales – only assets and possibly labor would seem relevant, since sales are being made to persons in third-party jurisdictions. Using assets requires addressing many definitional issues. Assets might be limited to fixed tangible assets (including rolling stock). Intangible assets, including particularly the value of the mine concession itself, are much more problematic. Fixed assets would be assigned to the jurisdiction in which they are located. Assets shared by the two jurisdictions, such as rolling stock, would have to be allocated among operations in the two jurisdictions, perhaps simply on the basis of mileage, but more complex factors could be imagined.32 Valuation would be further affected over time by the depreciation rules used, for example, economic depreciation per financial statements, tax depreciation (particularly distorted because of accelerated depreciation of expensing), or some agreed common rules.33 The purpose of using assets is to allocate taxable income not to determine it. Given this, book values (original cost less economic depreciation) may be the best among the measures readily available.

An especially knotty problem may exist in the taxation of gains when an interest in the project is sold directly or indirectly.34 This is a problem no matter what system is being used for the allocation of income, but if the two jurisdictions are employing some sort of formulary system for the division of income tax revenues, arguably the same formula should apply, since the gain is presumably attributable to some acceleration of income. On the other hand, the mine jurisdiction and the seller, depending upon the circumstances, may argue that the larger part of the value is attributable to the capitalized value of the resources, which would not be reflected in the invested capital.

Although dependent upon the particular facts and the definitions used, a formulary system and the allocation on the basis of assets in many instances could lead to the transit jurisdiction having higher income tax revenue than the mine jurisdiction, as the capital investment in rail and port will often exceed the investment in the mine. For instance, the investment in the downstream infrastructure for the Rio Tinto Simandou project in Guinea may be several times the value the investment in the mine itself. This somewhat uncomfortable result may be better understood and accepted by noting that the mine jurisdiction should also be recovering a royalty to capture the value of its resources. Nevertheless, one problem with utilizing a formulary system even if all of the parameters can be specified is that “taxable income” of the enterprise is likely to contain some element of rent or payment for the resource value. The interaction of the royalty and income tax streams and the capture of the resource value by the host country is discussed further in what follows.

Using a formulary system for allocating income between the two jurisdictions does not avoid the normal problems in determining taxable income. For the project as a whole, the issues of debt/equity ratios, cost of debt, and transfer prices all remain. Even the values of assets used to allocate income between the jurisdictions (if assets are used) are the result of capitalizing expenditures for services, equipment, and goods, some portion of which are acquired from affiliated entities (e.g., rails made from steel acquired from affiliated steel producer, as is the case of ArcelorMittal in Liberia).35

To avoid or mitigate these larger problems would require the application of the formulary system to the parent itself, including downstream activity and other segments not being part of the mine and its rail and port infrastructure. This larger application of the formulary system has some intriguing possibilities, but it raises a whole other level of discourse, including the need in a mining jurisdiction to have any formulary system for income tax coupled to a strong and independent royalty.

Although application of the formulary system requires careful definitions and faces all of the problems associated with taxation of an investment located in a single jurisdiction, it does not depend upon a set of values as widely dispersed as the estimates for the parameters used in the cost-plus method, and in that important particular is a superior method of calculation.

5 Royalties and transit fees

The interactions of the two jurisdictions are more complicated than simply determining and then allocating taxable income among them. The mining jurisdiction also expects and is entitled to a royalty for payment of the natural resource itself. In theory one can separate the royalty payment attributable to the “rent” element or the value of the resource as a factor of production from the income tax or other payments. Indeed, there ordinarily would be a royalty payment whether one or two jurisdictions are involved.

In principle the royalty payment should represent the rent or resource payment so that net income after subtracting the royalty payment would only be the normal return to capital. But royalties are at best an approximation and cannot be expected in practice to correspond to anything but a rough value of the rents or resource value. Thus some component of the income may be attributable to rent or the resource value, especially for highly profitable mines or during “high” price periods. Of course, the opposite can also be true especially during “low” price periods, and perhaps one could argue for their equaling out roughly. Still, proposals such as resource rent taxes, variable royalties, or other mechanisms to capture additional revenue during high-profit periods strongly suggest the belief that income net of royalty will often contain a component of rent. This belief is further supported by the wide range of ad valorem royalties that we see in practice.

The use of a resource rent tax36 in the mine jurisdiction in lieu of or in addition to an ad valorem royalty attenuates but does not fully eliminate this problem. In application a resource rent tax should only reach profits in excess of those necessary to provide a market return on the deployed capital. In that sense the tax is properly targeted only on the rent or resource value, although determination of the threshold used to divide the normal return to capital from the return to the resource (rent) in practice is more a matter of art or convention. Unless the tax captures the full rent or value, any profits above the threshold will also be subject to income tax, and through the income tax both jurisdictions will be capturing part of the rent or resource value.

The royalty problem has an analogue in the transit jurisdiction and the imposition of a “transit” fee. Transit fees, not to be confused with tariffs paid by shippers or the access and variable fees used to set a transfer price, are not universal, but in our limited experience in West Africa, both governments and investors expect the payment of such fee. Like a royalty, the higher the transit fee, the smaller the income for the project as a whole. In addition, when the royalty is calculated on the mine mouth value or other point in the mine jurisdiction, the transit fee will reduce the royalty payment. Thus a transit fee, like a royalty, can (against any agreed split and treatment of income tax) shift tax revenue from one jurisdiction to the other. Transit fees and examples are discussed in more detail in Chapter 9 of this volume at section 8 (Le Leuch).

In the first instance the transit jurisdiction may appropriately charge for leasing the right of way and land for the rail and port facilities. There is, however, a more subtle factor that also may be taken into account in computing a transit fee. The in-place value of the resource depends upon expected future market prices and the nature of the deposit (e.g., grade, impurities, or geology). Yet another part of the determination of the in-place value of the resource is locational, the cost of moving it to markets, and a principal component of that is its proximity to the sea. In the single-jurisdiction case, one does not need to take too refined an approach to the matter since the royalty, land rentals, and to some extent other taxes capture both the nature of the deposit and the location. There is a single “source” country. However, value determination is more complicated when two jurisdictions are involved.37

The importance of location can be illustrated by a case in which the mine jurisdiction does have an outlet to the sea but when shipping through a neighboring country is cheaper. This is the case for mining deposits located in south-east Guinea near the Liberia border, where shipment through Liberia (in part through established transport corridors) could be much cheaper. Some estimates indicated savings in the neighborhood of 40 percent in capital costs and 40 percent in operating costs for the rail and port.38 Where such an alternative in-country route exists, cost comparisons (including fiscal burdens) of using the in-country (or third-country) versus alternative route allow one to estimate an upper limit on location value of the alternative route.39 Where the mine country is landlocked and there is only one feasible exit country, the upper limit on the transit charge is the amount taking account of the royalty and other charges of the two governments that would otherwise make the project uneconomic (i.e., that would drive the return below the minimum rate of return on capital investment or, more realistically, the next best call on the resources of the potential investor).

These economics, which are due to simple monopoly power,40 do not imply that the transit jurisdiction should capture all the “location” value any more than that the mining jurisdiction should expect to capture through royalty all of the theoretical rent or resource value. However, they do illustrate that location affects project economics just as the characteristics of the deposits and that the transit jurisdiction does have some equitable claims to share in the overall project benefits. Geology or geography, both create claims, and within some range there is no principled approach to determining the allocation. Some view rail and port transport as simply a cost element, but extraction is also simply a cost item. Labels alone do not solve the problem.

6 Effect of third-party interests

Both tax and structure issues may be affected by third-party ownership or use of the infrastructure. There is a strong interest in utilizing infrastructure established for mining in broader economic development, although actual success in doing so has been limited.41 The efforts to do so have focused on either requirements for third-party access or separation of operation and ownership of the assets. Any effective access for third parties will generate additional revenues and costs. In some cases, for example, local passenger service, the service may be more a condition of getting the concession or right of way and certainly not intended to be profit making. In others the host government may attempt to force third-party access for other resource developments within either the transit or mine jurisdictions.42 Establishing third-party charges or tariffs may require the government and other parties to address in the context of the cost of service and cost recovery many of the same issues and items discussed in the cost-plus section. And while one could speculate on interrelationships, in the end for tax purposes it is probably simply a matter of recognizing and allocating revenues and additional costs when two jurisdictions are involved.

A more complicated situation arises when third-parties have ownership or operational rights to the infrastructure, especially when ownership in part or whole is vested in the host government.43 In addressing the matter, one must look at the economic reality and not focus on labels. Where the investor is responsible for capital costs and operating expenses and has full control over operations, for tax purposes “ownership” of parts of the infrastructure may not matter. Two other cases are of interest, however. In the first case the rail and port infrastructure is held by two or more independent investors, only one of which is the operator. If the ownership interests are not adjusted to reflect use, then the pricing for transport and port services is likely to have significant economic substance, and the prices will no longer be “transfer” prices. To the extent that the transport entity covers more than one jurisdiction, there would still be the issue of allocation but in a simpler form.

The second case, a variant of particular interest, is when the government holds an interest.44 The government then has a dual interest in the transport charges, which will determine both the value of its equity, as well as the size of its tax base. This must lead to a negotiated agreement about rates, a debate that may center on the same elements discussed earlier with respect to the use of cost-plus method for transfer pricing, at the risk of repetition, a debate for which governments may be ill prepared.

7 Conflicting interests

Most of the discussion has been framed as a negotiation between the transit taxing authority and the investor. But obviously the upstream taxing authority, the mine jurisdiction, has also a strong interest in the imputed charges. The investor will seek to deduct from its income in the mining jurisdiction the costs incurred in transportation in the transit jurisdiction. If both jurisdictions in fact accept the same calculation, higher transfer prices will shift income and tax revenue to the downstream jurisdiction; conversely, lower prices will shift income and tax revenue away.

The project as a whole has a single rent, and the project’s realization is dependent upon the cooperation and assent of both jurisdictions. Subject to the investor being satisfied, with the post-tax return for the project as a whole, the tax issues shift to a debate between the two governments with a focus on three items: the royalty regime of the mine jurisdiction, any transit charge by the transit jurisdiction, and the division of the income tax burden. The governments are not unconstrained in their bargaining. Royalty rates, tax rates, and the definition of taxable income may already be set by statute and subject to limited variation. Moreover, governments are subject to investment treaties and other legislation requiring them to act on a non-discriminatory basis or to observe other standards.

Although there is an assumed view that symmetrical treatment is required to avoid over- or undertaxation, that is not strictly speaking necessary given that many other factors unique to each tax system will affect the actual tax liability of the whole enterprise, for example, tax rates, allowable expenses, capitalization and depreciation rules. In any case, both jurisdictions have an interest in the total tax liabilities of the enterprise given the impact on the investment as a whole, and both jurisdictions have a common interest in the integrity of cost and revenue reporting, especially when transactions with affiliated entities are involved, as well as other necessary coordination.

8 Conclusion

The unitary nature of the mine and its cross-border infrastructure, each fully dependent upon the other, means that any exercise to separate the two into independent entities for tax purposes has a high degree of arbitrariness. The wide range of possible results is illustrated by our analysis of the cost-plus method and the assumptions required for its application. Notwithstanding the OECD‘s commitment to the arms-length principle, a better approach is to recognize the unitary nature of the project and to apply a profit-split or formulary allocation to apportion income between jurisdictions. The success of this approach requires as an ancillary matter that the mining jurisdiction apply a strong royalty or resource rent tax recognized in the computations of both jurisdictions. All of the traditional issues for determining taxable income, including the possibility of different calculations by each jurisdiction, remain.

References

    Association for Financial Professionals. (2011) Current Trends in Estimating and Applying the Cost of Capital.

    DaZ.R.J.Guo and R.Jagannathan. (2012) “CAPM for Estimating the Cost of Equity Capital: Interpreting the Empirical EvidenceJournal of Financial Economics103 (1) 204220.

    FamaE.F. and K.R.French. (1997) “Industry Costs of EquityJournal of Financial Economics43 (2) 153193.

    FrankM.Z. and T.Shen. (2012) Investment Q and the Weighted Average Cost of Capital Available at SSRN 2014367 (2014).

    HellersteinJ.R.W.Hellerstein and John A.Swain. (2012) State Taxation (Warren Gorham Lamont).

    LandBryan (2010) Resource Rent Taxes: A Re-Appraisal in DanielPhilipMichaelKeen and CharlesMcPherson (eds) The Taxation of Petroleum and Minerals: Principles Problems and Practice (RoutledgeLondon and New York) pp. 241262.

    Organisation for Economic Co-operation and Development. (2010) OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations.

    OttoJ.CraigAndrewsFredCawoodMichaelDoggettPietroGujFrankStermoleJohnStermole and JohnTilton. (2006) Mining Royalties: A Global Study of their Impact on Investors Government and Civil Society (Washington, DC: World Bank Publications).

    StrongJ. (2004) “The Development of Railway Concessions in West and Central AfricaThe Journal of Structured Finance9 (4) 6691.

    ToledanoP. (2012) Leveraging Extractive Industry Infrastructure Investments for Broad Economic Development Available at http://academiccommons.columbia.edu/catalog/ac:154081.

Notes

Although we use the labels “transit jurisdiction” and “mine jurisdiction,” it is important to remember that part of the transportation infrastructure lies in the mine jurisdiction, a factor that is particularly important if consideration is given to treating the transportation infrastructure as a separate entity.

Compare to the role of the host government agreements for multi-jurisdiction pipelines discussed in Le Leuch (2016), Chapter 9 in this volume.

For a survey of rail and port operations serving mining, see Toledano (2012). For experience in West Africa, see Strong (2004).

OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, OECD (July 2010), p. 26.

Ibid., p. 124.

Such a clause could also be a way to preserve income to the transit jurisdiction even if actual operations are suspended temporarily or permanently.

For more discussion on tariff rate setting, see Le Leuch (2016), Chapter 9 in this volume.

For this purpose, the project’s lifetime is the number of years during which the investor will own and operate the infrastructure. In many projects this period is specified in the contract up front; thereafter, ownership of the infrastructure often transfers to the government. To the extent that this period is renegotiated ex post, the access fee may need to be adjusted.

In the discussion that follows, we will assume annual discounting, such that the discount rate r is an annualized discount rate, giving rise to a per-annum access fee.

Actual capital expenditures incurred prior to the date that the access fee becomes effective would have to be adjusted upward using the agreed discount rate, r.

“Current trends in estimating and applying the cost of capital,” Association for Financial Professionals, 2011. In another illustration of this conundrum, a recent working paper constructs 440 different ways of calculating WACC based on 11 different measures of the cost of equity, 4 measures of the tax rate, 2 measures of the cost of debt, and 5 measures of the leverage ratio (Frank and Shen, 2012).

See http://pages.stern.nyu.edu/~adamodar/. Data from the same source suggests an average debt-to-capital ratio for “Metals & Mining” companies in emerging markets of 20.8 percent and for “Transportation” companies of 27.6 percent. Debt ratios are lower for industries in regions with less developed capital markets, such as the “Africa and Middle East” category.

See discussion in Shay (2016), Chapter 3 in this volume, regarding the problems in determining the appropriate treatment of debt when the debt is from or directly or indirectly guaranteed or supported by a related entity.

Other methods for calculating the cost of equity include the Fama-French three-factor model and the Carhart four-factor model, both of which are extensions of the CAPM that allow for more than one beta parameter. In addition, there is a class of equity valuation models such as the Gordon growth model that take the value of equity to be the discounted sum of expected future cash flows (dividends). In addition, numerous studies have found that various other firm characteristics not included in CAPM (such as firm size, the earnings-to-price ratio, the book-to-market value of equity, and others) also have power to predict equity returns (Da et al., 2012). Despite these challenges to the CAPM, it remains the most widely used model for calculating cost of equity in practice. Da et al. cite that about 75 percent of finance professors recommend using the CAPM to estimate the cost of capital for budgeting and that a survey of chief financial officers indicates that 74 percent of the respondents use the CAPM. They also argue that despite some problems with the use of the CAPM in estimating cross-section stock returns, it remains useful in estimating the cost of capital for projects.

For example, one paper quotes an annualized average market premium for 1963–1994 of 5.16 percent per year with a standard error of 2.71 percent (Fama and French, 1997). This standard error implies that the risk premium for a project with a beta equal to 1.0 could be anywhere from 2.45 percent to 7.87 percent.

Based on data compiled by NYU finance Professor Aswath Damodoran, the average beta of emerging markets metals and mining companies is 1.4, compared to 0.9 for transportation companies and 0.8 for railroad companies (data as of January 2013, available at http://pages.stern.nyu.edu/~adamodar/).

We note but do not analyze the interesting case in which the investment is in passthrough form so that the only tax is imposed at the level of the investor.

Despite recent controversies surrounding the accuracy and possible misrepresentation of the LIBOR rate in financial markets, it remains the widely used benchmark for pricing securities. Proposals to increase regulatory oversight of the LIBOR determination process are underway.

See Shay (2016), Chapter 3 in this volume, for a fuller discussion of transfer pricing issues.

See discussion in Section 4; also in Le Leuch (2016), Chapter 9 in this volume.

Compare a markup permitted to an operator for indirect costs in the context of joint operating agreements, commonly used in the petroleum sector between partners to a joint unincorporated venture to allocate responsibilities and liability.

This conclusion assumes that any allowable deduction for management fees would be less than the permitted markup.

See Shay (2016), Chapter 3 in this volume.

For simplicity of illustration, the calculations assume the project investment of US$1 billion occurs in one period, followed by a constant annual maintenance capital investment of US$50 million, fixed volumes with a simple estimate of operating expenses (of US$100 million annually), 30-year straight-line depreciation of capex and 15-year depreciation of rolling stock, full deductibility of interest, and taxable income equal to financial income less interest deduction (i.e., no loss carry-overs and other differentials).

All calculations presented here are in real terms. The actual amounts payable would be subject to inflation adjustment.

Again, because all calculations are done in real terms, the amount actually payable would be subject to an inflation adjustment.

An implicit assumption is that the project is “ring-fenced” for tax purposes.

Although the negotiated formulas and amounts would in effect be an effort to negotiate net income and the tax structure, we do not attempt to address arguments as to whether such a tax could be credited in a resident jurisdiction as a tax on net income.

Conceptually, any portion of the payment due as a transit fee should be a deduction, and the portion due as a payment for income taxes should be a credit. The size of the allowable deduction by the mine jurisdiction of the income tax payment to the transit jurisdiction would also require consideration of the implicit tax rate and some notion of the reasonableness of the imputed income in the transit jurisdiction. The latter could lead the mine jurisdiction and investor back into the same transfer pricing debate that the negotiated figure was intended to limit.

See generally Hellerstein (2012).

Proposal for a Council Directive on a Common Consolidated Corporate Tax Base, COM(2011) 12¼, 2011/0058 (CNS)(“EU Council Directive”).

In the proposed Nabucco gas pipeline in Europe (now cancelled, possibly to be restricted in simpler form), the parties provided for the allocation of net revenue (gross revenues less original costs) for national tax purposes on the basis of the proportional share of the total length of the pipeline falling in the particular jurisdiction – Article 11.2 of the Agreement Among the State Parties Regarding the Nabucco Project (July 13, 2009) available at http://www.mfa.gov.hu/NR/rdonlyres/8B0D4EA0–8FF1–46C3–8772-E16905FE29E9/0/090714_nabucco_agreement.pdf. Assuming equal costs, this is equivalent to allocation on a net asset basis.

Some of the factors that would need to be considered are identified in the EU Council Directive Articles 92–94.

For a general discussion of the issues associated with the taxation of capital gains, especially where indirect transfers of interest are involved, see Burns, Le Leuch and Sunley (2016), Chapter 7 in this volume.

A variation of the formulary system would be to have the governments and the investor agree to treat the entire project as a stand-alone entity for tax purposes subject to an agreed contractual regime, something like is done with joint development zones in which income is computed and allocated according to some agreed system – for more discussion, see Daniel, Veung and Watson (2016), Chapter 11 in this volume. The governing document could also deal with royalties, duties, transit fees (if preserved), customs, VAT, withholding, tax exemptions, exchange rights, employment, local procurement, and other issues typically dealt with in development agreements. The authors are not aware of any examples.

For the principles and issues affecting design of resource rent taxes, see generally Land (2010).

Within jurisdictions, the application of the royalty rate on the mine mouth basis partially takes account of the location aspect of valuation for royalty purposes. See generally Otto et al. (2006).

Calculations can be much more complicated. For instance, an investor in Guinea offered to construct a light rail to a port in Guinea from its Guinean mine location provided that it could have the right to export its ore across Liberia. The savings from the shorter route through Liberia, including the reduction in capital costs for engines and rolling stock and the reduced operating costs, would have more than offset the additional cost of the light rail across Guinea, but the capital savings is significantly reduced.

There is an intermediate case in which a mine in a country whether landlocked or not may have the possibility of exporting through either of several neighbors, for example, Guinean ore could move to the sea through Ivory Coast or Liberia. The potential location rent is then limited to the cost of the next best alternative.

Control of the monopoly power of the transit state and the right of landlocked states to access transportation has long been a subject of international concerns. For a brief summary of relevant international conventions, see Le Leuch (2016), Chapter 9 in this volume.

For a full discussion of the issue and impediments in practice, see Toledano (2012).

Even in economically advanced jurisdictions, the successful requirement of third-party access has been difficult. There are many operational problems with expanding local service, for example, high-density unit train traffic may not be readily compatible with more localized service that also has to be addressed but are outside of the scope of this discussion.

See Toledano (2012), Table at fn. 14 for a sample of ownership and operational arrangements.

The Simandou project in Guinea has such a structure. Rio Tinto holds the mining concession, but the rail and port infrastructure is to be held by a joint venture of Rio Tinto and the government.

    Other Resources Citing This Publication