Selected Issues


Selected Issues

Issues in International Taxation1

A. Executive Summary

1. The Uganda Ministry of Finance, Planning and Economic Development requested that the IMF Fiscal Affairs Department analyze issues in cross border corporate income taxation that may be contributing to the relatively low revenue performance of the Uganda corporate income tax. Although Uganda’s statutory corporate tax rate is the same as that of Kenya and Tanzania, for example, the percent of GDP raised by the tax is only about half of that in Tanzania, and less than a third of that in Kenya. Nearly 45 percent of corporate tax revenue in Uganda comes from fewer than thirty major corporations in the main economic sectors receiving the most foreign direct investment, other than mining and extractive industries 2—and the great majority of those companies are subsidiaries of multi-national enterprises. Thus, the structure of the Ugandan economy lends itself to revenue loss through cross-border tax planning and avoidance techniques.

2. The report focuses on three broad potential sources of corporate tax revenue leakage. These include: (i) the widespread use of tax incentives and exemptions granted in an attempt to increase foreign direct investment; (ii) aspects of the domestic tax code (ITA) that can facilitate cross border tax planning; and (iii) tax avoidance using Uganda’s network of double tax treaties (DTTs).

3. Uganda—in common with most developing countries—faces strong pressures to grant tax incentives and exemptions to attract investment, which should be resisted. These pressures arise from requests from potential investors, but are exacerbated by competition for investment among countries. Much evidence supports a view that such incentives are largely unnecessary—in that taxes are not among the most important factors in attracting foreign investment—and often are entirely redundant—that is, companies report that the investments would have been made without these tax incentives. Redundancy is particularly likely when the investment in question is made to produce for the local domestic market, or when the investment entails exploiting other location specific economic rents—as in the case of extractive industries, or telecommunications. And some types of investment tax incentives are more effective than others: tax holidays and reduced rates on ex post profits counterproductively reward most the most profitable (successful) firms, which likely need this benefit least, and do not incentivize actual new investments. Incentives for new investment are much better achieved by accelerated depreciation, and/or initial allowances that are specific to new investment. Uganda already has these latter, more effective, types of incentives in its basic tax system.

4. Cross-border tax avoidance possibilities arise through several means, including interest stripping, payment of management and service fees, and abusive transfer pricing:

  • The Ugandan ITA includes a “thin-capitalization” rule, restricting the deduction of interest payments where a company’s debt to equity ratio exceeds 1.5:1. The rule is seriously undermined, however, by a provision that provides that the restriction will not apply if the loans in question are on terms that would have been set between unrelated parties; and analysis in this report finds that for most large companies the ratio is not binding. The report recommends that Uganda replace or supplement this rule with a so-called “interest stripping rule,” which restricts the amount of interest payments that can be deducted relative to a before-tax, depreciation and interest measure of gross profits.

  • Payment of management and service fees to parent companies—a common tax planning device—is reportedly a major source of tax base leakage in Uganda. The ITA appropriately provides that such fees are subject to withholding in Uganda. However, this rule is largely rendered ineffective by the fact that most of Uganda’s existing DTTs do not include a provision permitting this withholding. Uganda should try to include a withholding provision for such services in its existing treaties—when and if these can be renegotiated—and insist on its inclusion in any new treaties.

  • Uganda has taken steps to reduce revenue loss through abusive transfer pricing, including by the introduction of rules based upon the OECD transfer pricing guidelines, the undertaking of training for staff of the URA, and the commencement of several complex transfer pricing cases. Going forward, Uganda should seek additional information to allow it to improve risk assessment for transfer pricing audits, through (i) requiring mandatory routine production of the documentation now available only on request; and (ii) ensuring that Uganda satisfies all the criteria (discussed in the report) which will allow it to obtain information newly required under the country-by-country reporting requirements of the BEPS project.

5. Uganda’s double taxation treaty network presents the opportunity for tax reduction by foreign investors. While any DTT will—by design, though it is advisable to minimize this—lead to some reduction in a claim on the tax base by the source country, treaties are also frequently used in tax planning by multi-national corporations to achieve “double non-taxation.” This is done by routing investments into source countries through intermediate companies in treaty countries that permit no, or very low, taxation on the income of those intermediary companies. Best practice anti-abuse provisions in treaties can help to minimize—if not eliminate—this, and should be adopted by Uganda in all of its treaties to the extent possible. The Ugandan Cabinet has recently approved a model treaty and treaty policy, which largely represent modern best practice for source countries seeking to enter into DTTs. It will be highly advisable for the Ugandan government to follow the model provisions in negotiating all new treaties. To the extent that treaty partners can be induced to enter into re-negotiation of some treaty provisions, this too should be undertaken. Uganda will also want to consider whether to enter into the new Multi-lateral Instrument (MLI) to adjust the provisions of its existing treaties. Whether this will be helpful—and not all desirable changes are covered in the MLI—will depend upon the positions already taken in the MLI by all of Uganda’s separate treaty partners. This needs to be carefully assessed in each case.

6. Finally, the report addresses an important issue that crosses over Uganda’s domestic tax law, and its DTTs: indirect transfers of interest in Ugandan assets that take place offshore. Recent amendments to the ITA include provisions that would tax in Uganda capital gains arising on such transfers where the underlying assets constitute “immovable property,” including mineral and petroleum rights. However, these provisions are not effective where transfers take place in a treaty country—as Uganda’s treaties do not now include a provision that would permit Uganda to assert the right to tax this base. Where possible, these provisions should be renegotiated, as any important such transfer would be structured through a country with such a treaty. Further, it would be advisable to extend in domestic law and in treaties the meaning of “immovable property,” to encompass other rights that give rise to economic rents based on a location in Uganda—for example, licenses and rights to domestic telecommunication assets.

B. Introduction

1. The Uganda Ministry of Finance, Planning and Economic Development requested that the IMF Fiscal Affairs Department analyze issues in cross border corporate income taxation that may be contributing to the relatively poor performance of the corporate income tax. At 13.3 percent of GDP in FY15/16,3 Uganda’s overall tax revenue collection remains at the lower end of regional peers—albeit representing a considerable improvement from the 10.8 percent tax ratio in 2013/14—and the authorities remain committed to increasing the tax-to-GDP ratio by ½ percent of GDP per year over the medium term, an objective that is supported by the IMF policy support instrument (PSI) and technical assistance (TA). As of 2015, the performance of the corporate income tax was still low by regional standards, shown in Table 1.

Table 1.

Uganda: Regional Corporate Income Taxes (Uganda 2015)

article image

Data refers to the latest year available. Efficiency ratio is defined as the CIT revenue ratio divided by the standard CIT Rate.

Source: IMF World and staff calculations

2. The analysis in this report points to three broad areas that potentially contribute to this deficiency: widespread use of tax incentives/exemptions for many sectors and firms; tax base leakage through tax planning using various cross-border income stripping techniques by multi-national enterprises (MNEs); and weaknesses in the Uganda double tax treaty (DTT) network. The URA recently estimated revenue foregone through certain tax expenditures—a positive step—but has thus far been unable to estimate the amount of revenue being lost as a result of cross border international tax planning. Given the limitations on data and time, it is also not possible for the mission team to estimate with reasonable accuracy the potential tax losses arising from tax avoidance through weaknesses in the provisions of the Income Tax Act (ITA), or through the use and abuse of tax treaties. However, the structure of investment and sources of tax revenue make clear that ample opportunity for cross border tax planning exists.

3. Uganda remains a “source” country for international tax purposes. Data provided from the URA with regard to the activities of large taxpayers across the major foreign direct investment (FDI) receiving sectors of the economy indicate that the vast majority of the largest taxpayers in those sectors (finance, agriculture, manufacturing, information and telecommunications, and wholesale and retail trade) are subsidiaries of MNEs—thus giving rise to considerable opportunity for cross border tax planning. And the stock of inbound FDI as a percentage of GDP has increased markedly over the past 10 years, as shown in Figure 1, indicating that Uganda’s economy is becoming more open to cross border capital flows.4 Uganda still has very little outbound FDI. Uganda is not unique among EAC countries as being a large net capital importer. For example, Kenya has also attracted increasing net FDI in recent years, and the growth in its total inbound FDI stock is still outpacing that of its outbound FDI.

Figure 1.
Figure 1.

Total Inbound and Outbound FDI Stock (as Percentage of GDP)

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A005

Source: IMF WoRLD and staff calculations

4. Nearly three-fourths of CIT revenue is derived from the five sectors that are the largest recipients of FDI,5 as shown in Figure 2.

Figure 2.
Figure 2.

Share of CIT Revenue for Top Industries receiving Inbound FDI, 2014

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A005

Sectoral CIT Revenue from LTU and MTU only; latest year for sectoral CIT breakdown is FY 2014-15.Source: UGA and IMF staff calculations

The average tax rate relative to gross profits (ATR) for the largest companies in these five principal economic sectors (29 unique companies, 89 separate observations) was less than 10 percent during FY 2012/13 – 2015/16.6 Figure 3 summarizes the ATR as a percentage of gross profits for this group of companies by industry (left blue bar). The ATR ranged from 3.4 percent in Agriculture to 9.1 percent in Manufacturing during the sample period.7

Figure 3.
Figure 3.

Average CIT Ratios for Selected Companies Named in Sectors, 2012-13 - 2015-16

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A005

Source: URA and IMF staff calculations.

As shown another way in Figure 4, there is a considerable gap between the gross and chargeable profits in all five of these key economic sectors in Uganda. While it is to be expected that gross profit would exceed taxable income, given normal deductible expenses, variations and such large differences can also reflect earnings stripping through tax avoidance. Wholesale and retail trade and agriculture exhibit lower gross profit margins compared to other industries. Other industries in the data reviewed also demonstrate on an aggregate basis a considerable gap between their average gross profit and the taxable profit.

Figure 4.
Figure 4.

Average Gross Profit Margin, 2012/13 - 2015/16

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A005

Source: URA and IMF staff calculations.

5. The Uganda authorities are well aware of the potential tax leakage that can result from cross border activity, and have taken several positive steps to try to combat this. Major improvements in this regard in the last few years include introduction of more detailed transfer pricing rules (and on the administrative side, the bringing of several major transfer pricing cases); the very recent adoption by cabinet of a new model DTT for Uganda, along with treaty policy guidance for government; and the adoption of new provisions governing the taxation of mineral and petroleum extraction.

6. This report identifies several areas where the law could be further tightened to better protect Uganda’s tax base from cross border base erosion and profit shifting. These include interest stripping provisions, tighter protections against excessive revenue loss from nonresident management and service fee charges, provisions to protect against transfers of interest in Ugandan assets without the payment of tax to Uganda, and ways to reduce inappropriate treaty reliefs and treaty abuse.

7. Given the current broad tax exemptions and holidays, and the low effective tax rates on inbound foreign investment even aside from these holidays, it would be unwise to consider reducing the statutory corporate tax rate at this time. The 30 percent statutory rate is in line with that of all neighboring EAC countries. Reducing it under these circumstances would result in a windfall to existing investment, both domestic and foreign, giving up much needed tax revenue for little or no spur to economic growth. Further, such a move would likely give rise to yet more damaging regional tax competition, at a time when tax revenue is sorely needed for investment in human capital and infrastructure.

C. Tax Incentives and Exemptions

8. The Ugandan Government faces a tradeoff between increasing the ratio of tax revenues to GDP and giving tax exemptions and incentives to potential investors in an attempt to create economic growth and, especially, to create new jobs. Pressure to award such tax exemptions seems, if anything, to be growing—yet based upon the government’s own estimates, the revenue cost of such incentives across all taxes and types constitutes 8-10 percent of tax revenue. Uganda’s tax ratio now stands at just over 13 percent, still below the regional average of 14 percent and clearly insufficient to meet the country’s needs for investment in human capital and physical infrastructure. Recent Fiscal Affairs Department analytic work finds that revenue ratios of less than about 15 percent of GDP are adversely correlated with economic growth8.

9. Uganda is not alone in facing these pressures. While all countries face them to some degree, they are particularly acute in developing countries. They generally seek to attract foreign investment to contribute (it is hoped) to economic growth—and they tend to suffer from a perceived lack of bargaining power relative to the MNEs by whom they are implored to grant exemptions. The problem is sufficiently widespread that the IMF, together with the World Bank, OECD and UN, were asked by the G20 to write a report on how such tax incentives could be made more “efficient and effective” (and less damaging to the revenue base) in lower income countries.9

10. Several lessons can be drawn from this report, and from the sources on which it in turn draws. These concern (i) what type of incentives are more effective and less damaging to the revenue base; (ii) how much are the tax benefits really needed in order to obtain investment in any given case; (iii) the need to measure the benefits hoped for, in a coherent way—and to assure that they will actually be produced; and (iv) the potential benefits of regional coordination.

Are tax incentives really needed?

11. Considerable evidence from surveys of MNEs themselves, including for Uganda, make clear that—whatever may be said in any specific instance—taxes are not generally a decisive factor in deciding whether to invest in a given location. As cited in the report referred to above: “In 2010, the United Nations Industrial Development Organization conducted a business survey of 7,000 companies in 19 sub-Saharan African countries active in agriculture, mining, manufacturing, utilities, construction, and services sectors. Investors were asked to rank the importance of twelve location factors and to assess how they might have changed, improved and worsened, in the last three years. The results suggest that tax incentives packages ranked 11th out of 12 in importance.”

12. Tax incentives are often found to be redundant in attracting investment in developing countries; that is, the same investments would have been undertaken even if no incentives had been provided. Table 2 (drawn from the cited report) shows redundancy ratios, based on investor surveys in various countries, measured by the percentage of investors who claim that they would have invested even without tax incentives. Redundancy levels thus obtained—subject to well-known caveats, such as a discrepancy between answers and actual behaviors under a counterfactual scenario—are high in most countries. For example, redundancy rates exceed 70 percent in 10 out of the 14 surveys listed in Table 2. The redundancy ratio for Uganda was found in this particular study to be 93 percent.

Table 2.

Uganda: Redundancy of Tax Incentives Based on Investor Surveys 1/

article image

Percent of affirmative answers to the question if an incentive was redundant;

51 percent for non-exporting firms outside free zones

13. The ability of MNEs to avoid source country taxation may blunt the impact of tax incentives. Thus the base protection measures discussed in the other sections of this report cannot be fully divorced from the issue of the granting of exemptions and incentives. As discussed later in this report, foreign-controlled groups have structural tax advantages in most CIT systems, even before considering avoidance opportunities, and Uganda’s is no exception. A MNE that can readily reduce Uganda taxes (by, for instance, using related party interest deductions to shift profits to a low tax jurisdiction) is in little need of additional benefit from corporate tax incentives such as holidays and exemptions.

Types of incentives

14. Many low-income countries—like Uganda—use costly tax holidays and income tax exemptions to attract investment, but investment tax credits and accelerated depreciation yield more investment per dollar spent. Uganda already has—appropriately—generous initial allowances and favorable depreciation schedules—as well as unlimited income tax loss carryforwards. These factors reward new investment. Tax holidays and exemptions provide the greatest benefit to the most profitable enterprises—which need them the least. Further, their benefit is derived only after the investment begins to make positive taxable returns. The holiday incentives are much less well targeted to attract actual new investment at the margin.

15. Tax incentives targeted at sectors producing for domestic markets or extractive industries generally have little impact. If companies’ goals are to produce for the Uganda market, or to extract Uganda natural resources, those markets and resources include location specific rents that themselves are a key attraction for the investors to come to Uganda.

The real social benefits to be derived from, and the costs of, the incentives should be carefully measured

16. A tax incentive serves a useful social purpose if the social benefits it actually generates exceed the associated social costs. Investment tax incentives ultimately aim to contribute to a country’s development and improved living conditions for its citizens. A number of elements are critical to determining the social benefits of an incentive. These include: (i) size of the net investment effect—the rise in investment should be corrected for the redundancy effects mentioned above, and displacement effects—the reduction in any other investments—to infer the net incremental increase in capital due to the incentive; (ii) net impact of higher investment on jobs and wages—new jobs can yield significant social gains if they reduce unemployment. However, if new jobs displace existing jobs, the social benefits depend on the productivity (and wage) differential between the new and old jobs; (iii) productivity spillovers—to the extent that new investment boosts productivity elsewhere in the domestic economy, such as in supplying or competing firms this magnifies social benefits by raising income levels more widely.

17. The social costs of tax incentives depend on various factors, and must be offset against the benefits as measured above. These include: (i) net public revenue losses—public revenue falls if tax incentives are redundant or create leakage and abuse. Taxes generated by additional net investment and jobs can be offset against this; (ii) administrative and compliance costs, which can rise due to tax incentives, especially if they are complex or create opportunities for rent seeking and corruption; (iii) scarcity of public funds. Often overlooked is the fact that $1 of tax revenue has a higher social value than $1 of private income, because it is the greater value of the public expenditure it finances that justifies transferring resources from private to public sectors through taxes. To compare changes in private income and tax revenue, the latter thus need to be weighted by the “marginal cost of public funds”, which will be greater than one; (iv) distorted resource allocation. Discrimination in favor of some and against other investment implies that taxes, rather than productivity differences, determine resource allocation.

Tax competition and regional cooperation

18. Demands for tax exemptions and incentives by potential investors lead to harmful tax competition among the countries targeted, particularly within regions. This “race to the bottom” has become more acute in sub-Saharan Africa over the past few decades. Uganda, like many other countries, understandably feels pressure to match the tax incentives granted by neighbors Kenya and Tanzania. Where the investments in question are mobile—such as investments in manufacturing facilities for export, as opposed to the extraction of local minerals, or production for the domestic market—this competition is particularly intense. The best approach to minimize this loss of revenue is to achieve a degree of harmonization and coordination among relevant neighboring countries in regard to the grant of these tax benefits. The EAC provides a vehicle for this coordination, which it would be advisable to further utilize.

19. Tax treaties should not be used as a form of tax incentive. This is discussed further in section IV of the report, below.

D. Three Sources of Tax Base Leakage

Structural Tax Advantages for Foreign-Controlled Businesses

20. It is challenging for every country to tax foreign-controlled groups or nonresident companies carrying on business within their borders. It is generally the case that some value is contributed to the local business of an MNE group from outside as well as inside the country. Capturing the appropriate share of the total value for the source country requires well-tuned cross-border tax rules. Further, as MNEs continuously find new innovative ways of reducing their tax burden, anti-avoidance rules are in constant need of being monitored and strengthened.

21. Statutory-based structural advantages exist for foreign controlled groups. There are clear tax advantages for MNEs to making deductible payments, including interest, royalties, and management fees, to low-taxed nonresident affiliates. This reduces the Ugandan tax base by benefitting from the deduction at a 30 percent rate, while (in the absence of even lower reduced treaty rates) paying tax to Uganda on the transfer at 15 percent withholding. This advantage for foreign-controlled business constitutes an integral part of the current international tax architecture, and is found in almost every country’s tax system. Thus, it would not be recommended to change this (other than to control abusive earnings stripping, as discussed below). However, this structure provides natural tax advantages to MNEs as compared to purely domestic enterprises, and serves to highlight the potential redundancy of providing additional tax exemptions and incentives to FDI. In an otherwise well designed tax system with fair recovery of expenses allowed to businesses, tax incentives to attract foreign investment generally are unnecessary and inappropriate.

22. Countering structural advantages of foreign-controlled groups. These structural tax advantages of foreign-controlled groups in avoiding source (host) country taxation may be mitigated in part only if the foreign-controlled business’s income is fully taxed, within the parameters of this architecture, by the source country, including by taxing capital gains on the exit from the business. They also are a reason why it is important to include in the law robust measures to prevent abuse of intercompany deductible payments for interest and management and service fees and to tax indirect offshore transfers of businesses realizing their value from economic activity in the source country. These issues are discussed in succeeding parts of this report.

Interest stripping and thin capitalization

23. The ITA includes an anti-avoidance rule designed to prevent reduction of the tax base through artificially inflating interest deductions. The present restriction is in the form of a “thin-capitalization” rule (ITA Section 89, as amended in 2015), which restricts interest deductibility based upon the ratio of debt to equity in a foreign controlled company’s capitalization. To the extent that the Ugandan subsidiary has a debt to equity ratio of more than 1.5:1, interest payable with respect to the excess is disallowed as a deduction. The exact form of this rule has varied over the past several years. The allowable ratio has gone from 1:1, to 2:1, and now to 1.5:1—along with changes in the definition of what debt and equity are involved in calculating the ratios.

24. A positive aspect of the current rule is that, in the case of corporations to which it applies (those which are owned more than 50 percent by a non-resident person), the ratio is to be calculated with regard to all debt and equity, not restricted to related party debt. While it is common to apply such rules only to related party borrowing, the current Ugandan approach is much preferable, and represents a more modern view. Such rules targeting all debt are more effective in restricting excessive corporate borrowing; from an administrative standpoint, anti-abuse rules restricted to related party borrowing are easier to avoid through use of related party guarantees of third party debt and back-to-back loan arrangements. Further, the approach of targeting all debt also focuses more generally on the risks for macroeconomic stability that excessive real leverage can cause, regardless of the source of the loans.10

25. An important problem with Uganda’s rule as presently drafted, however, is that it includes an exception for “arm’s length debt” which can easily undercut the protection afforded by the anti-abuse rule. Section 89(2) provides that the 1.5:1 ratio will not be binding if the debt ratio of the company does not exceed the “arm’s length debt” amount—defined as the amount of debt that an unrelated financial institution would be willing to lend to the Ugandan subsidiary. This exemption should be removed. It will be very difficult for the URA to prove that an unrelated party would have lent any specific amount to such an MNE subsidiary—it simply provides a way for the parent MNE group to avoid the binding rule.

26. Uganda’s provision against profit shifting by means of abusive interest deductions should be augmented (or replaced) by adopting an “earnings stripping” rule. This is the more modern approach—utilized for example by Germany, and recommended under the G20/OECD base erosion and profit shifting (BEPS) Action 4. This approach implements a “fixed ratio” rule that, regardless of the debt to equity ratio, limits net interest deductions claimed by a corporation to a fixed percentage of earnings before interest, taxes and depreciation (EBITDA).11 Interest deductions so restricted are allowed to be carried forward. The ratio is normally set at around 30 percent of applicable EBITDA, although recommendations for such a rule, including in the BEPS Action 4, propose a range of possible ratios from 10-30 percent. This type of rule avoids problems such as the need to define what constitutes debt (in a world with complex hybrid financing instruments), and serves as a backup in combatting the charging of an excessively high interest rate on related party debt. Thin capitalization rules such as that in Section 89 are subject to both those problems.

27. A very rough analysis of the possible effect of adopting such an “EBITDA” interest-stripping rule indicates that it could be more effective at restricting profit shifting than the existing 1.5:1 thin capitalization ratio, as well as better addressing the problems referred to just above. Table 3 and Figure 5 below compare the revenue impact of these two rules for the major MNE subsidiaries for which data was available, by sector. As shown in the figure, under the existing thin capitalization ratio test of 1.5:1 debt to equity, no company in this data set other than in the financial sector reaches the restrictive debt ratio. It is important to note that the structure of the financial sector – based upon borrowing and lending in various forms – naturally involves high debt ratios. For this reason, financial institutions are normally not subject to the same debt/equity/earnings stripping rules as non-financial enterprises. But even in the largest companies in the four non-financial sectors examined here, some would fall within the restriction of an earnings stripping rule measuring deductible interest against EBITDA. This of course varies depending on the level chosen—10 to 30 percent—as shown in the Table 3.

Table 3.

Uganda: Estimated Revenue Impact of the EBITDA Rule, 2014-15

article image
Source: URA and IMF staff calculations.
Figure 5.
Figure 5.

Average Debt to Equity Ratio, 2014-15

Citation: IMF Staff Country Reports 2017, 207; 10.5089/9781484309360.002.A005

Source: URA and IMF staff calculations.

28. As shown in Table 3, using a limit on interest deductibility of 20 percent of EBITDA, and not including financial sector companies in this limit, about one-quarter of the companies in the test data would be subject to restriction on interest deduction, resulting in the disallowance of about 18 percent of their deducted interest, for an increase in tax revenue from these companies alone of about 0.1 percent of overall tax revenues. This should not be taken as exactly correct, or as extendable in a linear way to all companies. It is just a very rough example of the potential impact of such a change on a single small sample of companies for a single year.

Service and management fees

29. It is common for multi-national groups to centralize charges for management overhead costs and general service fees (commonly through a centralized services company in a favorable tax jurisdiction) to charge out to members of the group. This practice raises numerous issues, including classification of the charges (as services, royalties or embedded in cost of goods), transfer pricing (including whether the charges sufficiently benefit the associated Ugandan company that they would be paid by an independent enterprise and, if there is a mark-up on the charge, whether the mark-up is appropriate), and whether the payment is eligible for relief from withholding tax by treaty. The income will be subject to a 15 percent withholding tax in Uganda.12 Where the fees are so subject to withholding, the result is (as in the case of interest payments) lowering the Uganda effective tax rate—since these fees are deductible at 30 percent. The same effect can be achieved with charges for royalties. And these benefits can be further increased through use of treaties and transfer pricing.

30. Excessive relief for management fees under existing DTTs. The withholding tax called for under Uganda’s statutory law is easily avoided by routing management and service fee charges through treaty jurisdictions. A number of treaties allow Uganda to impose withholding tax on technical services only if they are provided to a person in Uganda for a period or periods of 6 months or more in a 12 month period (or 183 days in a period of 365 days).13 If that provision does not apply, or if the treaty does not include such a provision, then services generally would be classified as business profits and would not be permitted to be taxed by Uganda if the service provider does not have a permanent establishment in Uganda. As a practical matter, this means that very few services provided from outside of Uganda will be taxed at all, which advantages non-local services over local services.14

31. Adoption of Uganda Model DTT technical services article. Uganda should seek to amend all of its existing treaties to add the technical services article set out in the model treaty approved by Cabinet. (Such an article will be added to the UN Model DTT in the near future).

Transfer pricing

32. There is broad scope for engaging in abusive transfer pricing. Transfer pricing refers to the prices that are charged for transfers of goods, services, property or other items of value in transactions between associates. The term “associates” is defined broadly in ITA Section 3 to reach a range of related persons that can include persons within a family, corporations or businesses that have a common owner (s), and also includes unrelated persons where one person can direct the action of another person in the context of the transaction in question. The term used in the Ugandan transfer pricing regulations for transactions among associates that are subject to transfer pricing rules is “controlled transactions.”

33. The revenue risk from transfer pricing is high. Transfer pricing can occur even in the context of transactions occurring solely within one country, where the parties to the transaction will be taxed differently. For example, this can happen where one company is subject to full income tax and the associated company enjoys a tax exemption. Where more than one country is involved transfer pricing is particularly important. In even the simplest two-country case it is likely that there are disparities in tax rates, and hence opportunities to derive advantage from abusive transfer pricing. And increasingly transactions and ownership of operations are structured through intermediary legal entities that do not bear a meaningful corporate tax because they are located in countries that either do not tax income or that facilitate very low effective tax rates on the income. These countries include a number of Uganda’s important trading and treaty partners, such as The Netherlands and Mauritius as well as prospective treaty (and trading) partners such as the United Arab Emirates (UAE) and Qatar. The benefit from income-shifting in these cases goes almost exclusively to the taxpayer while the revenue loss is suffered by Uganda. Moreover, local Ugandan businesses suffer the economic disadvantage relative to MNE subsidiaries of being unable to take the same advantage of income shifting to a lower tax jurisdiction.

34. The arm’s length standard is the measuring principle for determining whether a controlled transaction gives rise to inappropriate income shifting. Under the Ugandan transfer pricing regulations, the arm’s length standard is satisfied when the results of a controlled transaction are consistent with the results that would have been realized in a transaction between independent persons dealing under the same conditions. This is the internationally accepted standard.

35. URA has made progress in enforcing arm’s length transfer pricing. Transfer pricing compliance improves when rules are reasonably clear, incentives for aggressive transfer pricing are reduced, and enforcement is focused, tough and fair. Uganda has taken important steps in promulgating transfer pricing regulations and in putting resources into a large taxpayer office and advanced transfer pricing training for personnel. Moreover, the URA’s large taxpayer office is now pursuing transfer pricing cases through complex audits. The payoff from these steps is already starting to emerge in terms of staff understanding and capacity, and should bring in increasing revenues.

36. Continuing the progress made. Important elements of a robust transfer pricing enforcement strategy include obtaining relevant information and engaging in rigorous risk evaluation to target cases with a high enough payoff to justify the resource-intensive effort to support a transfer pricing adjustment. An initial starting point is to develop metrics for evaluating information currently reported on tax returns of Ugandan associates to identify transfer pricing risks.

37. Information and documentation. It is important to identify substantial controlled transactions and to obtain all the information relating to a controlled transaction, including information regarding the income from the transaction that is realized in the rest of the group. This is a key element in being able to fully evaluate transfer pricing risk. The transfer pricing documentation that has been required under the transfer pricing regulations’ practice note is fulsome and complete, but is only provided to URA on request. The penalties added to ITA for failure to provide this information are an important addition to the law. Sufficient information should be required to be provided with the tax return with respect to material controlled transactions to allow URA to screen for transfer pricing risk and thereby make more efficient use of its resources. In addition, the URA should be authorized to require the taxpayer to provide “see-through” profit information, and test the price in the controlled transaction against the reasonableness of the resulting profit split among relevant associates in relation to the actual activity carried on by those associates. Identification should be made easier by adoption of country-by-country (CbC) reporting called for by the OECD’s BEPS project.

38. The advent of multilateral CbC reporting is a signally important international tax development and Uganda should take full advantage of it. Under the G20/OECD BEPS Action 13, a MNE with substantial revenues (generally consolidated group revenue in excess of EUR 750 million annually for a prior year) would be required to engage in the three-tier CbC reporting. This means that it would have to make and maintain a Master file, a local file and a country-by-country report which would be exchanged by its home jurisdiction tax authorities with those of the jurisdictions where the multi-national group has activities—though, importantly only if such jurisdictions meet standards required to maintain confidentiality. Countries that require such reports by multi-national subsidiaries under their local law, that enter into appropriate information exchange agreements (described below), and which meet those standards are eligible to exchange and receive CbC reports.

39. The potential of CbC reporting for Uganda is considerable. A significant number of the MNEs that have investments in Uganda (including for example major MNEs such as AB Inbev, Barclays, Bhata Airtel, British American Tobacco, Diageo, Ericsson, Heinekens, LaFargeHolcim, MTN, Sabco, Standard Chartered Bank, Total)15 likely will be required by their home country or by a country in which they have a substantial operation to participate in CbC reporting. Moreover, a brief review of the current list of CbC information exchange relationships on the OECD website ( shows, for example, that both The Netherlands and Mauritius are engaging in CbC reporting exchanges. Accordingly, it is recommended that:

  • a. MOF pursue adoption of legislation that would require a Uganda reporting entity of an MNE Group to annually file a CbC Report; and

  • b. URA continue its progress in taking steps to be able to execute the Multilateral Competent Authority Agreement on the Exchange of Country-By-Country Reports (CbC MCAA, found at that will authorize such exchanges.

E. Source Rules

40. The source rules in section 79 of the ITA should be reviewed and modernized. The source rules were drafted over 20 years ago and have not fully kept pace with modern cross-border trade and investment. The Ministry has this on the agenda, but is waiting for pending legal cases to be resolved before changing the rules. The revenue implication of these rules may be quite large.

41. Services income is an example of problems with the existing rules. There are four possible bases under Section 79 upon which services income may be characterized as Ugandan-source income: (i) the income is derived from services rendered in Uganda (Section 79(c)); (ii) the services are rendered under a contract with the Government of Uganda (Section 79(d)); (iii) Uganda is permitted to tax the income under a tax treaty (Section 79(r)); and the income is attributable to “any other activity” occurring in Uganda, including an activity conducted through a Ugandan branch (Section 79(s)). As noted in prior advice,16 countries are increasingly? aligning the source rule for services income with that applicable to royalties; looking to the place of utilization rather than the place of performance (see Sections 79(j)(i) and (ii)). Thus, services income should have a Uganda source when paid by: (i) a Ugandan resident (other than as an expense of a business carried outside Uganda through a branch); or (ii) a non-resident as an expenditure of a Ugandan branch. This also would align the statutory source rule under the ITA with the taxing right under the technical service article in Uganda’s model tax treaty (Art. 13(5)). The source rule should apply regardless of whether the service fee is paid directly to the non-resident service provider, or through a recharge arrangement with a related nonresident company.

42. Source rules that look to the place of contract are also subject to manipulation and should be reviewed. The rule for sourcing income from sales of goods purchased by the seller under ITA section 79(a)(ii) looks to where the agreement of sale was made. This kind of rule is readily manipulated and should be reconsidered, in order to associate the sale with where the selling activity is carried on, rather than where the contract is executed.

F. Uganda’s Tax Treaties Introduction

43. Double tax treaties (DTTs) are concluded to eliminate double taxation and—to a lesser extent—to protect against the avoidance of taxation. Double taxation arises if both residence country and source country claim taxing rights over the same income. DTTs contain provisions that allocate income between the contracting countries, whereby the source country typically gives up (part of) its taxing rights in exchange, it is hoped, for more FDI, lower investment costs, better access to patents and knowledge-based capital with positive spillovers, or nontax benefits. In addition, a treaty should improve cooperation between taxing authorities in carrying out their functions, including by the exchange of information with a view to preventing or detecting avoidance or evasion of taxes.

44. Risks of double taxation only exist where there is a significant level of existing or clearly intended cross-border trade and investment between two States. In the circumstances of Uganda, which is only a source country in connection with all its current treaty partners, it is especially important to balance the revenue loss from reducing withholding tax rates and restricting source taxation of business income against evidence of concrete benefits that Uganda as the source country will receive in return. It is highly risky for a country to adopt a treaty in the hope of future investment that it is believed would not be made without the treaty, without engaging in a disciplined analysis to ensure with some confidence that this is true, and that, if so, the benefits outweigh the potential costs.

45. In all countries, a government policy and process regarding choice of treaty partners should preserve a central decision-making and negotiating role for the Ministry of Finance. Further, there should be a government policy addressing the factors described in the next paragraphs of this section of the report. Uganda has recently adopted guidance at the Cabinet level regarding such a treaty policy, which very well addresses these issues. Following that guidance consistently in negotiation and implementation of DTTs will serve Uganda well.

46. This section of the report addresses the following issues. (i) Best practices in entering into treaties; (ii) the existing Uganda treaty network, including treaties that are in the process of adoption or negotiation, examining specific issues which expose the tax base to revenue loss or which should be modified to conform to current best practices including withholding tax rates, limitation of benefits provisions, source taxation of technical services, and offshore sales of interests in immovable property entities; (iii) issues raised by the EAC treaty; and (iv) opportunities and challenges presented by the opening for signature of the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (the Multilateral Legal Instrument, or “MLI”).

Importance of a Tax Treaty Policy

Maintaining a transparent treaty policy

47. Uganda’s international tax policy is manifested in its tax law provisions (generally, the Income Tax Act (ITA)) governing taxation of non-residents (including international investors) on income derived from Uganda and of Ugandan residents on their foreign income. Under ITA Section 88, a Ugandan DTT modifies domestic tax law. Accordingly, once entered into, Ugandan DTTs become an integral element of Uganda’s international tax policy. Negotiating DTTs effectively requires the development of a tax treaty policy that aligns DTTs with Uganda’s international tax policy; it is this that is now embodied in Uganda’ recent protocol providing guidance on negotiating DTTs, and the accompanying model DTT for Uganda that is discussed below. These have not yet been made public, and we recommend that at least the model DTT and a summary of the treaty policy be made public. A country should not negotiate a DTT in isolation from other DTTs; each treaty negotiation should be understood as part of a country’s overall tax treaty program.

48. The new Uganda DTT model is based on the UN Model. That model is now being amended to take account of recent changes in best practice, including modifications recommended in the G20/OECD Base Erosion and Profit Shifting (BEPS) project, and provisions embodied in the text of the MLI discussed further below. This report includes comments on the Ugandan model treaty below, including the recommendation of updates to keep it in line with this current best practice.

Using cost-benefit analysis to assess the need for a treaty

49. Recent thinking about treaties in developing countries has begun to focus on the inefficiency of bilateral treaties generally. These are time consuming to negotiate, they absorb resources to administer (including processes for refunding treaty-based reclaims of withholding taxes), they tend to be based on one-sided investment flows—as in Uganda’s case—and therefore result in revenue loss. Treaties are hard to terminate as a diplomatic matter, even when the formal termination provisions only require a customary six-month notice. Treaties also are hard to change and can stifle tax law reform in the international context. A statutory reform affecting cross-border business can be effectively negated by a provision in a single treaty. The inflexibility of treaties once adopted is particularly significant currently as the OECD and UN have proposed the most significant changes in international tax rules in decades and changes may be expected to continue in coming years.

50. A dramatic expansion of Uganda’s tax treaty program may not be the highest and best use of government (especially Ministry of Finance and URA) resources. Policymakers should consider whether it is possible to achieve the same objectives as a DTT through a mix of (i) unilateral measures in the Uganda laws governing taxation, (ii) agreements governing specific investments or concessions,17 (iii) bilateral or multilateral tax information exchange agreements (TIEAs), and (iv) bilateral or multilateral investment protection agreements, including sector specific instruments such as shipping and aircraft agreements.

51. The decision to enter into each specific DTT should be based on an overall cost benefit analysis. This requires balancing the tax revenue loss from treaty source taxation concessions against gains in the form of increased investment (i) that would not be made absent the DTT, and (ii) that leads to job creation and increased incomes for residents. The latter point is critical; increased investment is beneficial to Uganda only to the extent it improves the quality of life of Uganda’s residents, so that its benefits do not accrue only to a nonresident investor.

52. The cost of administering a DTT program should not be underestimated. In addition to the time and cost of treaty negotiations, there are substantial ongoing costs. Implementation of a tax treaty network may require adjustment at several levels, including changing forms and filing requirements for non-residents who claim treaty benefits (it is critical to vet non-resident claims for treaty eligibility). Further, to avoid inadvertent treaty disputes with taxpayers, internal treaty coordination mechanisms are required from local office level up to central policy administration. Finally, ongoing monitoring is required to uncover tax avoidance schemes and enforcement is required to attack schemes that are uncovered (whether by URA or as often is the case from information shared by other governments).

53. Ugandan revenue loss is likely to be substantial from DTTs. For a country such as Uganda that principally is a recipient of foreign investment, each DTT results in revenue loss relative to the same investment without the treaty. The extent of the revenue loss depends on the specific terms of the DTT. In general, two broad factors affect the amount of revenue loss from a DTT. First, is the extent to which Uganda’s right to tax nonresidents is reduced by a DTT. The key source taxation issues affected by treaties are (i) withholding tax reductions on interest, dividends, rents and royalties, (ii) restricted net-based source taxation of business profits, including under many DTTs fees for technical, professional and management services, and (iii) limits on taxation of capital gains, including from offshore sales of stock in companies (especially those holding resource and telecom assets).

54. Taking the concession of source taxation just described as a given in the nature of DTTs, the second—and critical—factor affecting revenue loss is the extent that the treaty is susceptible to “treaty shopping” use by investors from a country other than the treaty partner. While it is possible to employ anti-treaty shopping (e.g., so-called limitation on benefit or principal purpose) provisions, it often is possible to circumvent these limitations with sophisticated planning or as a result of limited enforcement capacity. As a practical matter, each treaty should be viewed as a potential “treaty with the world”. It is therefore critical to preserve the right for Uganda in each tax treaty to prevent the abuse or misuse of the treaty.

55. Revenue loss can increase over a long period of time as investors shift investments to the weakest treaty in a country’s tax treaty network. It is difficult to measure the exact amounts of revenue loss that result from a treaty program, but the magnitude can be material. It is critical to understand that a country’s network of DTTs is only as strong as the DTT that permits the greatest reduction of source taxation and/or the greatest abuse. Adopting a treaty with the wrong treaty partner without adequate protections becomes a loophole for all inbound FDI affecting the country’s entire tax base. The data on Uganda’s FDI already shows that it is substantially funneled through the Netherlands and Mauritius (Table 4).

Table 4.

Uganda: Withholding Tax Rates

article image
Source: IBFD and Uganda ITA
Table 5.

FDI Stocks by Source Country, 2012-2015 (Values in US$ million)

article image
Source: Statistics Department, Bank of Uganda

56. Ugandan gains from entering DTTs need to be carefully evaluated against the costs of doing so. Balancing against inevitable revenue loss from a DTT are potential gains in the form of investment. Importantly, the gains that should be counted are from investment that would not be made in the absence of a DTT. Before entering into a DTT it is best practice for a country to assess: (i) what investment would come from the treaty partner country, (ii) whether the investment would occur without the presence of a DTT, and (iii) whether the investment could be encouraged by an instrument that does not involve the revenue loss of a DTT. It must be recognized that, if asked, investors routinely claim that a DTT is necessary for investment, but such claims should be carefully scrutinized. Comments from the private sector indicate that treaties are not necessarily a material factor.

57. The tax administration advantages of a DTT (such as exchange of tax information, resolution of tax disputes and assistance in collection) generally can be accomplished under other agreements, such as the Convention on Mutual Administrative Assistance in Tax Matters, recently adopted in Uganda, or by entering into other bilateral agreements (such as a tax information exchange agreement or a shipping and aircraft agreement) without the revenue loss accompanying a DTT. Moreover, it is possible in theory to enter into a limited treaty that includes information exchange and a mutual agreement procedure to address investor concerns about resolving tax disputes.

Issues in Existing Ugandan DTTs

58. Uganda has 10 DTTs in force, with: Belgium, Denmark, India, Italy, Mauritius, The Netherlands, Norway, South Africa, the United Kingdom, and Zambia. None of these treaties conforms to the Uganda model treaty or to current international best practices, highlighting the observation made above regarding the rigidity of DTTs as an instrument and difficulties of maintaining an up to date DTT network.

59. Uganda also has a number of treaties that have been signed, but for which notice has not been given or the treaty partner has not completed the formalities for the treaty to enter into force, as well as some under negotiation. The jurisdictions with which Uganda has signed a DTT but for which the DTT is not in force include China, Egypt and the UAE. The texts of the Egyptian and UAE DTTs have not been made public nor has the IMF mission team reviewed these documents. Uganda also has ongoing negotiations with several countries, including Qatar, Serbia, and Turkey. Succeeding sections consider areas of potential revenue leakage in the current and proposed Ugandan treaty network specifically.

Withholding tax rate reductions

60. In general, and except for the DTT with Zambia and the dividend rate with The Netherlands, Uganda’s DTTs provide only modest reductions in withholding tax rates from the statutory rates. While broadly speaking, there is limited risk of revenue loss from these withholding rates, even isolated reduced rates, if in a treaty with a country such as The Netherlands whose law facilitates its use as an intermediary country, can result in revenue loss. The DTT with The Netherlands limits dividend withholding 18 to 5 percent for companies holding at least 50 percent of the capital of a company. The FDI data show disproportionate FDI from The Netherlands, strongly suggesting that Dutch companies are used as an intermediate treaty shopping vehicle.

61. Viewed from a different perspective, however, even a reduction from a 15 percent to a 10 percent withholding rate is a substantial percentage reduction. The signed DTT with China (not yet in force) has a 7.5 percent rate on dividends and an effective 7 percent rate on equipment royalties (and under the Protocol to the DTT only 70 percent of equipment royalties (rents) are subject to the withholding). The very old treaty with Zambia includes a zero rate of withholding on dividends, interest and royalties if the item of income is taxed in Zambia.

62. The laws of Mauritius and The Netherlands present clear potential for treaty shopping advantage to be taken of lower withholding tax rates and other treaty benefits. A Mauritius or a Dutch company has potential to be used as a conduit company. The DTT withholding rates from Uganda are reduced (dividends 10 or 5 percent, interest 10 percent and royalties 10 percent) thus offering clear advantages over direct investment from a non-treaty country and even other treaty countries. These treaties do not have limitation on benefits provisions. There is no experience yet with the revised statutory treaty anti-abuse rule of ITA Section 88 and how its beneficial ownership and economic substance conditions to use of a DTT might apply to limit treaty shopping.

Fees for technical, management and professional services

63. Under ITA Sections 83 and 85, the Uganda withholding rate on management or professional fees paid to a non-resident company is 15 percent. It is common practice for global groups to consolidate group service expenses and to charge them out to the group, often with a mark-up (that sometimes is “parked” in a low-tax intermediary jurisdiction). In discussions with URA and private advisors, there appeared to be a consensus that use of management or technical services fees paid to nonresidents is one of the most substantial source of cross-border revenue leakage.19 Even where a reduced withholding rate does not apply on such payments, from the company’s point of view the reduction of the effective tax rate by virtue of paying the deductible service fee at 30 percent Uganda corporate tax, and being subject only to the 15 percent statutory withholding rate, saves 15 percentage points of tax. If a DTT can be used to further reduce the effective tax rate on management fees to a 10 percent withholding rate or even a zero rate, this base erosion tax planning can achieve 20 or even 30 percentage points of tax savings.

64. There will be circumstances in which treaties may be used to avoid withholding on management and technical service fees altogether. As an example, under the DTT with The Netherlands, a company that does not pay material tax in each of those jurisdictions (through features of their respective laws) could be used to charge out services fees. Because this existing treaty does not have the advisable technical services article found in the Uganda Model DTT, the fees would be classified as business profits and therefore exempt from Ugandan withholding tax under the business profits article of the treaties. 20

65. To address the technical services issue, it is advisable to include a separate technical services article such as that in the Uganda model DTT in all future treaties. The Uganda model DTT Article 13 is a good model. This approach is increasingly accepted internationally.

66. The MLI–discussed further below-does not provide an avenue to cure the DTT technical services issues. It does not include options to add technical services provisions to DTTs. The G20/OECD BEPS project did not reach an agreement on technical services, and these have been addressed almost exclusively through the UN.

Offshore indirect transfers of interest

67. “Offshore indirect transfers of interest” (OITIs) have become of increasing concern for the tax base of developing countries. These are discussed in the following section VI of the report, as they involve both domestic law and treaty issues. In short, analysis strongly favors a primary right to tax of the country in which the underlying immovable assets are located, with respect to capital gains realized by their indirect transfer offshore.

Limitation on benefits provisions

68. It now is recognized as a best practice for a DTT to include a limitation of benefits (LOB) provision that restricts use of a DTT to entities that have a sufficient economic nexus with the residence country to justify the source country tax concessions given under the DTT. This is particularly important for DTTs with countries whose tax systems either do not impose material income tax, such as Mauritius, Qatar and the UAE, or whose tax systems offer a tax-efficient conduit for income from Uganda to a third country investor, such as the Netherlands and the United Kingdom. The Uganda model DTT, with one significant modification, presents a good start on this problem, though it also would be advisable to add a principal purpose test as well.

69. Uganda recently revised a domestic law statutory anti-abuse provision. ITA Section 88 has been amended to provide that a nonresident entity (other than a publicly listed company) seeking to benefit from a DTT would not be eligible for benefits if: (i) it is not the beneficial owner of the income; (ii) it does not have unrestricted ability to enjoy the income and determine its uses; and (iii) it does not have economic substance in its country of residence. This is a shift from the earlier limitation of treaty benefit rules that restricted the application of treaty benefits to where 50 percent or more of the underlying ownership of the nonresident was held by resident individual or individuals of that other contracting state for purposes of that DTT. The effect of this new rule in practice remains open at this point, as does the legal status of the ITA provision in relation to treaties. For these reasons, it is important for Uganda to include limitation of benefits provisions in all its treaties.

70. The MLI may be used to add limitation on benefits provisions to DTTs. The MLI has options that could include needed limitation on benefits language in all treaties where both bilateral parties agree to do so; Uganda should make sure that when the time comes, this step is taken—which only can be effective, however, to the extent that its treaty partners will agree.

EAC treaties

71. The EAC multilateral DTT (MDTT) signed in 2010 is intended to promote intra EAC trade and investment. Because it does not include any LOB provision, the EAC MDTT could be used in theory by any third country investor that establishes a company in one of the EAC countries. In general, it appears that the laws of the EAC states would require payment of reasonably material tax in the event of conduit use of a member country entity, and therefore at this point this is not too much of a danger. Nonetheless, internal laws can change quickly, so it would be advisable that the EAC MDTT be updated with suitable LOB provisions before it is brought into force. Other provisions which would conform the treaty to current best practices, such as including authorization for source country taxation of direct and indirect sales of immovable property, also should be added. This would seem to be a reasonable path to propose, since the treaty was now drafted some time ago, and best practices in these regards have—particularly within the last two years—moved forward. It would be good to bring the EAC treaty into force in line with current best practice.21

Multilateral legal instrument (MLI)

72. The MLI was made public at the end of November, 2016 and opened for signature as of December 31, 2016. It includes 18 articles that each would permit a party to the MLI to modify existing in force bilateral DTTs, provided that the DTT partner agrees in each case to make the same modification.22 Uganda has not signed the MLI, as of this time.

73. The treaty modifications provided in the MLI range from the highly technical rules affecting hybrid mismatches in Part III, to the relatively less important change to treaty preambles in Article 6, to the much more important LOB provisions of Article 7, offshore sale of immovable property entities provision of Article 9 and mandatory binding arbitration rules of Articles 18 - 26. A signatory to the MLI must determine its position in respect of all the provisions in relation to each of the signatories of in-force DTTs. The MLI is open for signature and the OECD held an initial signing ceremony on June 7, 2017. The convention will remain open for additional signatories after that date. Of Uganda’s treaty partners, only Mauritius, Norway and Zambia did not sign the MLI and both Mauritius and Norway have sent letters to the OECD saying they intend to do so as soon as possible. The countries must indicate to which treaties they intend for their MLI positions to apply. Thus, for example, The Netherlands has identified its treaty with Uganda as one to which it would apply the MLI. Its proposed provisions include a “principal purpose test” type of anti-abuse rule. Where a partner has signed the MLI adopting this position, it will apparently not be possible for Uganda to have included in its MLI changes an LOB provision, as the principle purpose test is the default position in the MLI.

74. The MLI offers Uganda an opportunity to add important provisions to in force treaties, particularly with respect to LOB provisions and taxation of offshore sales of immovable property entities. These, and other changes, will only be effective, as described just above, with the agreement of DTT partners to accept the changes—so the MLI cannot be viewed as a complete panacea for dealing with DTT revenue leakage. Further, some of the issues discussed herein, most importantly the payment of technical services, are not covered under the MLI in any case.

75. It appears that the process under the MLI for determining a country’s positions involves at least the following steps:

  • Determine Uganda’s position in respect of each substantive article of the MLI. Is the position proposed one that Uganda would like to adopt as part of its treaty policy?

  • With respect to each in force DTT, determine

  • Whether Uganda would want to include the DTT within the ambit of the MLI as a Covered Agreement,

  • If so, how each substantive article would interact with the DTT in question so as to determine whether to accept application of the MLI rule or opt in to the MLI rule (as appropriate for that article) and what reservations if any are needed in relation to the article,

  • Determine whether the treaty partner has signed the MLI and what provisions it has notified that it is willing to accept, and

  • Make notifications are necessary for each MLI article in relation to the DTT in question.

76. In general, the substantive articles of the MLI (omitting the procedural dispute resolution provisions), would be favorable to Uganda, or at worst would do no harm. The LOB and offshore immovable property entity provisions in particular would improve existing treaties—to the extent that treaty partners will agree to include those.

G. Indirect Transfers of Interest

77. The Ugandan authorities have identified as a substantial concern the taxation, or lack thereof, of gain on offshore indirect transfers of interest (OITIs) in Ugandan assets. This is an issue that has become increasingly important to the tax bases of developing countries. The problem is the ability to sell interests in important domestic assets, without being subject to capital gains taxation in the country where those underlying assets are located. This is achieved by selling interests in foreign entities that own the assets (outside the location country) rather than selling the assets themselves. The central conceptual question raised by OITIs is that of how taxing rights should be allocated between the country where the underlying asset is located and other jurisdictions involved in the transaction.

78. Current understanding of best practice—as embodied in both the OECD and UN model DTTs—is to permit source country taxation of offshore transfers where the underlying asset is immovable property located in the source country. This is generally defined to include real estate, and, importantly, mineral deposits, licenses relating to them, and rights thereto.23 In order for such taxation to be effectuated, both domestic income tax law and relevant provisions of any tax treaties governing relations between the source country and the country in which the indirect transfer takes place must explicitly extend the reach of taxation to such transfers. The Uganda ITA includes a provision defining as sourced in Uganda income from indirect transfers of shares in a company the assets of which consist directly or indirectly primarily of interests in immovable property, where the interest or share constitutes a business asset.24 Amendments in 201525 provide a mechanism to ensure that tax can be collected where an indirect transfer of an underlying mineral license occurs, by requiring notification of the transfer by the licensee, making the licensee the liable agent for the income tax incurred by the non-resident transferor, and defining such a license as a business asset for source rule purposes.26

79. These provisions of the ITA could be extended and strengthened. It would be desirable to extend taxation of offshore transfers to indirect holdings beyond the resource sector. This can be done without going as far as to tax all indirect transfers of shares in any Ugandan company.

80. There is good economic justification for extending the definition of immovable property to reach assets embodying location specific rents that are clearly linked to national assets—e.g., telecommunication licenses; rights to operate public power grids or water systems; forests and fish. These types of assets derive their value from economic rents specific to the location country, and have been at the heart of some of the major disputes over such offshore taxation in a number of jurisdictions, including in Uganda.

81. It is also necessary to provide in any applicable double taxation treaty that such gains on indirect transfer of (immovable) assets are to be taxed in the source country (Uganda). The practice is embodied in the OECD and UN model tax conventions in Article 13(4). When more than 50 percent of the value of the asset being directly transferred is derived from immovable property in the source country, the right to tax capital gains (or losses) is, under treaties based on these models, allocated in full to that country; where between 20 and 50 percent of the value is so derived, the taxation is proportionate. It is critically important to note that this important provision will only apply if the right to so tax indirect transfers is not limited in applicable tax treaties to the residence country (as opposed to allowed to Uganda, the source country). Uganda’s treaty with India (and the treaty signed with China) does allow the source country (Uganda) to impose tax on these offshore sales. In considering whether and how to apply the provisions of the MLI, this is one of the items that the MLI could be used for to add to existing treaties. The proposed revisions would include the needed language in all treaties where both bi-lateral parties agree to do so; Uganda should make sure that when the time comes, this step is taken—which only can be effective, however, to the extent that its treaty partners will agree. Moreover, if any treaty does not have such a rule, it will attract use solely for that reason.

82. The new Uganda model DTT should be amended (i) in order to capture in Uganda gains on indirect transfers of immovable property as defined (presently) to include mineral rights –at present only direct transfers are included within the scope of the treaty; and (ii) to extend the reach of immovable property by defining it more broadly to include all instances of value derived from location specific rents arising in Uganda.


  • Burns, Lee, Philip Daniel, Diego Mesa Puyo, and Emil Sunley, 2015, “Uganda Fiscal Regimes for Extractive Industries: Next Phase,Technical Assistance Report (International Monetary Fund: Washington, D.C.).

    • Search Google Scholar
    • Export Citation
  • Burns, Lee, Philip Daniel, Diego Mesa Puyo, and Emil Sunley, 2015, “Uganda Implementing Fiscal Regimes for Extractive Industries: Technical Notes,Technical Assistance Report (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • Gaspar, Vitor, Laura Jaramillo and Philippe Wingender, 2016, “Tax Capacity and Growth: Is there a Tipping Point?Working Paper, No. 16/234 (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • Grigg, Jack, Victoria Perry, Stephen Shay, 2011, “Uganda Tax Policy: Issues in a Changing Economy,Technical Assistance Report (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (IMF), Organization for Economic Co-operation and Development (OECD), United Nations (UN), World Bank Group (WBG), 2015, “Report on Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment” (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (IMF), 2016, “Tax Policy, Leverage and Macroeconomic Stability,Policy Paper (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund (IMF), 2016, “Uganda Seventh Review under The Policy Support Instrument” (International Monetary Fund: Washington D.C.).

    • Search Google Scholar
    • Export Citation
  • Uganda Revenue Authority, 2015, DRAFT Tax Exemptions Report (Uganda Revenue Authority: Kampala).



Average tax rate


Base erosion and profit shifting

CbC -



Corporate Income Tax


Double Tax Treaty


East African Community


Earnings before interest, tax, depreciation and amortization


Foreign direct investment


Gross domestic product


International Bureau of Fiscal Documentation


International Monetary Fund


IMF Worldwide Revenue Longitudinal Database


Income Tax Act


Large taxpayer unit


Limitation of Benefits


Multilateral Competent Authority Agreement


Multilateral Legal Instrument

MNEs -

Multi-National Enterprises


Ministry of Finance


Medium taxpayer unit


Organisation for Economic Co-operation and Development


Offshore Indirect Transfers of Interest


Policy support instrument

TA -

Technical assistance


Tax Information Exchange Agreement

TP -

Transfer pricing


United Arab Emirates

UN -

United Nations


Ugandan Revenue Authority


Withholding tax


This report was produced by Prepared by Victoria Perry, Li Liu and Stephen Shay.


This report does not address issues specific to the extractives sector, as there has been much specialized technical assistance to the Ugandan authorities in that area.


Uganda’s overall tax-to-GDP ratio is estimated to be 14 percent of GDP in FY 16/17 using GDP projected in the main IMF staff report.


Table 5 further provides a breakdown of inbound FDI stocks by source country during 2012-2015.


Other than Mining and Construction; this report does not address specifically issues relating to fiscal regimes for natural resources, which have been extensively addressed by specialized IMF missions.


The analysis is based on anonymized detailed CIT tax records for a selected sample that include the largest 29 companies based on annual turnover in five key economic sectors in Uganda, including Agriculture, Forestry, Fishing; Manufacturing; Wholesale and Retail Trade; Information and Communication; and Financial and Insurance Services between financial years 2012-13 and 2015-16. Total CIT contribution from these companies accounted for approximately 45 percent of total CIT revenue in Uganda in 2014-15.


The average ATR was substantially lower than the average corporate tax paid as a percentage of pre-tax chargeable profits (right orange bar), which is very close to the statutory tax rate of 30 percent ; the issue is the reduction in chargeable profits.


International Monetary Fund, 2016, “Tax Capacity and Growth: Is there a Tipping Point? Working Paper”, November (Washington).


See “Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment,” REPORT PREPARED FOR THE G-20 DEVELOPMENT WORKING GROUP BY THE IMF, OECD, UN AND WORLD BANK, October, 2015.


International Monetary Fund, 2016, “Tax Policy, Leverage and Macroeconomic Stability, Policy Paper”, December (Washington).


EBITDA is a cash flow measure that is viewed as representing a reasonable commercial measure of an enterprise’s ability to borrow money and service debt. To this end, earnings stripping rules’ reliance on allowing interest deductions up to a fixed ratio of EBITDA creates link between allowable interest deduction and economic activity generated by the firm.


This is the case unless the management fees and constitute Uganda source income for the recipient carrying on business in Uganda through a permanent establishment.


See e.g., Convention Between the Republic of Mauritius and the Republic of Uganda for the Avoidance of Double Taxation and Prevention of Fiscal Evasion with respect to Taxes on Income (“Uganda –Mauritius Treaty”) Art. 13. Uganda’s existing treaties, other than those with The Netherlands and Zambia (and the signed treaty with China) have provisions with effects similar to the Mauritius treaty.


A recent change to the Ugandan VAT law attempts to generally advantage local services by imposing reverse charged VAT on imported services but not allowing businesses to take a VAT credit for that tax. This would offset the income tax advantage described here. However, it is an extremely blunt instrument to achieve local content requirements and distorts the operation of the VAT.


The structure of foreign investment in Uganda is readily available through publicly reported sources, including by using Orbis.


See Uganda Fiscal Regimes for Extractive Industries: Next Phase (Technical Assistance Report, IMF Fiscal Affairs Department), P. Daniel, L. Burns, D. Mesa Puyo, E. Sunley, June 2015


While specific arrangements or investment incentives are generally not advisable, they may (if unavoidable) cause less revenue damage than a treaty with provisions that open the opportunity for tax avoidance to the world.


It can be argued that reduced rates for dividends are less problematic than for interest (and other deductible payments) since dividends come from income that has already been taxed by the source country (in theory).


The mission was not able to verify this quantitatively given an absence of sufficiently reliable data specific to these kinds of charges.


In the absence of a Ugandan permanent establishment of the investing company.


Recent IMF technical assistance to the EAC Secretariat focuses on these issues.


A handful of the provisions may be applied to one and not both of the DTT parties.


Some countries have extended the reach of taxation of indirect transfers offshore to include interests represented by any interests in domestic companies (e.g., Peru; India), but this is not common.


ITA Section 79(g)


ITA Section 89GF


This is necessary in Uganda as only business assets are subject to capital gains taxation on transfer.

Uganda: Selected Issues
Author: International Monetary Fund. African Dept.