Tax Policy in MENA Countries
Looking Back and Forward
Author: Mario Mansour

Contributor Notes

Author’s E-Mail Address:

This paper reviews trends in taxation and revenue in MENA countries over 1990-2012, with a focus on non-resource taxes. On average, non-resource revenues declined slightly, while resource revenues soared. Country experiences vary: rates of main taxes and their revenues tend to be higher in the Magreb than in the Mashreq, except for the value-added tax, where lower rates are associated with equal or higher revenue; most oil producers raise little tax revenues—generally less than 5 percent of GDP—and most have reduced them since the late 1990s. But there are similarities: unlike common experience around the world, income taxes (not indirect taxes) have partially compensated for lost revenue from trade liberalization; revenues from indirect taxes have remained stable; personal income taxes have played an unimportant role as a revenue tool; and fees and stamp duties are significant revenue sources. Looking forward, tax reform challenges will also vary across countries: the Maghreb needs to focus on efficiency-enhancing reforms, especially in capital income and consumption taxes; the Mashreq have some room to increase revenue; and, there are ample opportunities to improve equity and reduce complexity of tax systems in all countries. Finally, the recent decline in oil prices and revenues is a reminder that even resource-rich GCC countries need to lay the basis of a tax system for the future.


This paper reviews trends in taxation and revenue in MENA countries over 1990-2012, with a focus on non-resource taxes. On average, non-resource revenues declined slightly, while resource revenues soared. Country experiences vary: rates of main taxes and their revenues tend to be higher in the Magreb than in the Mashreq, except for the value-added tax, where lower rates are associated with equal or higher revenue; most oil producers raise little tax revenues—generally less than 5 percent of GDP—and most have reduced them since the late 1990s. But there are similarities: unlike common experience around the world, income taxes (not indirect taxes) have partially compensated for lost revenue from trade liberalization; revenues from indirect taxes have remained stable; personal income taxes have played an unimportant role as a revenue tool; and fees and stamp duties are significant revenue sources. Looking forward, tax reform challenges will also vary across countries: the Maghreb needs to focus on efficiency-enhancing reforms, especially in capital income and consumption taxes; the Mashreq have some room to increase revenue; and, there are ample opportunities to improve equity and reduce complexity of tax systems in all countries. Finally, the recent decline in oil prices and revenues is a reminder that even resource-rich GCC countries need to lay the basis of a tax system for the future.

I. Introduction

This paper reviews revenue developments and taxation in the Middle-East and North Africa (MENA) region over the period 1990-2012. It analyses the levels and composition of government revenues, and assesses current taxation structures. The purpose is to draw some lessons for the region in terms of how well past reforms have performed against the standard yardsticks of tax policymaking (i.e. revenue, equity, and efficiency), and where future reforms may be directed. Although the paper takes stock of the evolution and relative importance of resource revenues for the region, its focus will be on non-resource tax policy.

Study of tax policy in MENA region often revolves around revenues from natural resources and the need to diversify them. Yet, the region is one of the most diverse in the world, politically, economically, and geographically. Endowments of natural resources vary significantly across countries; political systems include quasi-democracies, monarchies, and kingdoms; a number of countries are experiencing a difficult period of socio-political transformations (a result of the Arab Spring), which has been closely linked to strong perception of inequities in how governments have conducted economic policy, particularly taxation; and the number of fragile states has increased in the past decade, a result of unsuccessful political reforms following revolutions and wars. Closer to the technical aspects of tax policymaking lie also significant differences across MENA economies in terms of size, colonial inheritance of legal and tax systems, openness, migration, and sensitivity of countries’ main tax bases to the forces of globalization, to name just a few.

This diversity prevents general statements on how this group of heterogeneous countries should reform their tax systems. However, and as this paper will attempt to show, this should not preclude the possibility that certain tax policy changes or reforms may be appropriate for groups of countries with common economic characteristics. Thus, this paper pays particular attention to differences across country groups, attempting to draw some general lessons for reforms but keeping in mind that ultimately policy advice must remain country-specific.

A flavor of the differences across MENA countries is in Table 1, which shows selected population and key macro-fiscal indicators for five groups, identified according to three criteria: intensity of revenues from extractive industries, which is defined as having resource revenues exceeding 50 percent of total revenues for more than half of the period under study; geographical location; and trading blocs (in particular, the customs union of the Gulf Cooperation Council (GCC)). This partition of the region into relatively small groups allows possible general conclusions, and will be used throughout the paper (see Annex I for the list of countries in each group); it is also closely linked to income levels as GCC countries are all high-income, and all others are upper or lower-middle income. There are interesting differences across the groups, including in terms of changes between 1990 and 2012:

  • Population in the GCC group doubled over the period. It now exceeds the population of resource-rich Maghreb countries. A significant share of this increase is due to migration from other MENA countries, and more recently from Central Asian countries. This has tax implications both for the GCC region and the MENA countries providing labor to the GCC—the current debate in some GCC countries on how to tax expatriates’ outbound remittances is a direct implication of this migration.

  • The share of oil GDP in OME countries (Iraq, Iran, and Yemen) tripled. This helped push their per capita GDP from $3,925 to $11,704, and almost doubled government revenue per capita.

  • GDP per capita (at purchasing power parity) increased in all groups, and particularly in the non-resource group. At issue is the distribution of this increase in wealth and its implication for tax policy making, particularly in terms of equity.

  • The share of resource-GDP in total GDP in resource countries increased from less than 30 to nearly 45 percent. It is therefore not surprising that resource countries have increased their reliance on resource revenue to finance their budgets. Also of interest is the increase in real per capita revenues from USD 832 to USD 1,584.

  • In the non-resource country groups, government revenues increased slightly from 19.2 to 21.7 percent of GDP. In real terms, per capita revenues declined in the Mashreq group by about one-third, and increased in the Maghreb group by one-third.

  • Finally, tax revenues remained very low and stagnant in the resource group, and increased slightly in the non-resource group.

Table 1.

Selected Indicators; 1990 and 2012

article image
Sources: WEO, WDI, IMF Staff reports and other documentation, and author’s calculations.Notes: Government average total and tax revenues are weighted by GDP.OME: Other Middle-East countries, which include Iran, Iraq and Yemen.

The rest of the paper is organized as follows. Section II discusses general trends in revenues. Section III discusses these trends in the context of each main component of the tax system. Section IV combines results from II and III to identify key future challenges, including possible policy choices. Section V concludes.

II. Looking Back: Trends in Revenue Levels and Composition

In this section we review in detail revenue developments in MENA countries, based on a dataset built specifically for this paper, which draws from the International Monetary Fund staff reports and other documentation used by Fund staff surveillance and program activities (Appendix one provides details). These developments help form an understanding of how tax systems have scored on revenue and other important yardsticks, such as equity and efficiency, and provide a reference point relative to future objectives that policymakers may want to pursue.

At the regional level

In general for the region, and as we would expect, oil and gas dominated the evolution of government total revenues since 1990 (Figure 1)1—at 0.99, the correlation between total and resource revenues is remarkable. Resource revenues were already high at the beginning of the sample period, slightly below 15 percent of GDP, and shot up relatively quickly in the early 2000s to reach 31 percent of GDP in 2008. Although the impact of the 2008 crisis was significant, bringing down resource revenues to about 20 percent of GDP, they were back up to 29.5 percent in 2012. However, developments in oil and gas prices since June 2014, the prospects of a long-term decline in world demand, and growth in shale gas reserves outside the region, are likely to have a negative impact on resource revenues in MENA for the foreseeable future, with possible long-term consequences for fiscal policy in general, and tax policy in particular. Some of these consequences have already been felt in recent years, with a number of countries in the region scaling back their costly energy subsidies.2

Figure 1.
Figure 1.

Trends in Government Revenues, 1990-2012

(percent of GDP)

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Source: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages.

Perhaps less expected is the fact that tax revenues have remained relatively stable over the sample period. They fluctuated in a narrow range of 6 to 9 percent of GDP3, and the level at the end of the period was slightly below its beginning—although experience varies across countries as we will see later. Also of interest is the relative importance of non-tax (non-resource) revenues—an issue we will also examine in some detail later. These include various types of fees and stamp duties that can number in the hundreds, with most taking the form of transaction taxes on specific activities, such as registration of real estate property and motor vehicles, stamps on financial transactions, etc. The use of such fees is prevalent not only in the GCC, where taxation exists only in a rudimentary form, but also in other MENA countries which deploy standard taxes on income and consumption.

The impact on revenues of recent political instability and revolutions in the region, which started in Tunisia in December 2010, seems to have been negligible on average. At the country level, however, the results vary: in Syria, where 2010 is the last year for which data is available, the civil war may have had devastating impact on revenues and public finances (non-quantified yet); in Libya, although non-resource revenues have been weak historically, they dropped from about 4 percent of GDP before 2010, to less than 1.5 percent in 2012.

These developments in government revenues in the MENA region differ from those in the rest of the world. IMF (2011), for example, reports averages in high-income countries that are well above the MENA region for most of the sample period, but less volatile. This is because unlike MENA, other resource-rich countries, particularly in the upper-middle and high-income groups, generally have well developed non-resource taxes, whose contribution has the dual benefit of increasing government revenues and making them less volatile.4

Overall for the region, the small decline in the tax ratio between the early 1990s and 2011-12 is roughly equal to an uncompensated loss in trade taxes of about 1 percentage point of GDP (Figure 2).5 Unlike experience elsewhere (see IMF, 2011), sales and excise taxes have remained relatively stable. However, revenues from the corporate income tax (CIT) increased, but by less than the decline in trade taxes.

Figure 2.
Figure 2.

Changes in the Composition of Tax Revenues, 1990-2012

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages. CIT excludes extractive industries.

At the level of selected country groups

As noted earlier, it is interesting to look at country groups in the region according to not only endowment in natural resources, but also geography and trade relations. Figure 3 shows development in total and tax revenues for two non-resource groups (non-resource Maghreb and Mashreq) and three resource groups (resource Maghreb, GCC and OME).

In the non-resource groups (Maghreb and Mashreq), the level and evolution of total revenues have been broadly similar, but developments in recent years have differed. Revenues declined sharply in the Mashreq across all countries (Egypt, Jordan, Lebanon, and presumably Syria), by an average of about 5 percentage points of GDP.

Figure 3.
Figure 3.

Trends in Total and Tax Revenues by Group

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages.

Differences in tax revenues among these two groups are more significant. First, Maghreb countries tend to raise more than Mashreq countries—about 6.4 percentage points of GDP on average for the period. Second, Mashreq tax revenues have been stagnant for the period, while those in the Maghreb have increased over 2005-12. Relative to international norms, taxation levels in the Mashreq are comparable to those in Low- and lower-middle-income countries, while those in the Maghreb are comparable to those in upper-middle-income (see International Monetary Fund, 2011). While the first result is driven by Egypt, a lower-middle-income country, in the other group only Tunisia is an upper-middle-income country while Morocco and Mauritania are both lower-middle-income. The tax ratio in the Maghreb seems to be higher than what standard econometric analysis suggests on the relationship between the tax effort and income levels.6

For resource country groups (resource Maghreb, GCC, and OME), natural resources have been the main source of government revenues since 2000. Revenue levels in the Maghreb and GCC are significantly higher than in the OME group,7 but developments over the period are broadly similar in all groups. Tax revenues have been stagnant, and contribute little to the revenue effort, except in the Maghreb, where Algeria has a well developed tax system, yielding tax levels similar to those in low-income countries. On average, developments in tax revenues in resource countries are consistent with econometric findings suggesting that resource revenues crowd out partially non-resource revenues—about 20 cents on average for each dollar increase in resource revenue (Crivelli and Gupta, 2014). The exception is the OME group, where the mean tax effort increased from 2.7 to 4.5 percent of GDP over the period, 8 reflecting primarily an increase in non-resource revenue in Iran.

Differences among the selected country groups become more pronounced when we look at the evolution of the main components of tax revenues. Figure 4 shows such evolution for the two non-resource groups (Magreb and Mashreq). In the first, sales taxes have stagnated over the period, and the drastic loss of trade taxes in the 1990s was offset by an increase in income taxes in the 2000s. This explains the U-shape tax revenue ratio shown earlier, and suggests that Maghreb countries did not plan a timely replacement of their trade taxes revenue loss.9 With trade taxes currently at about 1 percent of GDP, trade liberalization in the form of reduction in import tariffs or bilateral trade agreements has been largely completed, and pressure to mobilize revenue will squarely fall on domestic taxes.

Figure 4.
Figure 4.

Trends in the Composition of Tax Revenues in Non-resource Groups

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages. CIT excludes extractive industries.

In the second group, a very different evolution took place. The loss of trade taxes was largely compensated with domestic sales taxes. PIT revenues remained relatively constant while CIT revenue declined by about 1.5 percentage points of GDP. Like in the Maghreb, trade taxes no longer constitute a significant source of revenues.

These differences in tax revenue developments reflect both different starting points in the two groups and different policy choices. However, the data clearly show some convergence over the period in the composition of tax revenues among the two groups.

Developments in the composition of tax revenues in the resource groups (resource Maghreb, GCC and OME), although not as important, are not insignificant (Figure 5). First, and except in the GCC where only limited use is made of the CIT and trade taxes, the evolution of tax revenue is more volatile, reflecting perhaps spillover effects from the volatility of the resource sector onto non-resource revenues. The exception is the personal income tax, which is unimportant and stable over time, reflecting the fact that it is mostly derived from wage withholding on government employees—its recent increase in the Maghreb is largely due to an increase in public sector wages in Algeria.

Figure 5.
Figure 5.

Trends in the Composition of Tax Revenues in Resource Groups

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted by GDP. CIT excludes extractive industries.

III. Key Policy Issues

This section reviews the main features of MENA’s tax systems, which in combination with the backward-looking analysis of revenue levels and composition allow us to look forward to how tax systems in such a diverse region might evolve in the next 10 to 15 years to respond to ongoing (and possibly other) challenges,10 with globalization and political transition being prominent ones.

A. Trade Liberalization and Tariff Policy

Tariff policy, a key element of trade policy, is distinct from domestic tax policy in that its primary role is to protect domestic production by creating a wedge between prices of imported goods and services and those of domestically-produced substitutes. Although this paper does not address trade policy issues, there are a number of interactions between tariff and domestic tax policies that should be reflected in the evaluation and development of tax policy choices.

First, tariffs generate revenues which can be important to government budgets, and hard to replace when governments liberalize trade, either unilaterally or by entering into free trade agreements (FTAs). Second, even when tariffs do not raise significant revenues, such as when they are prohibitively high or used jointly with quotas, reducing tariffs could have fundamental consequences on the composition of the various tax bases of the domestic tax system. Third, protection provided by tariffs to domestic production sectors may generate returns on investment above normal rates of return (i.e. those that would be required by investors in fairly competitive markets). Governments may want to tax such excess returns (i.e. economic rent), at rates above standard rates. Fourth, investment tax incentives may be ineffective when tariff rates are high on imported intermediate and capital inputs; for example, a CIT tax holiday may be of little use to a firm facing a high tariff rate on imported capital that has no domestic substitutes. In sum, domestic tax policy needs to respond to trade liberalization for revenue, efficiency and equity purposes, and tariff and tax policy need to be carefully coordinated when sectoral strategies are formulated.

From a revenue perspective, tariff policy has evolved toward more liberalization in MENA countries and less reliance on tariffs as a revenue source. Trade taxes today average about 1 percent of GDP in both resource and non-resource MENA, and they rarely exceed 2 percent of GDP. Their share in total tax revenues declined from about 26 in the early 1990s to 15 percent in 2012. Figure 6 provides a more detailed picture of this evolution; it shows, in addition to trade taxes (as a percent of GDP), the collected tariff rate (CTR)11 and imports (also as a percent of GDP)—emphasizing as before differences across the region.12

Figure 6.
Figure 6.

Trade Taxes: Tax Base and Collected Tariff Rates

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and other documentation; WTO, ICT, and UNCTAD (2014); and author’s calculations.

In resource countries, both trade taxes and the CTR have declined from already low levels in the early 1990s, while the tax base (imports) expanded significantly, from 33 to over 55 percent of GDP. The response of the base to the decline in the CTR could not have been very important given the already low level of the CTR in the early 1990s—even though Figure 6 suggests a strong relationship.13 This evolution begs the question why trade taxes have not increased or at least remained stable in resource countries. Part of the answer can be found in much lower tariff rates today than in the 1990s, especially in GCC countries where a single common external tariff of 5 percent applies to most imports; another part lies in the use of exemptions and other types of tariff preferences.

In non-resource countries (with Maghreb and Mashreq groups shown separately),14 there is also a clear and significant decline in both trade taxes and the CTR, with the paths of decline different in the two groups. Unlike resource countries, however, there is little sign of a significant expansion of the tax base—the increase in the early 2000s receded somewhat in recent years, particularly in the Mashreq.

It is tempting to conclude that the loss of trade taxes and the corresponding decline in the CTR are primarily due to trade liberalization; but the issue is more complex. Table 2 shows statutory tariff rates (bound and applied most-favored nation (MFN)), maximum rates, and the number of distinct rates used in MENA groups (data for some countries were not available from the World Trade Organization), along with the loss of trade taxes over the sample period. Average bound rates are very high in most countries, and average MFN rates are not particularly low, especially in non-resource countries. Maximum rates are also very high, and more importantly perhaps, the number of tariff rates is high in all countries, except in Kuwait. All this suggests that the low revenue take from tariffs can be explained, at least in part, by the use of preferential treatments, such as exemptions and reduced rates.

Table 2.

External Tariffs: Bound, Maximum and MFN Rates

article image
Sources: IMF staff reports and other documentation; WTO, ICT, and UNCTAD (2014); and author’s calculations.Note: Bound, maximum and MFN (most-favored nation) rates are for 2013 or 2012.

The loss is estimated as average trade taxes in 2011-12 minus the average in 1990-91.

This complexity in the tariff may play a role in the wider issue of why trade in the MENA is well below its potential. Behar and Freund (2011) find that MENA exports have increased since the early 1990s, but the region’s exports are two-thirds below its potential, estimated according to fundamentals—and the pace of growth is even slower in non-resource countries. These authors also find that intra-MENA trade is well below its potential, suggesting that regional trade agreements have not worked as well as expected. Bhattacharya and Wolde (2010) explore the empirical determinants of constraints to trade in MENA; they find high transportation costs, and inefficient and slow customs clearance procedures to be significant.

From a revenue standpoint, this analysis suggests that MENA countries can maintain, and perhaps even increase their trade tax revenues, by flattening the tariff structure, reducing the number of rates, and reducing maximum rates. This would lessen the need to provide tariff exemptions or reductions, and may improve the overall progressivity of the tax system—since imported goods tend to be consumed disproportionately by well-off individuals. Details matter, especially in this area, where careful attention should be paid to a number of factors while designing such reforms, including: existing and contemplated FTAs, and the interaction of tariffs with domestic taxes, in particular consumption taxes.

B. Taxation of Consumption

Taxation of consumption in the MENA region takes one of three forms: general sales taxes; excise taxes; and other specific transaction taxes. This section addresses issues with the first two. The third, which could have elements of fee-for-service, and raise different but related issues, is covered in III.E.

Value-added taxes (VATs)

General sales taxes in MENA are primarily of the type destination-based value-added tax (VAT). They have two origins: (1) cascading sales taxes that applied to goods prior to the 1990s; this is the case of Maghreb countries and Egypt,15 which transformed old sales taxes into VATs by modifying their laws to expand the tax base to services and reduce cascading by enhancing refunds of taxes on intermediates and capital goods; (2) new VATs designed without any historical heritage; this is the case of Lebanon, Jordan, and the OME group. These origins are important in that they have had implications for a number of key VAT design issues, including the choice of the level and number of rates, the registration threshold, exemptions, and possibly others. They have also affected the dynamics of national policy debates on VAT reforms.

VAT rates vary widely across countries (Table 3). Mashreq and OME rates tend to be lower than Maghreb rates. Low-rate countries also happen to be those which designed a VAT without a historical heritage, as a new source of finance (e.g. Iran) or to replace revenue loss from trade liberalization (e.g. Lebanon). Standard rates have increased in all countries, as has the number of rates (except in Algeria). The level and number of VAT rates in MENA span the spectrum of international practice, except in relation to standard rates in the European Union (EU), which are typically higher than the highest MENA rate of 20 percent in Morocco; EU countries also tend to use multiple rates more than other countries (see European Commission, 2014).

Table 3.

Value-added Taxes: Rates and Revenue Contributions

article image
Sources: IMF staff reports and other documentation; International Bureau for Fiscal Documentation; and author’s calculations.

The low rate in 2013 applies to certain food items and petroleum products. The threshold shown applies to service activities; for other activities, the threshold is about USD1,700.

The other rates apply to tobacco and petrol respectively.

The other rates apply to semi-processed gold (2%), gold jewellery (3%), and mobile telephony and international telecommunication services (10%). VAT revenues include revenue from excises.

Retailers do not have to register unless their turnover exceed USD232,000 approximately.

Retailers do not have to register unless their turnover exceed about USD64,000.

Egypt’s General Sales Tax functions as a VAT on goods, and a series of excises on business-to-business services applied in most cases at the standard rate.

The threshold applies to traders. Other thresholds appy to manufacturers (USD14,000 or USD85,000), and services (USD42,000).

C-eff is C-efficiency, estimated using the standard VAT rate. All things equal, this would overestimate (underestimate) true c-efficiency in countries with lower (higher) non-standard VAT rates.

Registration thresholds are low in some countries relative to international norms (e.g. Algeria, Egypt, Morocco and Tunisia),16 and their definition is problematic. Although defined in terms of turnover, they vary according to activity and the legal form of the business. The argument frequently justifying the use of multiple thresholds is that some businesses have a higher value added than others (e.g. services), and therefore should be subject to the VAT at lower turnover levels. There are two major flaws with this argument. First, the use of a registration threshold responds to two realities: the capacity and resources of the tax administration, which are often limited; and the accounting capabilities of taxpayers who must meet minimum filing requirements, which could also be limited by the availability and use of accounting standards, and familiarity of enterprises with them. The share of value-added in relation to turnover has little connection with these realities. Second, multiple thresholds are significantly more complex to administer, and create tax planning opportunities that taxpayers can exploit, and unscrupulous tax inspectors can use for rent seeking.

The contribution of MENA VATs to tax revenues ranges from 30 to 55 percent (except in Iran and Egypt, where it is much lower). Although this may seem significant, c-efficiency17 estimates suggest that the yield can be improved through base broadening, especially in Egypt, Algeria, Tunisia, and Iran.18 These countries provide, in addition to low rates, extensive exemptions of final consumption items (see appendix table A1 for a list of main exemptions under current VAT laws).

The cost of VAT tax expenditures (exemptions and lower rates) provided under current VAT laws in MENA is possibly sizable, and should concern policymakers for two reasons. First, the equity implications of such policies are far from trivial since targeting using an indirect tax, especially one of general application such as the VAT, is too costly—i.e. poor consumers may benefit, but much less than the well-off—and raise administration problems that are hard to address, and which are amplified by opportunities for tax avoidance (e.g. classification of imports). Second, and particularly for those countries in need of revenue, increasing revenues from other tax sources could be more problematic and distortionary (in terms of negative impact on growth and employment) than scaling back VAT exemptions and low rates.

Studies on the distribution of VAT tax expenditures abound for developed countries, but are much less common for the rest of the world.19 These studies generally find that VAT tax expenditures, although generally progressive (but not always), are poorly targeted and very costly. For example, a study by the Organization for Economic Cooperation and Development (OECD, 2014) find that this is the case for low-VAT rates on food in 15 OECD countries. Interestingly, the same study finds that low VAT rates targeted to social or cultural objectives (e.g. restaurants and hotel accommodation) are regressive. Bird and Gendron (2007) review the available evidence for developing countries. They report interesting results: depending on country circumstances, VATs can be slightly regressive or progressive; zero-rating can make VATs more regressive (consisting with the findings of the OECD study cited above); and, replacing excises and import duties by a general consumption tax such as VAT may actually improve tax progressivity in most low-income countries. These results are country specific; they depend on consumption profiles and VAT design. Nevertheless, they should be of some use to MENA countries in thinking about the costs and equity implications of VAT policy.

Table 4 shows the distributional impact of low VAT rates in Morocco, and low rates and exemptions in Tunisia, using the same methodology in OECD (2014). The results are consistent with expectations and broadly similar to those obtained for OECD countries. The top quintile of the income distribution reaps a much higher share (about 40 percent) of VAT tax expenditures, while the first quintile receive less than 10 percent, and the poor even less in the case of Morocco.

Table 4.

Distributional Impact of VAT Tax Expenditures in Morocco and Tunisia

article image
Sources: Fouzi Mourji (2011) for Morocco; and author’s calculations for Tunisia (based on 2010 household survey data from the Institut National de la Statistique).Note: Estimates assume no behavioural responses to removing tax expenditures.

These results suggest that more efficient policies can be designed whereby revenue from streamlining VAT tax expenditures can be used to increase support to the poor—and perhaps even revenue to the government. Social safety nets (SSNs), which are conditional cash transfers, are examples of how subsidies can be better channeled to the poor. SSNs are however poorly designed and underfinanced in MENA (Silva, Levin and Morgandi, 2012), and can suffer from the same political economy issues that perpetuate the use of poorly targeted VAT preferences. Although there has been progress recently in improving them, more work is needed. In the short term, they are unlikely to be able to cope effectively and efficiently with the need to expand support to the poor. But even poorly targeted subsidies can be better than VAT tax expenditures, as Keen (2013) concludes from a recent survey of the literature. In addition, SSNs can be a powerful tool to improve acceptance of tax reform.

Other options which take a more gradual approach to reforming VATs could also be contemplated. These would involve reducing the scope of VAT exemptions and low rates to a limited list of items, rather than categories, which share in the consumption basket of the poor is very high. But even this approach has proven politically difficult in MENA countries, and elsewhere.20

Excise taxes

Excise taxes are levies on specific goods and services, which generally apply in addition to a VAT. Consistent and comprehensive data on excise revenues are scarce; here, we report data on the main ones applied in MENA countries: tobacco, alcoholic and non-alcoholic drinks, petroleum products, cars, and mobile telephony. Excise revenues in percent of GDP declined significantly in all MENA countries since 2000, except in Egypt and Lebanon, where they remained relatively constant (Table 5).21 In all countries, and particularly those where the yield is below 2 percent of GDP, excises are untapped, potentially significant, revenue sources.

Table 5.

Excise Tax Revenues ; 2000-01 vs. 2011-12

article image
Sources: IMF staff reports and other documentation; WEO; and author’s calculations.Notes:

The two-year average is weighted by GDP.

In the resource group, Libya does not apply excises, neither do GCC countries; and data for Yemen are not available.

The 2011-12 figure for Syria is for 2009-10. Jordan’s figures include a real estate registration fee which could not be carved out.

The decline in excise taxes reflect a number of policy and other factors that cannot be disentangled due to data limitations. Buoyed by other revenue sources, and perhaps out of other concerns, some countries reduced their excise tax rates or did not index them to inflation. For example, in 2006 Mauritania abolished all major excises (tobacco, non-alcoholic beverages, and cars) in anticipation of oil revenue by 2010. But revenues from oil turned out to be significantly lower than expected, and Mauritania reintroduced some of these excises recently—though at very low rates. In Tunisia, excises on alcoholic drinks are specific for low-alcohol content drinks such as beer and wines, and not indexed to inflation. Excises on high-alcohol content drinks such as whisky are so high, reaching over 650 percent, that very little legal trade takes place. The result is that Tunisia raises some revenue from an important domestic consumption of beer (some 0.3 percent of GDP), but virtually nothing from other alcoholic drinks. In Lebanon, excise tax rates on alcoholic drinks are specific and very low, but tariffs on imported drinks are high; the result is that alcohol excises raise very little revenues.

Excises in MENA countries are sometimes used to provide protection to domestic production over and above that provided by the tariff. In Tunisia for example, the tobacco excise differentiates between brands, with higher rates applicable to imported brands that the government monopoly cannot produce—due principally to technology constraints. Such practice is inconsistent with country obligations under the World Trade Organization (WTO).

Other non-policy factors also affect the yield of excise taxes, in particular market structures and price controls. Tobacco is a good case in point. In most Maghreb and Mashreq countries, tobacco production and wholesale distribution is a government monopoly, with prices fixed typically according to a formula that distributes the total tax take between Treasury, tobacco monopoly, and wholesale and retail distributors. One of the results of such a system is that tobacco variety is very low, especially on the high side of the price spectrum, and illegal trade is very important. Another result is that revenue from this source is underestimated as part of it end up in the form of monopoly rent.

In the GCC, there are no excise taxes. But the GCC common external tariff (CET) applies a rate of 200 percent on alcoholic drinks and 150 percent on tobacco. Given religious restrictions on the consumption of alcohol in many GCC countries, and CET exemptions for diplomatic missions, the tariff on alcohol yields little revenues; and so would an excise. However, GCC countries should be able to mobilize important revenue from excises on other consumptions: tobacco, cars, and non-alcoholic drinks are obvious candidates.

C. Taxation of Business Profits and Investment Incentives

The corporate income tax plays an important role in MENA’s tax systems, with the exception of GCC countries, where it applies mostly to foreign companies. It currently generates between 1 to 1.5 percent of GDP in revenues in resource countries (excluding GCC), and 2.3 to 4.3 percent in non-resource countries (Figure 7).22 In this latter group, the CIT contribution is high (just below 20 percent of tax revenues), and comparable to its contribution in developing economies.23

Figure 7.
Figure 7.

The Corporate Tax: Rates, Revenues, and Implicit Bases

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Sources: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages. GFCF is gross fixed capital formation.

Corporate tax rates in MENA have declined significantly, and there are signs of rates convergence. In the early 1990s, average top CIT rates across country groups ranged from about 35 percent in the Mashreq to over 55 percent in OME. In 2012, rates stood between 18 to 28 percent. This trend has been noted elsewhere around the world, and interpreted as a form of corporate tax competition.24 In recent years, however, this decline has slowed, and Egypt even increased its top rate from 20 to 25 percent in 2012.

Despite the decline in statutory CIT rates, CIT revenues have increased in all groups since the early 1990s, except in the Mashreq. A number of factors specific to MENA, and experience elsewhere, may help explain this and inform future policy debate on how countries should further reform CIT rates.

First, the implicit CIT base (i.e. the ratio of CIT revenues to the product of the standard tax rate and GDP) 25 doubled since the early 1990s in groups where the increase in CIT revenues has been important (resource groups, excluding GCC, and the Maghreb) (Figure 7). This suggests that corporate profits as a share of GDP have increased, and that the corresponding revenue gain more than compensated for the loss from the reduction in tax rates.

Second, the relationship between top CIT and PIT rates may have changed. Empirical studies have found that entrepreneurs’ choice of the legal form of carrying business depends on this (among other factors): if the CIT rate is higher (lower) than the PIT rate, business activity is more likely to be carried through unincorporated (incorporated) entities.26 The ratio of CIT to PIT rates in MENA has increased slightly during the 1990s to exceed one, but declined in the past decade to levels below or close to 1 in most countries—of particular interest is the fact that this ratio is the lowest in the non-resource Maghreb, where the increase in the implicit CIT base and revenues is the most significant. Although this does not establish causality, it does suggest that the relationship between the corporate and personal tax rates may be the cause of distortions in the choice of the legal form of doing business—and the fact that PIT revenue has been rather weak lends some support to this hypothesis (more on the PIT later).27

Third, and especially in the Maghreb, CIT rates on selected sectors were not reduced in line with the standard CIT rate. Algeria taxed trading activities at 25 percent rather than the standard 19 percent until January 2015—when a single 23 percent rate starts applying; Jordan and Tunisia tax banking and telecoms at 35 percent, instead of 14 and 25 percent, respectively;28 and Morocco taxes banking at 37 percent instead of 30 percent. These sectors tend to be regulated with barriers to entry that may be the source of important economic rent—telecoms in particular have grown significantly in the past decade in these countries.

Finally, it is possible that lower CIT rates may have improved investment incentives, and hence broadened the corporate tax base. This is an empirical question that is difficult to establish, given the complexity of investment decisions and the myriad of factors that can affect them.29 Figure 8 shows for the Mashreq some correlation (not causation) between the CIT rate decline and private fixed capital formation. However, the relationship between the CIT rate and foreign direct investment as a percent of GDP (FDI) is ambiguous—the 2008 crisis and political turmoil in the Mashreq are likely more important factors that explain the collapse of FDI inflows. Moreover, the relationship between the CIT rate and investment is complicated by the fact that all countries in the region provide investment tax incentives, making the standard CIT rate of questionable relevance to investment decisions—and suggesting that effective tax rates, particularly forward looking average effective tax rates (AETRs), are more relevant for real investment decisions than headline CIT rates or effective marginal tax rates (EMTRs). This is an issue we now turn to in some more detail.

Figure 8.
Figure 8.

Corporate Income Tax Rates and Private Investment in the Mashreq

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Source: IMF staff reports and documentation, WEO, and author’s calculations.Note: Figures are weighted averages. GFCF is gross fixed capital formation.

Investment tax incentives

All MENA countries provide tax incentives in investment laws or free-zone laws. With a few exceptions, free-zone laws have spread mainly since the 1990s (see appendix Table A2), while investment laws are older. Tax incentives are only one aspect of these laws, although an important one; other aspects include guarantees of non-expropriation, free repatriation of profits for foreign investors, liberal foreign exchange rules, one-stop shop for foreign investors to lessen the administrative burden from dealing with government agencies, etc.

A cursory reading of income tax laws in MENA suggests that many countries provide additional incentives in their general tax laws, including non- or low-taxation of certain types of incomes, such as agriculture, small business income, capital gains, and reinvested profits. It is not uncommon that such incentives overlap with those provided in non-tax laws. The issue of tax incentives is therefore more complex than the discussion in this section.

There are more common features in the use of tax incentives than differences across MENA countries. First, with a few exceptions, all incentives invariably take the form of temporary (most used) or permanent tax exemptions (less used). Second, the holiday period is generally 5 to 10 years, but goes up to 20 (Egypt, Iran, Yemen), and 50 years (UAE). Third, only a few countries explicitly reserve incentives to foreign investors (Iraq, Libya, and de facto GCC countries since they do not apply their CIT to domestic or GCC-national enterprises). And fourth, incentives are broadly targeted to a wide range of economic sectors, and are based on criteria such as the location of the investment, job creation, and the use of domestic inputs.

Tax competition is often cited by policymakers as a key reason for providing tax incentives. The concern is losing “tax-competitiveness” to other countries, which could translate into lower FDI inflows. It is likely that this argument has led to a dual-economy in MENA countries, where the primary form of tax incentives is the tax holiday, as noted above—certain investments benefit from CIT exemptions while others are taxed under the general tax regime at relatively high standard rates. This dualism causes two major difficulties for policymaking. First, it creates inequities between tax-favored firms and taxed firms, which may discourage entrepreneurship and risk-taking, and reduce rather than enhance competition. Taxes are not the only source of this uneven level-playing field; the World Bank 2015 MENA Development Report (World Bank, 2015) describes several other policies that have been used in MENA countries to provide privileges to politically-connected firms, and found empirical evidence (in Tunisia and Egypt) that such policies have reduced competition and employment. Second, it creates rigidities that delay the opportunity of reducing standard CIT rates, which until recently remained high relative to developing and emerging economies. Recent Arab revolutions should have, in principle, provided impetus for reforming CIT systems toward more neutrality, but this has proven more difficult than originally thought (see Box 1 for the case of Tunisia)—entrenched privileges are very hard to take away it seems, even when political systems that supported them fail.

Another argument that could explain the generosity and the type of tax incentives provided by MENA governments is a political economy argument. Legislating tax incentives in non-tax laws with complex eligibility criteria and oversight by committees external to the tax administration may provide discretion to politicians in selecting the enterprises that can benefit from tax incentives.30 In fact, in some countries (e.g. Egypt) the tax administration does not have access to the accounting statements of exempt firms, not even to verify transactions with taxable firms.

The empirical evidence on the effectiveness of tax incentives to attract foreign direct investment is mixed. It points to large variations across countries, sectors, and types of investment.31 Generally, incentives that reduce directly the cost of investment (e.g. investment allowances and accelerated depreciation) have been found to be relatively effective in some advanced economies. Tax holidays, which are mostly used in developing countries, rank low in business climate surveys, and are often redundant—i.e. the investment would have been undertaken without the incentive, implying that their cost in forgone revenue is high relative to their effectiveness. Finally, and aside from cost-effectiveness issues, certain tax incentives such as those favoring exports may be inconsistent with countries’ obligations under the WTO—since they amount to export subsidies.

Reforming the CIT and Investment Tax Incentives in Tunisia: Politics and Signaling

Until 2007, two elements characterized the Tunisian CIT: a standard rate of 35 percent; tax holidays (mainly for export profits, referred to as the offshore sector) of 10 years or more, followed by CIT taxation at half the standard rate. In practice, tax holidays were frequently renewed and the half-rate rarely applied. These holidays, provided primarily in the Investment Code and other sectoral laws, were perceived as having played a major role in building the tourism and textile sectors (both geared toward European markets). But no government empirical investigation of tax holidays in Tunisia had ever been undertaken, nor an estimation of their costs. Dividends paid by either taxed or exempt firms were also exempt from taxation in the hands of shareholders.

The 2007 budget proposed a reduction in the standard rate from 35 to 30 percent, except for banking, telecoms and hydrocarbon (BTH) sectors, and taxation of activities benefiting from tax holidays at 10 percent (with grandfathering of existing investments). The idea was to improve neutrality of the CIT by bridging the gap between the onshore and offshore sectors from 35 to 20 percentage points as a first step, and later to tax both at a single rate of about 20 percent—roughly half of the pre-2007 rate. It was thought that this strategy would be politically feasible and not very costly in terms of revenue given that the sectors excluded from the reform were major contributors to the CIT.

While the reduction to 30 percent was approved, the 10 percent rate on the offshore sector was postponed every year until 2014. The offshore sector lobby argued that it needed to keep the tax holiday given competition from China and India in the textile sector (interestingly, not tax competition, but lower wages); it also needed to keep a comparative advantage to other sectors in the Tunisian economy—hence, the reduction in the post-holiday rate to 50 percent of the standard rate should, if any, be linked to the reduced rate (30 percent) and not to the 35 percent that remained applicable to the BTH sectors. The fact that the offshore lobby won the argument sent a clear message that tax holidays under sectoral codes were the government’s preferred policy tool for providing tax incentives; lower CIT rate for all sectors was not yet considered good policy.

In the 2014 budget the post-revolution transitional government proposed and approved a reduction of the standard rate from 30 to 25 percent (keeping the 35 for the BTH sectors); it also proposed and approved a 10 percent rate on the offshore sector. The authorities also announced publicly that the policy objective of the CIT rate in the medium term should be to have a single rate of 15 to 20 percent applicable to all firms (again, except the BTH sectors).

Like most developing countries, and a number of developed, MENA governments do not estimate and publish regularly the cost of tax expenditures—investment incentives or others. Morocco is the only exception; it has published its tax expenditures since 2005 (with estimates for 2003 onward),32 with a presentation by major tax type (i.e. CIT, VAT, PIT, excise taxes, registration fees, etc). For CIT investment incentives, the cost (in terms of forgone revenues) has increased slightly over time, from 0.7 percent of GDP in 2003 to 0.8 percent in 2013, with a peak of 1.2 percent in 2012.

Governments rarely study the benefits of investment tax incentives, and to our knowledge no government does it in the MENA region. Although technical capacities may be a barrier, poor policy management seems to play a bigger role. Empirical studies have considered the relationship between tax incentives and incremental investment;33 the implication often drawn is that if such relationship is positive, it must be beneficial (e.g. good for growth, employment, etc.) But this is somewhat misguided because growth has multiple sources, and the incremental investment that can be had from tax incentives may displace other more efficient investment (e.g. those generated by the forces of competition). It is for this reason that tax incentives are usually recommended where markets fail to generate a socially-desirable outcome. Otherwise, neutrality is a better guide to a competitive tax system than distortions caused by incentives. In the MENA region, non-tax policies affecting investment, such as barriers to entry, capital controls, public infrastructure, high statutory tariff rates, seem to be more important for investment than the CIT. The analysis in the World Bank 2015 MENA Development Report tends to support this view, as do other studies on the importance of tax factors in the presence of weak non-tax factors.34

As a first step towards reforming their tax incentives, MENA countries need to improve transparency and evaluate the costs and benefits of their incentives. Policymakers and their advisors can agree or disagree on the usefulness of various types of tax incentives, particularly in the absence of strong empirical evidence. But tax incentives should be subject to the general rules of good budgetary practice. This means relatively simple eligibility criteria, a minimum of information that beneficiaries must report to the tax administration, and an appropriate penalty regime for non-compliance. It also means that the cost of tax incentives, both past and in the medium term, must be estimated and made public, and that incentives must be subjected regularly (say every three to five years) to a thorough evaluation to assess whether they accomplish their objectives.

D. Taxation of Individuals

Personal income taxes

With the exception of GCC countries, all MENA countries deploy PITs, with progressive rates on wage and business income, and flat rates (often schedular) on portfolio and real property income, including capital gains. This gives MENA tax systems elements of dual income taxation.35

From a revenue perspective, taxation of individuals’ income is one of the weakest areas (if not the weakest) of MENA tax systems. On average, PITs generate about 2 percent of GDP in revenues in non-resource MENA (11 percent of tax revenues) and less than 0.5 percent in resource MENA (9 percent of tax revenues). Top PIT rates have declined substantially in all MENA countries, except in the non-resource Maghreb. Despite this decline, PIT revenues have increased (Figure 9).

Figure 9.
Figure 9.

Average Top PIT Rates and PIT Revenues, 1990-2012

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Source: IMF staff reports and documentation, WEO, and author’s calculations.Note: for the Resource Countries group, the average PIT rate includes zeros for GCC countries (they do not apply a PIT).

Weaknesses in PIT revenues suggest that MENA countries do not fully exploit the tax that is most suitable for building progressivity in taxation. Measuring progressivity is more complex than simply looking at PIT systems. It involves assessing, among other things, how budgetary revenues are spent—issues that are beyond the scope of this paper. However, progressivity in income taxation is important for perceived equity and to build credibility in the tax system.

The non-resource Maghreb (Morocco, Mauritania and Tunisia) is the only region that derives a substantive amount of revenue from its PITs. Higher than average top rates is one factor explaining this, but advances in tax administration, including more effective taxation of non-wage income (particularly in Morocco), are also important factors. Detailed data are scarce, but PIT taxation in most MENA countries amounts to withholding on wage income; other types of income are not apprehended effectively by tax administrations—taxes on income not classified as CIT or PIT, which include withholding on interest and dividends to residents and non-residents, account for less than 10 percent of income taxes.

Differences in PIT rate structures in MENA are more significant than in any other tax areas (Table 6). First, the number of rates tends to be large in some countries (e.g. Lebanon and Syria). Although complexity in PIT systems is rarely due to the number of rates, a large number is unnecessary to achieve a desired progressivity,36 and may be a sign of other more fundamental problems.

Table 6.

Personal Income Tax Rates; 2014

article image
Sources: IMF staff reports and other documentation; International Bureau for Fiscal Documentation; and author’s calculations.

Includes the zero rate.

Highest to lowest positive rate, divided by the number of positive rates.

The lowest rate bracket is computed for the average of the lower and upper bounds.

Second, the lowest positive rate in the rate structure tends to be very low relative to international standards, well below 10 percent. This is particularly the case of countries with a high number of rates, but also some with flatter rate structures (e.g. Jordan and Libya). This is usually a symptom of a low exempt thresholds—i.e. the zero-rate bracket. Structures with very low positive rates for the first bracket of income, combined with low top rates, may be appropriate for resource countries, for which non-resource revenue and progressivity of income taxes are not important considerations—e.g. Libya or Iraq.

Third, the average linear progression—average increase in the rate from one bracket to another—is very high in some countries (e.g. Iraq, Lebanon, Mauritania). Again, this seems to be primarily a symptom of relatively low rates on the first brackets of income.

Fourth, the zero-rate bracket applies at very low income levels, measured as a percentage of GDP per capita. Iran and Morocco are exceptions, where the first untaxed tranche of income is slightly above per capita GDP. Jordan, on the other hand, provides a much larger exemption from income, over 3 times per capita GDP. While it is not atypical in a developed country that the exempt threshold is around 0.25 to 0.5 per capita GDP, higher ratios should apply in developing countries in order to shield a higher proportion of the population from income taxation, given that public services are lesser and of poorer quality—this has the additional advantage of limiting administration costs by eliminating from tax a large number of the taxpayers population, which contribution to tax revenue is potentially very low. This means that lower rates should be consolidated at slightly higher levels.

Finally, the taxation of investment income (interest, dividends, and capital gains) is patchy in most MENA countries. Capital gains on real and financial properties are generally exempt—as discussed later registration fees are used instead, and act as imperfect income taxes on capital gains. Interest income is exempt in some countries, or cannot be apprehended by tax administrations due to information constraints and bank secrecy laws. Dividend income is also exempt in many countries (e.g. Egypt, Iran, Jordan, and until recently Tunisia). The main issue with the taxation of investment income is not so much the rate (most countries impose rates between 5 and 15 percent, and Morocco goes up to 30 percent on certain interest income), but the exceptions, rate differentiation, and inconsistencies in the treatment of various types of income. For example, Iran taxes capital gains from real property at 5 percent, but exempt interest and dividend income; Jordan exempts dividends and capital gains, but taxes interest at a 5 percent final withholding. These inconsistencies lead to three fundamental problems: distortions across saving instruments; weakening of the CIT in its role as a withholding tax for the PIT;37 and ample opportunities for tax planning by high-income individuals.

Social security contributions and the tax-cost of labor

The tax-cost of labor in MENA, reflecting in addition to the PIT social security contributions (SSCs), and its impact on labor force participation and unemployment, should be considered more extensively by policymakers. The tax-cost of labor, and more generally fiscal policy, have been found to have a significant impact on employment and unemployment.38

Although there is no evidence for MENA countries on the relationship between the tax cost of labor and employment and unemployment—studies have so far focused on advanced economies—, statutory rates indicate that such cost could be significant in some countries. Table 7 shows combined PIT and SSC rates considering the top and average rates of the PIT. The latter also happens to be roughly the median rate of the PIT, and is more representative of the effective rate—which cannot be calculated due to lack of data on income distribution. The combined rates tend to be reasonable in relation to advanced economies, but are close to 50 percent in a number of countries (e.g. Algeria, Egypt, Morocco, and Tunisia).

Table 7.

Combined PIT and SSC Rates; 2014

article image
Sources: IMF staff reports and other documentation; International Bureau for Fiscal Documentation; and author’s calculations.

Total of employees’ and employers’ rates.

These averages should be interpreted with caution. They say little about the marginal effects of the tax-cost of labor, which can be very important, particularly for certain segments of the wage distribution—and hence for certain types of employment, such as skilled versus unskilled workers. They concern a fraction of the employed population given that the PIT is not applied widely in most MENA countries due to administrative capacity and other non-tax factors; SSCs could also suffer from similar problems. The combined PIT-SSC rate overstates the true rates due to wage ceilings over which SSCs do not apply—available data indicate that a number of countries apply such ceilings, and that they tend to be low.

Another issue to consider is the implication of differences in the taxation of labor and capital. As noted earlier, all MENA countries provide significant tax preferences which purpose is to lower the tax-cost of capital, in some cases to zero. But as has been recognized extensively in the literature, and from country experiences, the cost of labor is equally important, perhaps even more important in some sectors. The general policy of providing tax preferences to capital while keeping higher tax rates on labor may have adverse consequences on the composition of marginal investment, and hence economic growth.

E. Fees and Stamp Duties

In addition to general taxes, most MENA countries levy fees on the transfer of immovable and movable property, and stamp duties on various types of contracts, deeds, and other official documents. A detailed description of these systems, which are very complex, is outside the scope of this paper. However, Appendix Table A3 provides a general overview of the use of these systems.

Although nomenclature varies across countries, in general, “fees” are used for levies on the transfer of immovable (mostly buildings and land) and movable (mostly publicly- and privately-traded securities) property. Rates are proportional to the contractual value of the transfers, and can be as high as 6 percent—and start as low as 0.5 percent. Most countries impose a fee on immovable property without distinguishing between land and buildings (i.e. on the total value). Only a few countries impose fees on shares and bonds. Moreover, a number of countries have “capital levies” (not documented in the table), which apply to capital contributions (first time and additions) to a corporate entity. These levies are generally fixed amounts.

Stamp duties apply to various types of official documents and private contracts, including life and property and casualty insurance. The rates on official documents are generally specific (often a fixed monetary amount per page, which may depend on the type of page and document issued); those on contracts are proportional to the value of the contract (as in the case of fees) or insurance premiums.

Revenues from fees and stamp duties are very difficult to estimate. However, a good proxy from the dataset constructed for this paper is the sum of “other taxes” and “non-tax revenues”. This proxy is especially robust for non-resource countries, but is tainted by investment income of sovereign funds in resource countries. Figure 10 shows this proxy for the two groups. Revenues are significant, and have been stable in non-resource countries. The decline in resource countries since the early 2000s is mostly due to the impact of the resource sector on GDP.

Figure 10.
Figure 10.

Fees and Stamp Duties Revenues

Citation: IMF Working Papers 2015, 098; 10.5089/9781484364789.001.A001

Source: IMF staff reports and other documentation.

Fees on immovable and movable property, and stamp duties on insurance contracts, act as imperfect substitutes for general taxes on these sectors (i.e. income and sales general sales taxes). As such, they raise a number of issues related to their application to the gross value of transactions, including the following: they distort real estate and capital markets; they discourage transparent pricing as parties have an incentive to understate contract values; and, effective tax rates can be significantly higher than statutory rates.

Despite these shortcomings, fees and stamps on insurance contracts can be justified on tax-capacity grounds. For example, insurance contracts are very hard to tax on a net basis; GST is very difficult to apply to financial transactions; capital gains taxation is also difficult to muster in the absence of detailed historical records on property profiles and values; etc. However, the fee rates in most MENA countries seem very high, given profit tax rates, and may be having detrimental effects on the development and transparency of real estate and financial markets.

While the use of fees on immovable and movable properties can be justified, the use of stamp duties on official documents and contracts is highly questionable. Aside from its lack of economic underpinnings, the administrative costs of such duties often exceed their revenue yield, and create a culture of rent-seeking. In many cases, they should simply be eliminated.

IV. Looking Forward

This section brings together the analyses in sections II and III, and identifies policy choices that MENA countries can pursue to address ongoing and future challenges. Keeping in mind that there is no “one-size-fits-all” approach for the region, we make some general proposals and distill more refined policy choices for each country group, paying particular attention to the relative weights that governments should give to revenue mobilization versus efficiency and equity considerations.

A. Tax Policy Choices for the Future

Experience of tax policy reform in MENA over the past two decades suggests that changes to tax systems have been marginal in most countries, particularly as they relate to tax bases. More fundamental reforms should be considered, that re-evaluate long-standing practices in tax policy, particularly in the area of VAT and excises, personal income taxation, and investment tax incentives. This demands stronger political commitment, transparency in the conduct of tax policy, and a longer perspective. This is not to say that marginal policy changes are not useful,39 but they tend to focus on short-term considerations, and do not address fundamental structural problems in MENA’s tax systems.

Differences in revenue levels across the region, from well above 40 percent of GDP in most resource countries, to moderate or even low in non-resource countries, imply that revenue mobilization should not figure equally on the policy agenda of all groups. For most resource countries, revenue policy should continue to focus on efficiency in managing resource revenues, and addressing equity issues on the spending side of the budget, including better targeted and more cost-efficient transfers to individuals and enterprises.

In resource-rich countries (GCC, OME, and resource Maghreb), tax policy should be directed at building simple and effective tax systems at very low tax rates and broad bases to pursue multiple objectives, including the following:

  • Building institutions and capacity. Establishing now a simple tax system with low rates can be seen as investing in the future through building the capacity to design and administer taxes.40 Ultimately, resource revenues will have to be replaced, at least partially, either because of exhaustibility or long-term commodity shocks. Experience in countries with well established and strong tax institutions shows that this could take decades. A good parallel in the GCC is the time it took to build effective institutions to manage resource revenues; most observers would agree that this was a difficult, lengthy, but valuable investment. Building tax institutions could prove more difficult and may take longer, because it requires stronger political commitment and deals with a larger population of taxpayers than taxing oil companies—e.g. it took Iran two decades to double its non-resource revenues from a little over 2 percent of GDP in 1990 to about 5.5 percent in 2012.

  • Capturing non-resource economic rent. Non-resource taxes, mainly profit-based, can be usefully used to capture part of the rent in the non-resource sector that accrues as a spillover from spending resource revenues. Resource countries invest significantly in non-resource activities such as infrastructure and government procurement of goods and services, where markets tend to be highly regulated (e.g. barriers to entry for foreign investors; licenses; etc.) It can be argued that such activities are to a large extent captive to the presence of the resource rent and regulations, and should be taxed irrespective of revenue needs. The case for taxation is especially strong given the low or weak role played by PIT systems. For example, a tax on business profits would capture part of the significant rent that accrues to individuals in GCC countries.

  • Taxing non-nationals. The composition of the population in many GCC countries, with a very high share of expatriate residents, is another argument why taxation in resource-rich countries may be desirable—the argument is somewhat similar to the previous one, but with rent accruing to resident non-nationals rather than resident foreign companies. However, taxing non-nationals only, such as through a withholding on remittances, is discriminatory and may affect a large number of low-income non-national residents—hence inequitable. Alternative policy options can be contemplated, such as an income tax with a very high exemption threshold.

  • Taxing real property. Oil wealth has been heavily invested in residential and commercial property in the GCC and to a lesser extent the OME region. It is efficient to tax real property, not only because it is immobile, but also to finance the cost of public infrastructure and other services necessary to preserve the value of real property.

  • Selective consumption taxes. Excise taxes on certain widely consumed items, such as soft drinks and motor vehicles, can be both efficient and revenue generating. This is particularly the case in GCC countries, where the broader issue of introducing a common VAT has suffered technical and political setbacks for nearly a decade.

In non-resource country groups, equity and efficiency considerations should figure prominently in tax reforms along with additional revenue needs, with the latter depending on current revenue levels relative to potentials and absorptive capacity to spend additional revenue effectively and efficiently.

In the non-resource Maghreb, tax rates are relatively high and revenue levels exceed 20 percent of GDP, which arguably is close to potential given per capita GDP. These countries should emphasize equity and efficiency in their tax reform, more than revenue mobilization. Policy options for this group include:

  • Reducing the number of VAT rates. Two rates should be sufficient to address, even though imperfectly, the regressive effect of VATs. This, however, should be done with a fundamental review of the use of exemptions and their targeting, with a view to narrowing their application to fewer items than currently provided. Countries may want to consider limiting the use of lower rate and exemptions.

  • More effective and efficient use of excises. This could include higher tax rates on some consumption items, in particular tobacco, alcohol and soft drinks, petroleum products, and motor vehicles. Higher excise rates should absorb other taxes applicable to excised items, such as fees and stamps duties—in some countries (e.g. Tunisia and Algeria), additional excise-like levies are earmarked.

  • Reducing CIT rates. Rates in the range of 15 to 25 percent are probably appropriate. But countries should undertake a fundamental review of the use of tax incentives for investment, in particular tax holidays. Two CIT rates may be justified in some countries, with a higher rate applicable to sectors that exhibit country-specific rent.

  • Restructuring PIT rates and bases. The bracket taxed at zero could be increased in some countries, and the number of rates could be reduced to no more than 4—without any loss of revenue or rate progressivity. This should be combined with a fundamental review of the PIT base, in particular the use of deductions to relieve certain expenditures from the PIT and exemptions of certain types of income (e.g. portfolio and pension income). Considerations should be given to streamlining deductions given their regressive nature, or converting them to tax credits calculated at the lowest positive PIT rate, with appropriate caps to limit the benefit to high income individuals.

In the Mashreq, tax rates seem adequate relative to international and regional practice, and are probably on the low side in some areas (e.g. VAT rate in Lebanon; GST rate in Egypt; PIT rates in Jordan)—this could be positive, especially from an efficiency and tax administration perspective. In these countries, tax revenue levels are also low, generally below 15 percent of GDP, which is probably a reflection (at least in part) of low tax rates. If needed, tax revenues could be increased in the Mashreq, but equity and efficiency issues should figure as prominently in tax reforms as in the non-resource Maghreb. Policy options for this group include:

  • VAT rates could be increased. Additional revenue needs could be met with an increase in VAT rates, which tend to be low relative to regional and international standards. It is however preferable first to streamline exemptions—low tax rates carry significant benefits, including lower distortions caused by exemptions and other non-neutralities in the definition of the VAT base, such as the registration threshold.

  • PIT rates could be increased. This is the case for example in Jordan, where a higher tax bracket could be added to the current structure;41 or, changing the rate structures in Egypt, Lebanon and Syria with higher rates on high incomes. The number of rates in Lebanon and Syria is also unnecessarily high, and could be halved. However, base broadening options should be given priority, as should enforcing the PIT on self-employed and closely-held enterprises.

  • Excises could be used more effectively and efficiently. This is particularly the case in Lebanon, where domestic products are treated more favorably than imports.

  • CIT rates seem appropriate. However, the use of CIT incentives for encouraging investment, in particular tax holidays in Egypt and Jordan,42 should be curtailed. As in Maghreb countries, two CIT rates may be justified with a higher rate applicable to sectors that exhibit country-specific rent.

In addition, all non-resource countries need to consider more carefully the implications for employment and unemployment of labor taxes, including SSCs. The issue could be particularly important when SSCs are considered together with middle and higher PIT rates on particular population cohorts—e.g. women and young educated individuals. Countries should also reconsider rate differentiation and exemptions in the taxation of investment income; they cause distortions in the choice of saving instruments and provide opportunities for high-income individuals to reduce substantially their income tax (or avoid it altogether). Tax policy in this area should aim for low rates and greater neutrality to encourage compliance and facilitate enforcement. Even at low rates, this will likely generate some revenue and improve progressivity of income taxation.

Finally, lowering transfer and registration fees on immovable and movable property (to, say, no more than 2 percent), together with better enforcement should improve transparency with little or no revenue loss. Also, countries should introduce or improve the taxation of recurrent taxes on real property. Stamp duties, particularly those on documents and contracts should be eliminated in most cases (exception to consider include real estate and shares), or consolidated with fee systems where a case can be made for keeping them (e.g. driver’s license and passport fees) and valued at cost or market value.

B. Policy Process and Public Participation

In thinking about the policy process, it is useful to distinguish three elements of tax policy making at the broadest level: tax policy advice, policy interpretation, and tax administration policy.43

Tax policy advice deals with analyzing and evaluating policy ideas to enhance the role that tax policy plays in an economy. Ideas for policy changes or new policies arise from many sources, internal and external to government. Among internal sources, three play a major role: the ministry of finance as the guardian of fiscal policy identifies and estimates revenue sources for the budget; other ministries attempt to achieve policy objectives through the tax system (e.g. health, environment, industrial development, attracting foreign direct investment, etc.); and the revenue administration seek to improve tax compliance by ensuring that tax rules are relatively easy to administer and comply with. Sources external to government include civil society organizations, business associations, the authorities of other countries, and international organizations.

Policy interpretation, a core function of the tax administration, is directed towards interpreting the provisions of existing tax legislation, and providing guidance and direction for taxpayers and revenue administration officials on how the existing tax code and regulations are to be applied in practice in specific situations. In this area, the judiciary plays a key role in interpreting tax laws, and may even influence policy changes.

Tax administration policy deals with the role of the tax administration’s headquarters in relation to tax administration core activities: taxpayer services, returns and payments processing, collection enforcement, and audit.

While MENA countries exhibit to varying degrees weaknesses in the second and third elements, which are beyond the scope of this paper, the first element, a core function of tax policy making, is practically absent. The function of tax policy advice is often relegated to a small number of key individuals, with the consequences that policy proposals are frequently made as legislative proposals with no analytical content on main issues such as: revenue impact; possible economic and social consequences; behavioral response to changes in policy; implications for interactions with other policies. A cursory reading of MENA countries’ proposals for policy changes in their annual budget laws (or other vehicles, such as decrees and decisions) is sufficient to establish the absence of analytical underpinnings to tax policymaking. To our knowledge, none of the countries covered in this paper has established a unit or a structure within the ministry of finance dedicated entirely to the design, monitoring, and evaluation of tax policy. Some countries have made progress (e.g. Morocco with their publication of tax expenditures), but a structured approach to tax policy analysis within ministries of finance remains largely an absent necessity.

One of the consequences of this is that often consultations about policy (if undertaken) happen too late in the process of policy development and does not allow for proper discussion of the implications of policy on various economic sectors and stakeholders. Important steps are skipped in the process, pushing stakeholders to discuss the details of how each article of draft legislation should be framed, how many exceptions should be given and to whom, etc. The end result can therefore be very different from the original intention. This is not to say that similar results could not occur in the presence of solid analytical underpinnings, but the implications of such results would at least be known, including how they differ from the implications of the original scenario.

MENA countries can benefit from the experience of many developed countries in establishing tax policy analysis capacity within their ministries of finance. Such capacity should be able to integrate the analysis of all aspects of policymaking: economic, legal, accounting, and international relations. It should have access to a wide range of macro and micro data available in the country, and build the necessary analytical tools to exploit such data for policy analysis. It should also handle cross-cutting issues in tax policy, such as the tax implications of trade, health, and environmental policies. Tax systems in MENA are as complex as in developed countries—perhaps even more complex given existing administrative and compliance capacities. It is imperative that countries pay more attention to the analytical dimension of policymaking to enhance public participation and improve outcomes.

V. Concluding Remarks

This paper has provided an overview of revenue and tax policy developments in MENA since 1990. While countries differ in their economic and political structures, and the current situation of their tax systems, there are common features among country groups, as defined in this paper, that are useful for identifying some general reform options. Beyond these generalities, however, tax policy options must be grounded in detailed country-specific analyses. At this detailed level, policy options are likely to differ significantly across countries, including in the weights governments may attach to revenue mobilization, versus equity and efficiency issues.

In resource-rich countries (GCC, OME, and resource Maghreb), tax policy should be directed at building simple and effective tax systems at very low tax rates and broad bases. An important step to these ends is building capacity for designing, monitoring and administering modern broad-base taxes.

In non-resource country groups, efficiency considerations should figure prominently in tax reforms, and there exists ample opportunities for equity-enhancing reforms, along with additional revenue needs. The latter depends on current revenue levels relative to potentials, and absorptive capacity to spend additional revenue effectively and efficiently. One of the main differences between non-resource Maghreb and the Mashreq is that tax rates are relatively high and revenue levels arguably close to potential in the first group, while both rates and revenue levels are relatively low in the latter group.

All countries can benefit from more effective and efficient use of real property taxation. Fee rates on transfer of real property should be reduced in a number of countries, and recurrent taxes on real property should be increased or administered more effectively where they exist.

Finally, MENA countries can benefit from the experience of developed countries in establishing tax policy capacity to better integrate the analysis of all aspects of tax policymaking, which in turn is likely to improve the quality of interactions between government and citizens, and policy outcomes.


  • Abed, George T. 1998, Trade Liberalization and Tax Reform in the Southern Mediterranean Region, Working Paper WP/98/49, (International Monetary Fund: Washington, DC).

    • Search Google Scholar
    • Export Citation
  • Barrand, Peter, John Norregaard, and Mario Mansour (unpublished), Establishing a Tax Policy Capacity, Fiscal Affairs Department, International Monetary Fund.

    • Search Google Scholar
    • Export Citation
  • Behar, Alberto, and Caroline Freund, 2011, The Trade Performance of the Middle East and North Africa, Working Paper No. 53 (World Bank: Washington, DC).

    • Search Google Scholar
    • Export Citation
  • Bhattacharya, Rina, and Hirut Wolde, 2010, Constraints on Trade in the MENA Region, Working Paper WP/10/31 (International Monetrary Fund: Washington, DC).

    • Search Google Scholar
    • Export Citation
  • Bird, Richard, and Eric Zolt, 2010, “Dual Income Taxation and Developing Countries,Columbia Journal of Tax Law, 1:174217.

  • Bird, Richard, and Pierre-Pascal Gendron, The VAT in Developing and Transitional Economies, Cambridge, 2007.

  • Besley, Timothy, and Torsten Persson, 2011, Pillars of Prosperity: The Political Economics of Development Clusters, Princeton University Press.

    • Search Google Scholar
    • Export Citation
  • Bräutigam, Deborah, Odd-Helge Fjeldstad, and Mick Moore, eds. 2008, Taxation and State-Building in Developing Countries: Capacity and Consent, Cambridge University Press.

    • Search Google Scholar
    • Export Citation
  • Brun, Jean-François, Gérard Chambas, and Mario Mansour, 2015, « Effort fiscal des pays en développement: une mesure alternative » forthcoming).

    • Search Google Scholar
    • Export Citation
  • Crivelli, Ernesto, and Sanjeev Gupta, 2014, “Resource blessing, revenue curse? Domestic revenue effort in resource-rich countries,European Journal of Political Economy, 35, 88101.

    • Search Google Scholar
    • Export Citation
  • De Mooij, Ruud, and Gaëtan Nicodème, 2007, Corporate tax policy, entrepreneurship and incorporation in the EU, Discussion Paper TI2007-030/3 (Tinbergen Institute, Amesterdam)

    • Search Google Scholar
    • Export Citation
  • De Mooij, Ruud, and Sjef Ederveen, 2008, “Corporate tax elasticities: a reader’s guide to empirical findings,Oxford Review of Economic Policy, 24(4), 680697.

    • Search Google Scholar
    • Export Citation
  • European Commission, 2014, Tax Reform in EU Member States 2014, available online at:

  • Keen, Michael, 2013a, Targeting, Cascading, and Indirect Tax Design, Working Paper WP/13/57 (International Monetary Fund, Washington, DC).

    • Search Google Scholar
    • Export Citation
  • Keen, Michael, 2013b, The Anatomy of the VAT, Working Paper WP/13/111 (International Monetary Fund: Washington, DC).

  • Klemm, Alexander, and Stefan Van Parys, 2012, “Empirical evidence on the effects of tax incentives,International Tax and Public Finance, 19:393423.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2013, Energy Subsidies Reform: Lessons and Implications (Washington, DC).

  • International Monetary Fund, 2012, Fiscal Policy and Employment in Advanced and Emerging Economies (Washington, DC).

  • International Monetary Fund, 2011, Revenue Mobilization in Developing Countries (Washington, DC).

  • James, Sebastian, 2009, Tax and Non-Tax Incentives and Investments: Evidence and Policy Implications (World Bank: Washington DC).

  • Keen, Michael, and Mario Mansour, 2010, “Revenue Mobilisation in Sub-Saharan Africa: Challenges from Globalization II—Corporate Taxation,Development Policy Review, 28(5): 573596

    • Search Google Scholar
    • Export Citation
  • Mansour, Mario, 2014, A Tax Revenue Dataset for Sub-Saharan Africa: 1980-2010, Working Paper I19 (FERDI: Clermont-Ferrand).

  • Mourji, Fouzi, 2011, L’incidence fiscale au Maroc–Cas de la TVA. Presentation at the FERDI conference on taxation and development, October 24-25, 2011; downloaded from on December 16, 2014.

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, 2014, The Disbtributional Effects of Consumption Taxes in OECD Countries, Tax Policy Studies No. 22 (Paris).

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, 2006, Fundamental Reform of Personal Income Tax, Tax Policy Studies No. 13 (Paris).

    • Search Google Scholar
    • Export Citation
  • Van Parys, Stefan, and Sebastian James (2010), “The effectiveness of tax incentives in attracting investment: panel data evidence from the CFA franc zone,International Tax and Public Finance, 17:400429.

    • Search Google Scholar
    • Export Citation
  • Silva, Joana, Victoria Levin, and Matteo Morgandi, 2012, Inclusion and Resilience—The Way Forward for Social Safety Nets in the Middle East and North Africa (World Bank: Washington DC).

    • Search Google Scholar
    • Export Citation
  • Sorensen, Peter Birch, 2005, Dual Income Taxation: Why and How? CESifo Working Paper No. 1551.

  • World Trade Organization, International Trade Centre, and United Nations Conference on Trade and Development, 2014, World Tariff Profiles 2014, downloaded from

    • Search Google Scholar
    • Export Citation
  • World Bank, 2015, MENA Development Report-Jobs or Privileges: Unleashing the Employment Potential of the Middle East and North Africa (Washington, DC).

    • Search Google Scholar
    • Export Citation
  • Zee, Howell H., 2005, “Personal Income Tax Reform: Concepts, Issues, and Comparative Country Developments,” working paper 05/87 (International Monetary Fund: Washington DC)

    • Search Google Scholar
    • Export Citation

Appendix I. The Dataset: Sources and Description

The main source used for building the revenue dataset is IMF staff reports and statistical appendices, produced by IMF surveillance and program missions. The methodology follows closely that described in Mansour (2014).

Total revenues: This is the total of all tax and non-tax revenues. Where detailed data were available, certain stamp duties and registration fees, which can be non-negligible, were recorded as taxes (mainly real property registration and stamps on financial transactions and insurance contracts), under Other Taxes; otherwise, they were reported under Non-tax Revenues.

Direct Taxes: These include taxes on all income sources (i.e. business profits, wages, portfolio income, and capital gains, etc). Social security contributions are not included, since they are not paid into the consolidated fund. This is a significant weakness of the dataset since SSCs can be substantive both in terms of revenue and their incentive effects on employment.

Direct Taxes are not always the total of corporate and personal income taxes. The difference is due to taxes that cannot be allocated to either of these two sub-categories; these include recurrent taxes on real property, which are insignificant in MENA countries, and taxes on portfolio income that are not reported as part of individual or corporate income.

Corporate Income Tax: This is the tax imposed on corporate income in countries that have a corporate tax. This can include taxes on profits from individual enterprises or any other business entities liable to a schedular profit tax in countries that have schedular tax systems. Data to split the profit tax according to the legal status of the business entity are not usually available in IMF staff reports.

A great deal of effort went into separating from the corporate tax, tax revenues from profits of corporations and other business entities from extractive activities in the oil and gas and mining sector. These were included in the series Resource Taxes.

Personal Income Tax: This is the tax imposed on income of individuals from all sources in countries that apply a comprehensive income tax on individuals, including tax on business profits. In countries that apply a schedular income tax, the Individual Tax includes primarily taxes on wages (other than SSCs, which as noted earlier, are not paid into the consolidated fund).

Indirect Taxes: Indirect taxes were separated between General Sales Taxes, Excise Taxes, and a residual that could not be accounted for. VATs are included in indirect taxes on a net basis, where data on refunds are available. Some excise-like taxes could not be accounted for separately, and are included in the category Other Taxes.

Trade Taxes: These include the following: (1) revenues from a country’s external tariff on imports and exports; and (2) ad-valorem service fees. It is not always possible to distinguish the various fees applicable—it can be argued that certain fees that do not finance the budget of the country that levy them should not be accounted as own revenue.

Another category that may be included is withholding on imports for purposes of income taxes. This is applied in some countries as a way to tax the informal sector; it is levied on imports, but can be deducted against a taxpayers’ income tax liability (tax on income from a business activity). Revenues from this withholding are accounted for as direct taxes (levied by customs on behalf of the domestic tax administration).

Resource Taxes: These include primarily revenues from the Corporate Income Tax that are collected from extractive activities in the oil and gas, and mining sectors, and royalties. Revenues from production sharing agreements were included in this series to the extent that they were paid into the government general fund—which typically occurs in the case of oil and gas through a national oil company.

Other variables: The dataset contains a number of other variables that are used in the trend analysis. These are the following: GDP at current prices; population; annual average inflation (as measured by the consumer price index); income classification. The source of all these variables is the IMF World Economic Outlook database, and in some cases the World Bank development indicators database.

Appendix II. List of MENA Countries and Group Classification

The countries included in this study (19) are listed below along with the five group classification. Maghreb countries are separated in two groups, one belonging to the resource-rich group (Algeria and Libya) and the others to the non-resource group.

Countries are selected according to the IMF Middle-East and Central Asia (MCD) Department coverage, which follows closely (but not completely) geographical location. Hence, Mauritania, which belongs to MCD, is included, but not Israel and Turkey. Also, Djibouti and Soudan, which belong to MCD, were excluded due to data limitations.

article image
Note: GCC: Gulf Cooperation Council; UAE: United Arab Emirates.

MENA: VAT Main Exemptions under Current Laws

article image

MENA: Key Features of Investment Tax Incentives

article image
Sources: Compiled from IMF staff reports and other documents, International Bureau for Fiscal Documentation, and countries’ laws.

MENA: Summary of Main Fees and Stamp Duties

article image
Source: Internatinal Bureau for Fiscal Documentation.

Total revenues are equal to resource revenues plus non-resource revenues, with the latter equal to tax and non-tax revenues. Annex I provides further detail on the definition of the main categories used throughout the paper.


IMF (2013) estimated energy subsidies in the MENA region for 2001 to account for about 50 percent of worldwide total, 8.5 percent of regional GDP, or 22 percent of total government revenues. These estimates are based on pre-tax international prices. IMF (2013) also reports estimates based on post-tax prices, with taxes reflecting the environmental and health damage of energy consumption. Whether subsidies should be measured on a pre- or post-tax basis is debatable, and depends on government policy objectives.


This is true whether total GDP or non-oil GDP is used as a scaling variable.


Non-resource taxes can dampen the volatility of total revenues when their tax bases are not strongly linked to commodity prices, and when their share in total revenues is important.


The left-hand panel allows for easier international comparisons, which are available mainly on the basis of total GDP. Using non-oil GDP provides a broadly similar picture.


Econometric analysis suggests a strong relationship between the non-resource tax effort (expressed in percent of GDP) and the log of per capita GDP. See, for example, Brun, Chambas, and Mansour (forthcoming), and Crivelli and Gupta (2014).


One has to be cautious in interpreting these differences in levels since off-budget fiscal activities may vary significantly across countries, including, for example, from sovereign wealth funds and state enterprises.


The correlation over the entire sample period between resource and tax revenues is -0.73 for resource countries as a whole, but +0.3 for the OME group. The latter goes against recent econometric findings.


This is despite analyses that clearly indicated the fiscal (and other) challenges that trade liberalization would create for these countries (see Abed, 1998).


An important recent development is the increase in the number of armed groups challenging the authority of the state, and in some cases establishing their own tax systems.


This is the ratio of total trade taxes divided by the value of imports.


The ratio of trade taxes-to-GDP (or any tax-to-GDP ratio) is the product of two ratios: a backward looking effective tax rate calculated as revenues divided by the respective tax base (in this case imports); and the tax base as a share of GDP. This decomposition of the tariff-to-GDP ratio allows for a better understanding of how revenues (and hence the effective tax rate) evolved over time relative to the base.


Naturally, the response of imports to tariff rate reductions and revenue implications differ across goods.


We do not show the resource groups separately since there are no significant differences among them in relation to this issue, with the exception of Iran, where tariff revenues increased over the sample period from about 0.5 percent of GDP to 1.5 percent.


The Egyptian Goods and Services Tax acts as a VAT on goods, but only a partial VAT on services.


Algeria recently increased its VAT registration threshold to 30 million dinars (roughly USD 310,000, valued at exchange rates prevailing at end of March 2015), in an effort to simplify the taxation of small businesses, for VAT and income tax purposes.


This is the ratio of actual VAT revenues to potential revenues, estimated as the product of the standard VAT rate and final consumption taken from National Accounts statistics.


Revenue productivity is calculated as actual VAT revenue divided by the product of the standard tax rate and GDP. C-efficiency is an estimate of the overall tax gap of the VAT (i.e. actual revenues as a share of potential revenues, estimated based on aggregate final consumption). Both measures reflect policy and compliance issues that are hard to disentangle without detailed data on the composition of VAT revenues and the tax base. For an overview of measurement issues with C-efficiency, and how it can be decomposed into policy and compliance gaps, see Keen (2013). Given these issues, it is prudent to take the C-efficiency estimates in Table 3 as rough indicators of the total revenue gap, without giving them precise normative meaning.


Lack of data is a key reason for this, but deeper political issues may be at play as well; the provision of VAT tax expenditures tend to be a powerful political tool among the poor, whose number in developing countries (and in most MENA countries with a VAT) is very high.


For example, Table A2.8 in European Commission (2014) shows that between 2000 and 2014 the number of VAT low rates in the EU increased, despite research showing that such policies are not cost-effective in targeting support to the poor.


Data on excise revenues are not available from IMF staff reports on a consistent basis across countries and over time. In some countries, like the Maghreb, excises are still levied on a large number of goods (in some cases exceeding 50). Data shown here is our best attempt to provide the magnitude of key excises, which also tend to be the main contributors to this revenue category. See Appendix I for more detail.


Here as elsewhere in this paper, we ignore the taxation of profits from upstream oil and gas. CIT revenues from such activities, along with other taxes specific to the sector, are included in resource taxes. Also, for countries which had or still have a schedular tax system, the CIT is applicable to industrial and commercial activities earned by corporations; corporations may be taxed at different rates on other sources of income.


IMF (2011) estimates the average (unweighted) contribution of the CIT, including the resource sector, at 10 percent for OECD countries (pre-2008 crisis) and 17 percent for developing countries (in this calculation, CIT revenues from resource activities are included but not other revenues, such as royalties and production sharing). Mansour (2014), using a measure of CIT revenues similar to the one in this paper (i.e. excluding the resource sector), estimates its share in Sub-Saharan Africa at about 18 percent of tax revenues in 2010.


Devereux and Loretz (2012) review the empirical literature on CIT competition and conclude that it is far from clear that the significant reduction in CIT rates around the world since the early 1990s is due to strategic interactions between countries. Nevertheless, practical experience suggests that policymakers attach significant importance to headline tax rates in neighboring countries and (perhaps to a lesser extent) elsewhere.


The implicit CIT base is the base implied by revenues collected from the CIT, and is used as a proxy to the actual taxable base. If P/GDP is the actual tax base as a share of GDP, the implicit tax base is (c*P)/(c*GDP), where c is the corporate tax rate. This can be rewritten as (C/GDP)*(1/c), where the first term is potential CIT revenues as a share of GDP, which we approximate by actual revenues to GDP. The weakness of such a measure is the difference that can exist between potential and actual CIT revenues due to compliance factors.


De Mooij and Nicodème (2007), for example, find that income shifting from the personal to the corporate tax occurred in the EU as CIT rates declined relative to PIT rates starting in the early 1990s.


Because taxation of dividends is light, porous, or inexistent in MENA, double taxation of corporate profits is not an issue.


Starting in 2014, Jordan’s standard CIT rate increases to 20 percent, and the telecom rate decreases to 24 percent. The rate on banks remains at 35 percent and the rate on industry at 14 percent.


De Mooij and Ederveen (2008) review the literature on the sensitivity of the corporate tax base to tax rates, and find that it varies according to a number of dimensions. For example, profit shifting is more sensitive to differences in the CIT rate than real investment decisions, which are more sensitive to effective (marginal and average) tax rates. They conclude that on aggregate the semi-elasticity of the corporate tax base with respect to the tax rate is -3.1—i.e. one point increase (decrease) in the CIT rate reduces (expands) the base by 3.1 percent.


MENA is not unique in this regard; see Keen and Mansour (2010) on Sub-Saharan Africa, and James (2014) more generally.


For a review of effectiveness and efficiency of tax incentives for investment, see IMF, World Bank, OECD and UN (2015), Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment, forthcoming.


The Moroccan tax expenditures report is annexed to the annual budget law, with estimates for the year preceding the budget year.


See de Mooij and Ederveen (2008) and James (2009) for a review of the empirical literature. In a unique study on tax holidays in Latin America and the Caribbean, and Africa, Klemm and Van-Parys (2012) find that tax holidays have some impact on FDI in the former region, but none in the latter, suggesting that it is not always clear that the strategy of providing tax holidays work to attract FDI.


Van Parys and James (2010) find that tax factors may be important to investment when countries are competitive in non-tax factors—suggesting that tax incentives cannot be used as substitutes for the absence or weakness of other (non-tax) factors.


Dual income tax systems combine a progressive tax rate on wage income with a flat tax rate on capital income, set at the lowest tax rate on wage income. For a review of dual-income taxation in developing countries, and experience in developed countries see Bird and Zolt (2010). Sorensen (2005) is also a useful review of design issues with dual-income tax systems.


See discussion in Zee (2005).


One of the roles usually attributed to the CIT is a withholding tax for the PIT. Since corporate profits are ultimately distributed to individual shareholders as dividends, it is sometimes argued in the literature that dividends should be exempt from tax in order to prevent double taxation. However, to the extent that corporate profits are exempt from the CIT under tax incentives (as discussed earlier), or that they can be paid to shareholders in other lightly-taxed forms (interest or capital gains), the role of the CIT as a backstop to the PIT is questionable.


For a review of the literature, see IMF (2012).


Examples of marginal changes include: recent increases in excises on tobacco in Tunisia and Egypt, and on petroleum products in Lebanon; increase in the CIT rate from 20 to 25 percent in Egypt, which reverses some of the large rate cut (from 40 to 20 percent) undertaken in 2006.


Besley and Persson (2011) refer to this as fiscal-capacity investments, an integral part of state building. On the relationship between taxation and state building see for example Bräutigam, Fjeldstad, and Moore (2008).


In its 2015 budget, Jordan restructured its PIT by adding a third rate of 20 percent, and lowering the amount of exempt income (taxed at zero percent). These changes would apply to taxation years starting in 2015.


Jordan increased the CIT rate to 20 percent, effective for taxation years 2015 onward.


For further detail, see Barrand, Norregaard, and Mansour (2008).

Tax Policy in MENA Countries: Looking Back and Forward
Author: Mario Mansour