Information about Sub-Saharan Africa África subsahariana

VI Macroeconomic Aspects of Trade Liberalization

Arvind Subramanian
Published Date:
September 2000
  • ShareShare
Information about Sub-Saharan Africa África subsahariana
Show Summary Details

This section deals with two questions: whether initial macroeconomic conditions influenced the extent of trade liberalization during the 1990s, and whether trade liberalization was associated in identifiable ways with the macroeconomic performance of ESA countries.26 The description below does not purport to assign causality: there is the customary difficulty of isolating the impact of macroeconomic conditions on trade liberalization from other factors, and vice versa. The analysis aims at highlighting the stylized aspects of the relation between macroeconomic conditions and trade liberalization, drawing upon the experience of individual countries.

Countries are classified into two groups according to the level of trade liberalization achieved at the end of 1998.27 Countries that by 1998 had lowered their index of trade restrictiveness to 4 or less are classified as part of the “substantial liberalization” group (i.e., the bottom two categories in Table 3.1, also referred to as fast reformers), while the remaining countries are placed in the “moderate-or-no-liberalization” (i.e., the top two categories in Table 3.1, also referred to as slow reformers) group.

Initial Macroeconomic Conditions

Initial macroeconomic conditions do not appear to have played a decisive role in ESA governments’ decisions to liberalize trade (Zimbabwe is one country where macroeconomic factors may have played a role in slowing, even reversing, liberalization). Countries in the moderate-or-no-liberalization group did not display consistently larger macroeconomic imbalances in the early 1990s. At the same time, a number of ambitious reformers—Uganda, Zambia, Mozambique—faced severe macroeconomic imbalances in the early 1990s, with inflation rates varying between 36 and 135 percent and very high fiscal and current account deficits. Far from constraining reform, these imbalances may even have provided a spur to the reform process.

The substantial liberalization group had larger initial current account deficits than the slow liberalizers, but managed to improve their external position over the 1990s. The average current account deficit declined by about 4 percentage points for the former set of countries, but deteriorated by a further 2 percentage points for the latter. This may have resulted in part from better exchange rate management. Countries that liberalized, on average, saw their real effective exchange rate depreciate by about 6 percent (cumulatively) during the period 1990–98, while that of the slow reformers appreciated by about 11 percent. The six countries—Angola, Congo, Eritrea, Kenya, Rwanda, and Tanzania—that experienced the most adverse movements in the exchange rate were in the group of slow liberalizers. Thus, to some extent, a depreciating exchange rate in liberalizing countries may have averted or mitigated short-term pressures on their external position.

Trade Liberalization and Macroeconomic Outcomes

Even though not every country in the “substantial liberalization” group had a superior economic performance during the 1993–99 period, the figures show that these countries have, on average, fared much better than the countries in the “moderate-or-no-liberalization” group. This is particularly the case for GDP growth, investment ratios, and fiscal and inflation performance (Appendix II, Table A18). In the case of South Africa, there is evidence that trade liberalization contributed to the dynamic efficiency of the economy, as reflected in faster growth of total factor productivity (Box 6.1).

Box 6.1South Africa: Dynamic Gains From Trade Liberalization

South Africa undertook significant trade liberalization in the 1990s and has experienced a large increase in the openness of its economy. The impetus for liberalization started gaining momentum in the late 1980s, reflected in a consultative process under the auspices of the tripartite National Economic Forum involving government, labor, and organized business. As a result, South Africa adopted a two-pronged approach to trade liberalization, consisting of multilateral trade liberalization in the context of the Uruguay Round of trade negotiations and unilateral trade liberalization. Neither fiscal nor macroeconomic considerations influenced trade reform in South Africa. Trade taxes accounted for a very small share of total revenue—and in any event much of the revenue derived from trade was transferred to SACU partners. Furthermore, inflation steadily declined during the 1990s while the external position—although subject to episodic shocks—was generally sound.

Virtually all quantitative restrictions have now been eliminated, including those operating through agricultural marketing boards. The tariff regime has been rationalized, with the number of lines reduced from over 13,000 in 1993 to about 7,900 in 1998 and the number of tariff bands reduced from well over 100 to 72. These data mask the progress made in reducing the actual complexity of the tariff regime, as the number of lines carrying formula duties (that acted like variable import levies) was reduced from 1,900 in 1993 to 28 in 1997 and the number of lines facing specific tariffs was reduced from 500 to 227. The standard deviation of tariff’s declined from 43 to about 15, while the range of effective protection was also narrowed.

In a recent study, Jonsson and Subramanian (2000) show that South Africa’s increased openness and the recent trade liberalization have had a beneficial impact on dynamic efficiency in the 1990s, in particular on growth in total factor productivity (TFP). The study uses both a time-series approach at the aggregate level and a cross-section approach across different manufacturing sectors. Both approaches yield results supportive of a strong and statistically robust relationship between trade and total factor productivity. Furthermore, this improvement in overall efficiency was not a consequence of reduced employment levels.

The time-series analysis yields a co-integrating relationship between TFP. openness (defined as the sum of exports and imports of goods and nonfactor services divided by GDP), and the share of machinery and equipment in total investment, with the evidence suggesting that causality runs from openness to TFP. On average, a 1 percentage point increase in openness results in a long-run increase in annual TFP growth of about 0.5 percent. Similarly, an increase in the share of investment in machinery and equipment of 10 percentage points is associated with higher TFP growth of about 3 percentage points. Furthermore, both the openness and the investment in machinery and equipment variable are important for the co-integration result, indicating that the two sources complement each other in influencing productivity. Although research and development (R&D) augments productivity, it is also important to provide an open and liberal environment in which gains from R&D can be maximized.

In the cross-section analysis, the trade variable is measured in terms of domestic policy liberalization and the results indicate that those manufacturing subsectors that witnessed greater trade liberalization during the 1990s achieved higher TFP growth. On average, a 10 percent reduction in prices induced by trade liberalization in a sector, led to a 3 percent increase in its annual TFP growth.

Typically, countries in the “substantial liberalization” group embraced a wider range of economic reforms during the 1980s than countries in the other group. This is important insofar as trade liberalization is likely to be just one of the factors affecting economic performance, and trade reforms are likely lo be more successful in an environment where a comprehensive set of reforms is in place or being implemented. Mozambique (Box 6.2) and Uganda provide positive evidence on the impact on growth of bold and comprehensive reforms. Countries that had problems implementing comprehensive reforms experienced macroeconomic instability that undermined any positive effects of trade liberalization on economic growth (for example, Zambia, Box 6.3).

Per capita GDP growth was also—on average—higher in the “substantial liberalization” group than in the “moderate-or-no-liberalization” group (Figure 6.1). Growth in export volumes was also positively correlated with trade liberalization when countries affected by war were excluded, and current account balances in most liberalizing countries improved, while the opposite was true for countries with little or no liberalization (Figure 6.2).

Figure 6.1.Eastern and Southern Africa: Trade Liberalization and Economic Performance

Figure 6.2Eastern and Southern Africa: Trade Liberalization, Fiscal Performance, and External Balance

(Changes in percent of GDP)

Mauritius is an interesting case because it registered high rates of export and GDP growth despite having a restrictive trade regime. The lessons of the Mauritian experience of “heterodox opening.” however, are not clear, given the arguably unique configuration of internal and external factors that contributed to Mauritius’ growth performance (Box 6.4). The replicability of the Mauritius model of export processing zones is also unclear given their limited success as a means of accelerating exports and growth in other countries (Box 6.5).

Box 6.2Mozambique: Successful Trade Liberalization

Mozambique’s trade liberalization began in the late 1990s and gathered momentum in the mid–1990s. The reforms undertaken—within the context of programs supported by the Enhanced Structural Adjustment Facility—involved the removal of quantitative import and export restrictions, the simplification of the tariff structure, the narrowing of tariff dispersion, the curtailment of exemptions, and the reduction of import and export taxes (see below). By the end of 1997, all significant non-tariff barriers had been eliminated, and only standard prohibitions relating to firearms and public health remained.

In 1996, the number of tariff rates was reduced from 12 to 5. and the average import-weighted tariff rate was lowered from 18.4 percent to 11 percent. A further reduction in the top tariff rate from 35 percent to 30 percent was implemented in April 1999. The current tariff structure is as follows: 0 percent on essential goods, 2.5 percent on raw materials, 5 percent on capital goods. 7.5 percent on intermediate goods, and 30 percent on consumer goods. Remaining distorting factors are generous exemptions under the investment law and import surcharges on cement, steel tubes, and sugar.

The government plans to reduce the top import tariff rate to 25 percent by January 2002. A ban on exports of raw cashews (Mozambique’s second-largest export product) was replaced by a 40 percent export tax in 1995. The tax was subsequently reduced to 14 percent in 1996, although a new law was passed in October 1999 stipulating an increase to 18–22 percent. No other export taxes are in place.

Trade liberalization has been complemented by a comprehensive package of reforms undertaken during the last five years. These reforms comprised the development of a medium-term expenditure framework, a civil service reform, the rationalization of the tax system, the liberalization of interest rates, the introduction of interbank foreign exchange and money markets, and the privatization or restructuring of over 1,300 state-owned enterprises. The results obtained so far have been impressive: substantial investment has been channeled toward import substitution and nontraditional export activities, and the investment-to-GDP ratio rose from under 15 percent in the early 1990s to an average of 25 percent during 1997–99. Economic growth averaged 7 percent a year during 1993–95 and 10 percent a year during 1996–99. and inflation fell from over 50 percent in the early 1990s to single-digit levels in 1999.

Box 6.3Zambia: Bold Liberalization but Low Growth

In the 1990s, Zambia implemented an ambitious macroeconomic stabilization and structural adjustment program, which included privatization, liberalization of the agricultural sector, and trade reform. Nevertheless, growth rates remained low owing to poor implementation of complementary policies and an unfavorable external environment. A major first step toward a more open trade system was taken in 1990 with a reduction in the maximum tariff from 100 percent to 50 percent. Following a phased reduction in nontariff barriers, most of the exchange restrictions on current international transactions were eliminated in 1994, and the dispersion of tariffs was further narrowed by raising the minimum rate to 20 percent while lowering the maximum rate from 50 percent to 40 percent. In 1995, there was some reversal of trade liberalization with the introduction, for fiscal reasons, of an across-the-board 5 percent import declaration fee (IDF). The effects of the IDF were, however, mitigated by the elimination of the 20 percent uplift factor applied to the sales tax in the same year. Further trade liberalization took place in 1996 when exemptions for government imports were abolished, and the average level of tariffs was brought down considerably through a 15 percent cut in most tariffs. The remaining nontariff barriers were removed in 1996 (except those kept for environmental, health, and security reasons), and the IDF was eliminated in 1998. Zambia now has a tariff regime with four rates, ranging from 0 to 25 percent, a statutory average rate of 14 percent, and specific rates on some products.

Despite some progress on macroeconomic stabilization, which was reflected in the reduction in fiscal deficits and inflation since the early 1990s, economic growth has remained subdued, mainly because of repeated droughts and delays in the privatization of the copper mines, which play a key role in the economy. Nonetheless, the liberalization effort has pushed the economy into a transformation process that has created favorable growth opportunities for a large number of enterprises—often newly established after 1993 or successfully restructured following privatization—as evidenced by the strong growth of nontraditional exports since the beginning of the decade. In 1992, metal exports—mainly copper and cobalt—constituted some 90 percent of total exports. However, reflecting generally weak copper prices and the falling productivity of the state-owned mining company, the total value of metal exports steadily declined in the following years from some US$1 billion in 1992 to only US$574 million in 1998. During the same period, nontraditional exports almost tripled, rising to US$288 million. The strongest growth, albeit from a low base, was recorded in floricultural and horticultural products, primary agricultural products, processed foods, engineering products, and textiles. As mentioned in Section V, Zambia was one of the few countries in sub-Saharan Africa that witnessed significant diversification of its exports.

Box 6.4.Mauritius: Heterodox Opening

At the beginning of the 1990s, Mauritius had a highly complicated trade regime, ridden with exemptions. There were 60 tariff bands, the top rate was about 600 percent, and the average tariff rate was about 80 percent. Some liberalization was implemented during the 1990s, beginning with the reduction of quantitative restrictions in 1991. By 1994, the number of nonzero rate bands was reduced to 7, and the maximum tariff rate was lowered to 80 percent. The export duty on sugar and all controls on exports were also eliminated, except for some items that still require export permits or prior approvals (for example, tea, spices, and pharmaceuticals). In July 1998, however, an additional rate band of 10 percent was added. There are also additional import excise levies ranging from 15 to 400 percent. In 1998, the average tariff was about 19 percent.

Despite these reforms, Mauritius’ import regime remains restrictive (with a restrictiveness rating of 7 out of 10), as reflected in its still substantial QRs, high and dispersed tariff rates, and complexity. Another significant feature of the trade regime is the high level of unclassified exemptions, exemptions to the EPZ and other investment incentive schemes, and concessions by ministries. The restrictiveness of the import regime has not translated in a tax on the two major export products—sugar and textiles—which have been sheltered from the effects of import protection through the export processing zones and by government measures that until recently prevented spillovers of wage pressures from the importable to exportable sector. In addition, preferential quotas in export markets for sugar and textiles have ensured high returns (and rents) to exportable that have helped finance higher investment levels. It remains to be seen whether such a strategy will be sustainable in a changing external environment where rents from export quotas begin to decline and where domestic wage spillovers become increasingly difficult to prevent.

Fiscal Aspects of Trade Liberalization

A recent study (Ebrill, Stotsky, and Gropp,1999) examines the theory and experience of the revenue implications of trade reform. On the theoretical side, some important conclusions can be drawn. First, revenue is less likely to be affected by trade reform when the initial position is highly restrictive; when trade liberalization involves the tariffication and/or elimination of quotas, and elimination of exemptions, simplification, and reduction of tariff dispersion; when there are simultaneous reforms of customs and tax administration; and when trade liberalization is supported by macroeconomic policies that maintain external balance and avoid a concomitant real appreciation of the liberalizing country’s currency. Second, beyond a certain point, and even allowing for the above factors, a reduction of tariffs will lead to a decline in the importance of trade tax revenue, highlighting the need for domestic tax reform to enhance the domestic tax base.

Box 6.5.Export Processing Zones

In early 1998, the following ESA countries had in place legislation for export processing zones (EPZs): Botswana, Kenya, Madagascar, Mauritius, Mozambique, Namibia, Uganda, and Tanzania (Madani, 1999). Although the features of the various EPZs in the region differ from country lo country, most of them have some key elements in common: (i) generous corporate income tax exemptions, with tax holidays of up to 10 years and concessional lax rates afterward; (ii) dutyfree access to inputs and capital goods; and (iii) relaxed labor laws. In some countries, export processing firms that are not located in designated geographical areas are also eligible for the incentives. Despite their appeal, however, EPZs have major disadvantages, including potentially large tax revenue losses and leakages, distortive impact on the competitiveness of domestic Firms that do not benefit from the incentives, and ample scope for abuses. As a result, their overall success in the ESA region and in sub Saharan Africa in general has been rather limited. The anticipated gains in foreign direct investment, production, and employment in EPZs have not materialized in most countries, with the exception of Mauritius, where the establishment of an EPZ in 1971 contributed to a substantial increase in exports of clothing and textiles. In addition, the development of EPZs in most ESA countries has been negatively affected by implementation problems (notably weak and/or nontransparent administration) and the slow progress in some of the key areas of adjustment, including macroeconomic stabilization and infrastructure development. The experience of these countries underscores the view that sound economic policies, coupled with a strong structural adjustment effort (including trade liberalization) are a first-best policy, and that special tax incentives may not be able to offset the negative impact of an unfavorable and unstable policy environment.

What has been the experience of the ESA countries that have undertaken significant trade and tariff liberalization? Three facts stand out:

  • First, at the start of the reforms, ESA countries (with the exception of SACU countries)28 depended to a great extent on trade taxes for their total revenues. In 1990, the trade tax revenue to GDP ratio of ESA countries was 5.7 percent, significantly higher than that for other groups of developing countries (Table 6.1). On average, the slow reformers were considerably more dependent on trade taxes than the fast reformers (Table 6.2), This dependence on trade taxes may have conditioned the nature and extent of trade liberalization. In the case of South Africa, fiscal considerations did not play a significant role in the decision to liberalize trade because of the country’s limited dependence on customs duties for revenue generation (customs duties accounted for just about 4 percent of total revenue, of which a substantial part was transferred to BLNS countries under existing revenue-sharing arrangements). In the case of Malawi, trade liberalization in the mid-1990s was reversed because of a deteriorating fiscal situation, which, in turn, led to the imposition of an export tax. Similarly, in Tanzania, extensive leakages stemming from weak administration and rampant exemptions led to a slowdown of trade liberalization and efforts to address the tax policy and tax administration problems.
  • Second, there was a substantial reduction in the reliance on trade taxes during the course of the 1990s. On average, trade taxes decreased from 34 percent to 23 percent of total revenue, with a broadly similar rate of decrease in both fast and slow reformers. But total tax revenue-to-GDP ratios remained rather stable during the 1990s, reflecting efforts by ESA countries to reform their domestic tax system and broaden the base of trade-neutral taxes. The impact of fiscal reform has varied within the ESA countries. In the case of Tanzania, dependence on revenue from trade taxes increased, partly as a result of efforts in the latter part of the 1990s to strengthen customs administration; the outcome, however, also reflected a more modest effort of tariff liberalization. In Zambia, although tariff reform proceeded at a rapid pace, reliance on trade taxes did not diminish substantially because there was a concomitant effort to tackle the problem of pervasive exemptions.
  • Third, on average, trade liberalization is associated with an improvement in the overall fiscal position in 1993–99 (Figure 6.2). Two-thirds of the countries in the substantial liberalization group had improved their fiscal balance since the early 1990s, despite a reduced reliance on trade taxes, pointing to the success of offsetting tax and expenditure measures that were part of broader programs of fiscal consolidation.

Table 6.1Selected ESA Countries: Tariff Revenue(Percent of GDP)
Country196019851990Last Available Year1
Congo, Dem. Rep. of31.943.173.261.33
South Africa0.720.730.980.18
Average (Sub Saharan Africa)24.825.815.205.39
Average (Eastern and Southern Africa)4.006.405.725.73
Non-OECD Asia and Pacific2,43.153.534.163.42
Middle East24.183.953.273.48
Non-OECD Western Hemisphere 22.782.512.432.36
Sources: IMF, Government Finance Statistics, various issues, and World Economic Outlook (October 1997); and OECD, Revenue Statistics, various issues, except as noted.

Last year for which data are available is 1996

Data are unweighted averages

Data provided by the country authorities; and IMF staff estimates

Korea joined the OECD in December 1996.

Excluding the Czech Republic, Hungary, Luxembourg, and Poland. The OECD data for the countries of the European Union (EU) are for tariff collections at the borders of those countries and accordingly differ from the tariff collections on goods used in each EU country to the extent that there is transshipment within the union.

Sources: IMF, Government Finance Statistics, various issues, and World Economic Outlook (October 1997); and OECD, Revenue Statistics, various issues, except as noted.

Last year for which data are available is 1996

Data are unweighted averages

Data provided by the country authorities; and IMF staff estimates

Korea joined the OECD in December 1996.

Excluding the Czech Republic, Hungary, Luxembourg, and Poland. The OECD data for the countries of the European Union (EU) are for tariff collections at the borders of those countries and accordingly differ from the tariff collections on goods used in each EU country to the extent that there is transshipment within the union.

Table 6.2Eastern and Southern Africa: Trade Tax Revenue as a Share of Total Revenue1(Percent)
Moderate or no liberalization
Substantial liberalization
South Africa84
Sources: IMF, African Department database, and World Economic Outlook database.

The averages exclude Botswana, Lesotho, Namibia, and Swaziland.

Sources: IMF, African Department database, and World Economic Outlook database.

The averages exclude Botswana, Lesotho, Namibia, and Swaziland.

    Other Resources Citing This Publication