Building Integrated Economies in West Africa
Chapter

Chapter 25. Macroeconomic Implications of Tariff Changes

Author(s):
Alexei Kireyev
Published Date:
April 2016
Share
  • ShareShare
Show Summary Details
Author(s)
William Gbohoui, Karim Barhoumi, Qiang Cui and Monique Newiak 

The common external tariff (CET) for the Economic Community of West African States (ECOWAS), which includes all West African Economic and Monetary Union (WAEMU) countries, became effective in January 2015. The macroeconomic effects of a replacement of the WAEMU CET with the ECOWAS CET on WAEMU member states could be substantial. The new tariff could have an impact on trade levels, government revenue, and growth. It is likely to increase WAEMU imports from other ECOWAS countries but reduce imports from the rest of the world. We estimate that the tariff’s impact on government revenue will vary widely by country, ranging from a loss in most countries by up to 2½ percent of current revenue to an increase in a few countries by up to 3 percent. Finally, an increase in relative import prices induced by an increase in the average tariff rate could yield a negative effect on GDP.

Regional Trade and Tariffs

The CET for the ECOWAS was adopted at a Heads of State Summit in October 2013 and became effective in January 2015. As a consequence of this CET, WAEMU countries, a subset of ECOWAS countries, will be subject to a new tariff structure. Currently WAEMU countries (all members of the ECOWAS), impose tariffs within the range of 0 to 20 percent on goods from all non-WAEMU countries, including ECOWAS countries, with a simple average import tariff of about 12 percent. After adopting the CET, WAEMU countries will eliminate tariffs on goods from all ECOWAS countries, but tariffs on products from non-ECOWAS members will increase. The revenue implications of this policy change could be significant, as many WAEMU countries rely heavily on import duties (Figure 25.1, panel 3).

Figure 25.1.Stylized Facts: Trade within ECOWAS

Sources: Country authorities; Direction of Trade Statistics; World Trade Organization Tariff Analysis Online Facility; IMF staff estimates.

Note: ECOWAS = Economic Community of West African States; WAEMU = West African Economic and Monetary Union; RECOWAS = rest of ECOWAS, i.e., ECOWAS without WAEMU member states.

While the literature on trade agreements in West Africa is vast, the revenue effects of implementation of the CET have not been quantified. This chapter contributes to the existing research by estimating the revenue implications of the implementation of the ECOWAS CET on WAEMU member states. First, the chapter takes stock of the current tariff structure in the WAEMU and current intra- and extra-ECOWAS trade flows. It also provides a description of the implied changes to the tariff structure with the ECOWAS CET. Second, it estimates the elasticity of import demand for individual members of the WAEMU. Third, a partial equilibrium framework is used to assess the potential trade and revenue effects of the tariff change. Finally, a panel vector auto regression (VAR) model assesses the dynamic macroeconomic effects of changes in trade tariffs in the WAEMU.

Trade between WAEMU countries and the rest of ECOWAS is rather low. The WAEMU’s interregional trade as well as trade with the rest of the ECOWAS remains weak, with some variation across countries. Burkina Faso, Guinea-Bissau, and Mali are the only WAEMU countries for which intra-WAEMU imports constitute more than one-fifth of the total import value, and only Côte d’Ivoire, Niger, and Senegal import close to or more than 10 percent from non-WAEMU ECOWAS countries (Figure 25.1, panel 1). Trade balances between the WAEMU and the rest of the ECOWAS thus remain low, with only Guinea-Bissau individually showing a trade surplus with the rest of ECOWAS of almost 7 percent of its GDP.

WAEMU countries in general rely heavily on import duties (Figure 25.1, panel 3). These duties constitute a substantial source of revenue for WAEMU countries, representing at least 8 percent of total government revenue in any WAEMU country over the period 2000-13. In 2013, import duties represented at least 13 percent of total government revenue in all WAEMU countries, with the highest share of 34 percent in Benin. With the exception of Côte d’Ivoire, more than half of imports to any WAEMU country face a tariff rate of 10 or 20 percent.

The introduction of the ECOWAS CET eliminated tariffs on the WAEMU’s imports from other ECOWAS countries but increased tariffs on imports from the rest of the world. The ECOWAS CET is organized in five bands. The first four bands are taken from the WAEMU CET (see Table 25.1). However, the ECOWAS CET includes a fifth tariff band of 35 percent for specific goods for economic development, which implies a customs tax increase for WAEMU countries and a reduction for non-WAEMU ECOWAS countries, such as Nigeria and Guinea.

Table 25.1.Structure of ECOWAS and WAEMU Common External Tariff
CategoryDescriptionWAEMU Duty Rate

(In percent)
ECOWAS Duty Rate

(In percent)
0Essential social goods00
1Goods of primary necessity, raw materials, and specific inputs55
2Inputs and Intermediate goods1010
3Final consumption goods2020
4Specific goods for economic development35
Average11.912.3
Source: Roquefeuil and others (2014); and the West African Economic and Monetary Union Commission.Note: The structure of the Economic Community of West African States common external tariff as in the version agreed upon by the region’s ministers in Praia in March 2013; details on the tariff bands applied to each tariff line are shown in WAEMU Regulation 23/2002/CM/UEMOA.
Source: Roquefeuil and others (2014); and the West African Economic and Monetary Union Commission.Note: The structure of the Economic Community of West African States common external tariff as in the version agreed upon by the region’s ministers in Praia in March 2013; details on the tariff bands applied to each tariff line are shown in WAEMU Regulation 23/2002/CM/UEMOA.

Tariff Effects

To quantify each WAEMU member country’s import responsiveness to price and tariff changes, the price elasticity of the volume of trade is estimated in a first step. To this end, an import function is postulated in which the quantity of imports depends on the price of imports, the price of other domestic consumable commodities, and domestic income. Controlling for further determinants of imports, the following relationship is estimated econometrically

in which in each year t, Mt is the import volume index, PtmCPIt is the ratio between the import price index and Consumer Price Index, TAXMt is the level of import duties, and GDP is the real GDPt level (constant 2005 prices). β0, β1, β2 β3, and εt represent a constant, the price elasticity of import demand, the income elasticity of import demand, the elasticity of import to collected duties revenue, and the error term, respectively. The estimated import demand elasticities are presented in Tables 25.2 and 25.3.

Table 25.2.Price Elasticities of Imports: Individual Countries
BeninCote d’IvoireMaliTogo
Import price-CPI ratio−2.43***−0.22*−0.87**−0.63***
GDP−4.00***1.19***
Import duties1.11***0.25**0.35**
GDP-export1.08***0.71**
Volume of exports0.86***1.12***
Constant33.91***−10.65**−4.32**−5.58**
p-value; F test0000
Source: Author’s estimates.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively; Heteroskedasticity corrected by White (1980) estimator; Serial Correlation corrected by Cochrane-Orcutt (1949) estimator.
Source: Author’s estimates.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively; Heteroskedasticity corrected by White (1980) estimator; Serial Correlation corrected by Cochrane-Orcutt (1949) estimator.
Table 25.3.Price Elasticity of Imports: WAEMU
WAEMU
Import price-CPI ratio−0.39***
Real GDP0.73***
Import duties0.46***
Constant−0.79
Source: Author’s estimates.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively; Panel corrected standard errors (control for heteroskedasticity/cross sectional dependence); Serial correlation corrected by Prais Wintsen (1954) estimator with panel specific AR(1) coefficient.
Source: Author’s estimates.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively; Panel corrected standard errors (control for heteroskedasticity/cross sectional dependence); Serial correlation corrected by Prais Wintsen (1954) estimator with panel specific AR(1) coefficient.

The price elasticity of import demand varies strongly across WAEMU countries. The results indicate that the percent decrease in import demand due to a one percent change in the import price varies from about 0.2 (in Côte d’Ivoire) to 2.4 (in Benin), with an average of about 0.4 for the WAEMU region as a whole. The high elasticity in Benin may be explained by the informal reexports of Benin to Nigeria, estimated to represent 50 percent of imports going through the Port of Cotonou (see Box 25.1).

The change in the tariff structure is expected to increase the WAEMU’s imports from other ECOWAS countries, but decrease its imports from the rest of the world. The trade effects are estimated using the methodology described in Box 25.2. With the exception of Togo, for which the share of non-ECOWAS imports in total imports is very high, trade creation is estimated to be larger than trade diversion in all WAEMU countries. In other WAEMU countries, the tariff change would imply an increase in imports from non-WAEMU ECOWAS countries by approximately 6 percent, with the highest increase, of 52 percent, in Benin (Figure 25.2, panel 1). The increase in the average tariff for non-ECOWAS countries is estimated to slightly reduce WAEMU countries’ imports from the rest of the world, with imports decreasing by less than 0.2 percent for most WAEMU countries, and the highest effect, of a ¾ percent decrease, expected in Benin (Figure 25.2, panel 3).

Figure 25.2.Trade and Revenue Implications of the Tariff Change

Source: Country authorities; own estimates; IMF staff estimates.

Note: BEN = Benin; BFA = Burkina Faso; CIV = Côte d’Ivoire; GNB = Guinea-Bissau; MLI = Mali; NER = Niger; SEN = Senegal; TGO = Togo; WAEMU = West African Economic and Monetary Union; ECOWAS = Economic Community of West African States.

The estimated trade effects of tariff elimination should be considered as the upper bound of the potential effects. Similarly to the effects after the introduction of the WAEMU CET, trade creation may be lower because some nontariff barriers may remain in place after the implementation of the ECOWAS CET. Trade diversion may also be lower because some WAEMU imports from non-ECOWAS countries have no substitute in other ECOWAS countries, so that the decrease in imports from non-ECOWAS countries could be lower than the effect estimated in this chapter.

Box 25.1.Revenue Impact in Benin

The estimated revenue impact related to trade liberalization is much stronger in Benin than it is in other WAEMU countries due to Benin’s unique trade patterns, which are dominated by informal trade with neighboring Nigeria (Figure 25.1.1). This box provides more details on such trade and its potential revenue impact.

Figure 25.1.1.Revenue Implications of a Full Liberalization of Nigeria’s Trade Regime for Benin1

(2011, in percent of GDP)

Sources: Beninese authorities; and IMF staff estimates.

1The tax revenue after liberalization is estimated based on the assumptions that potential losses due to full liberalization of Nigeria’s trade regime are distributed as 1/3 for customs revenue and 2/3 for goods and services tax revenue

Benin’s trade with Nigeria mainly takes the form of informal reexports to Nigeria due to trade restrictions on certain products. The proximity to Nigeria and a relatively porous border has made Benin the preferred platform for importing certain goods that Nigeria forbids either through bans or through prohibitive tariffs. These products include frozen poultry, rice, used cars, and textiles. It is estimated that about 50 percent of imports going through the Port of Cotonou are destined for Nigeria. Overall, at least 20 percent of Benin’s GDP is generated through informal trade (World Bank 2014).

These reexport activities are a significant source of tax revenue for Benin. The Beninese authorities reached an agreement with the Nigerian government to not reexport certain goods, so that the only way for Beninese importers to reach the Nigerian market is to first declare imported goods for domestic consumption, pay the value-added tax, and then reexport the goods informally to Nigeria. According to Geourjeon, Chambas, and Laporte (2008), the tax revenue gain stemming from this activity could be about 2 percent of GDP or 14 percent of total tax revenue, respectively. This highlights Benin’s fiscal dependence on informal reexport activities with Nigeria.

Trade liberalization in Nigeria would thus result in significant revenue losses in Benin. Based on customs data from 2009–12, the impact of a full trade liberalization scenario in Nigeria is estimated to result in a revenue loss of at least 2 percent of GDP (Sola and others 2013).

Box 25.2.Assumptions and Analytical Framework

Assumptions (Figure 25.2.1.)—The world is divided into three blocks from which WAEMU countries import: WAEMU countries, other ECOWAS countries that are not members of the WAEMU (the rest of ECOWAS), and the rest of the world. The supply of products to the WAEMU has perfect supply elasticity (overall supply to WAEMU countries for products is infinite at a given price). In each WAEMU country, local consumers substitute imperfectly products from these regions, but all products from the three alternative sources are equally sub-stitutable. The elimination of tariffs within the ECOWAS is assumed to have negligible competition effects between intra-WAEMU exports to a given WAEMU country.

Figure 25.2.1.Composition of Domestic Demand

Source: Author’s illustration.

Trade creation—This refers to the substitution of domestic production by imports from the rest of the ECOWAS, resulting from tariff elimination. Trade creation between WAEMU country i and the rest of ECOWAS is

in which Mi, RECOWASis the value of imports from RECOWAS by country i, τWAEMUis the WAEMU’s CET average tariff andηimis the import demand elasticity.

Trade diversion—This is the substitution of imports from the rest of the world by imports from the rest of the ECOWAS, resulting from the elimination of tariffs between WAEMU and ECOWAS countries. It is defined by

Other trade effects—The ECOWAS CET against non-ECOWAS members is higher than those applied initially by the WAEMU CET to non-ECOWAS countries. The implementation of the ECOWAS CET by the WAEMU thus eliminates their tariff against other ECOWAS countries but increases their tariff against non-ECOWAS countries. It may thus lead to a trade loss within WAEMU countries and the rest of the world. This other trade effect (OTE) is the change in WAEMU country i’s demand for import from non-ECOWAS country j, resulting from the increase in tariffs associated with the adoption of ECOWAS CET. The OTE can be written as

where Mi,ROW is the value of imports from the rest of the world (ROW) by the country i, τECOWAS is the average tariff of ECOWAS’s CET, τWAEMU is WAEMU’s CET average tariff, and ηim is the import demand elasticity of country i.

The overall net trade effect can then be computed as

Revenue Implications

Based on the projected trade effects, the revenue implications are estimated. The change in revenues in the WAEMU would be the combined effect of changes in tariff income from imports from non-ECOWAS countries (higher tariff versus lower import value) and tariff income from non-WAEMU ECOWAS countries (lower tariff versus higher import value):

In this equation, Mi,ROW and Mi,RECOWAS represent the value of imports from non-ECOWAS countries and non-WAEMU ECOWAS by country i, respectively. TDi,RECOWAS is the trade diversion from the non-ECOWAS countries to non-WAEMU ECOWAS countries. τECOWAS and τWAEMU are the average tariffs of the ECOWAS CET and the WAEMU’s CET, respectively.

The changes in tariffs could have ambiguous effects on revenue across WAEMU countries. The elimination of tariffs on imports from other ECOWAS countries would decrease government revenue. However, the tariff increase for products from the non-ECOWAS countries could have a positive impact on revenues. Based on the assumptions in Box 25.1 and the preceding paragraph, revenues in Benin, Burkina Faso, Cote d’Ivoire, Niger, and Senegal could decrease by ½ to approximately 2½ percent from their 2013 levels (Figure 25.2, panel 4). However, consistent with current trade profiles implying a low share of imports from non-WAEMU ECOWAS countries, revenues could increase by ½ to 3 percent in Guinea-Bissau, Mali, and Togo.

Data limitations prevent accounting for informal trade in the WAEMU. The decrease in government revenue may be higher in some countries if informal trade is taken into account. For instance, this may be the case in Benin, where informal reexports to Nigeria are estimated to contribute 2 percent of GDP to fiscal revenue (Box 25.2). The implementation of the ECOWAS CET is more likely to imply a stronger fall in imports for reexports and hence the reexports, implying a loss in import duty revenue as well as value-added tax revenue. Due to trade policy restrictions, Beninese importers can only reach the Nigerian market by declaring imported goods for domestic consumption, paying the value-added tax, and reexporting informally the goods to Nigeria.

Macroeconomic Effects

To assess the dynamic effects of changes in tariff structure on imports in the WAEMU, the traditional VAR model is combined with a panel data approach. A first order Panel VAR model is specified as

in which Yt is a four-variable vector (import duties, real GDP, relative import price, import). An impulse response function then determines the transmission channels of an exogenous change in import tariffs to imports based on a Choleski decomposition assuming the same ordering.

An increase in relative import prices induced by an increase in the average tariff rate could yield a negative effect on GDP. Table 25.4 depicts the results of the estimation of the previously mentioned model for a panel of WAEMU countries for the period between 1995 and 2013. Column 2 implies a statistically significant negative relationship between import prices and GDP. An increase in the average tariff rate, which implies an increase in import prices, would thus be expected to affect GDP negatively as well. In terms of the dynamic effects (Figure 25.3), the impulse response function shows that import prices adjust to an increase in tariff with a one-period lag, and imports simultaneously induce a decrease in imports. Current increases in relative import prices dampen economic growth within one period. Increases in GDP result in higher demand, including the demand of foreign goods and higher domestic production of importable goods later on. However, as domestic production of importable goods rises faster than domestic consumption of importable goods, lagged GDP decreases imports.

Table 25.4.Estimated Coefficients of Panel VAR
Response toResponse of
Import duties(t−1)Import Prices(t−1)Import (t−1)GDP(t−1)
Import duties (t)−0.007−0.120.051−0.1750.131−0.132−1.624−1.104
Import Prices (t)0.132*−0.08−0.046−0.1040.037−0.1010.577−0.68
Import (t)0.05−0.0910.149−0.144−0.085−0.123−2.367**−1.014
GDP (t)0.004−0.018−0.055**−0.023−0.003−0.021−0.069−0.183
Source: Author estimations.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively, with standard errors in parentheses; First lags of dependent variables in the model are instrumented by the same lags; Stability test shows that all eigenvalues are inside the unit circle, i.e., the PVAR is stable. VAR = Vector Autoregression; PVAR = Panel VAR; t = time.
Source: Author estimations.Note: ***; **; * significant at 1, 5, and 10 percent level, respectively, with standard errors in parentheses; First lags of dependent variables in the model are instrumented by the same lags; Stability test shows that all eigenvalues are inside the unit circle, i.e., the PVAR is stable. VAR = Vector Autoregression; PVAR = Panel VAR; t = time.

Figure 25.3.Transmission Channels of Changes in Tariff Structure to Imports1

Source: Own estimations.

1RMF Notes: RMP = relative import price; TAXM = import duties; M = imports. IRF of response variables to one standard deviation increase in impulse variables. Confidence intervals computed using Monte Carlo Simulations.

References

    GeourjonA. M.G.Chambas and B.Laporte.2008. “République du Bénin. Modernisation du Système Fiscal. Orientations et Stratégie de Réformes.September. International Monetary FundWashington.

    • Search Google Scholar
    • Export Citation

    RoquefeuilQ. deD.Plunkett and F.Ofei.2014. “The Impact of Closer Regional Economic Integration on Food Security in West Africa: Focus on the ECOWAS Common External Tariff.Discussion Paper No. 154.European Centre for Development Policy ManagementBrussels.

    • Search Google Scholar
    • Export Citation

    SolaS.G.ParentJ.RussellA.Ould Boilil and G.Barker.2013. “République du Bénin. Bilan des Reformes et Poursuite du Renforcement de l’administration Douanière.International Monetary Fund Fiscal Affairs Department June.

    • Search Google Scholar
    • Export Citation

    World Bank. 2014. “Benin Diagnostic Trade Integration Study (DTIS) Update: From Rents to Competitiveness.World BankWashington.

    Other Resources Citing This Publication