Kenya: Selected Issues and Statistical Appendix

Like most Sub-Saharan African countries, Kenya’s economic growth appears to have been primarily driven by factor accumulation. The Selected Issues paper and Statistical Appendix for Kenya examines economic developments and policies. During the last two decades, Kenya has been plagued by pervasive problems of internal conflicts, constitutional crises, and corruption scandals. The governance agenda focuses on several reforms, including upgrading the public budget and financial management systems, strengthening the anticorruption institutions, and improving the judicial framework.

Abstract

Like most Sub-Saharan African countries, Kenya’s economic growth appears to have been primarily driven by factor accumulation. The Selected Issues paper and Statistical Appendix for Kenya examines economic developments and policies. During the last two decades, Kenya has been plagued by pervasive problems of internal conflicts, constitutional crises, and corruption scandals. The governance agenda focuses on several reforms, including upgrading the public budget and financial management systems, strengthening the anticorruption institutions, and improving the judicial framework.

V. Trade Integration in the East African Community1

A. Introduction

1. The multilateral trading system is guided by the nondiscrimination principle. The Fund has stressed that nondiscriminatory trade liberalization on a most-favored-nation MFN basis is the first-best policy. Despite the well-documented superiority of MFN liberalization, regional trade arrangements (RTAs) have always been part of the economic relations between countries. RTAs have proliferated in Africa to such an extent that the issues of overlapping membership in regional integration arrangements need to be addressed (see Figure V.1).

Figure V.1.
Figure V.1.

Regional Integration Arrangements in Africa in Africa

Citation: IMF Staff Country Reports 2009, 192; 10.5089/9781451821253.002.A005

Note: Tanzania withdrew from COMESA in 2000.Source: Adapted from DeRosa et al (2003)

2. Kenya, Tanzania and Uganda have consisted liberalized their trade regimes at both the regional and global levels. As they have promoted more open and liberal trade policies, the three countries have simultaneously embarked upon a process to integrate their economies through the creation of the East African Community (EAC). 2 The formation of the EAC customs union is an important step in the process of deepening regional integration. The EAC treaty provides for the formation of a customs union by 2004. 3 The formation of a customs union requires the removal of all internal tariffs, the establishment of a common external tariff, Rules of Origin and a variety of administrative arrangements including a harmonized customs administration, a customs valuation system and customs procedures and documentation.

3. The Chapter has three main tasks to accomplish: First, it identifies the key features of EAC member countries’ trade flows and trade regimes (Section B). The paper then describes the new EAC customs union (CU), particularly the EAC common external tariff (CET), analyzes its impact on the trade regimes in EAC member countries, and attempts to gauge its potential impact on trade by conducting simulations for Kenya (Section C). Finally, it discusses factors other than trade that could make regional integration in the East African region a desirable policy for Kenya (Section D), and offers conclusions in Section E.

B. Trade Flows and Trade Regimes in the EAC

Trade Flows

4. Overall, the trade data in Table V.1 indicate that the direction and pattern of trade of the three EAC members is consistent with their level of development. They export primary products,4 mainly to Europe, and to a lesser extent, the Middle East. In 2001, the EU received 37.1 percent and 64.5 percent of Tanzania’s and Uganda’s exports. The exception is Kenya, whose exports to African countries, particularly the EAC sub-region, are substantial. Kenya’s exports to the EU were 31.9 percent of the total whereas exports to other African countries accounted for 35.9 percent and to the EAC, 22.6 percent. Imports from Africa and the Middle East (mainly Egypt) are 35.7 percent of Kenya’s total imports and EU imports, 27.3 percent. Tanzania and Uganda received a large share of their imports from Africa and the Middle East (35.0 percent and 60.2 percent respectively of total imports).

Table V.1.

Kenya: EAC Countries: Exports and Imports, 2001

(in million of U.S. dollars)

article image
Source: World Bank estimates.

5. In the last decade intra-regional trade has grown, with the share of intra-regional exports increasing from about 6 percent in 1991 to 16 percent in 2001 and imports rising from 2.7 percent in 1991 to 10.5 percent in 2001. Despite these gains the trade linkages between the countries could be stronger. Although Kenya sends a significant share of its exports to the EAC, it sources only 1.4 percent of total imports from the sub-region (Table V.1). Tanzania sends only 9.9 cent of total exports to the sub-region and receives from it 7.2 percent of its total imports. However, while Uganda’s exports to the EAC are similarly low, it receives a substantial 48.8 percent of total imports from the EAC (mainly from Kenya).

6. The commodity composition of intra-regional trade reveals that unlike trade with the rest of the world, manufactures play an important role. Table V.2 indicates that for Kenya 11.5 percent and 43.4 percent of its imports from respectively, Uganda and Tanzania are manufactures. For Uganda, 33.8 percent and 71.3 percent of its imports are from Kenya and Tanzania, and for Tanzania 56.8 percent and 16.6 percent of its imports are from Kenya and Uganda. In short, the expansion of intra-regional trade has provided a market for the manufacturing sectors in the EAC member states, particularly Kenya. The challenge is to transform these industries to produce internationally competitive exports and go beyond the regional market.

Table V.2.

Kenya: EAC Countries: Regional Trade by Commodities, 2001

(percent of total)

article image
Source: United Nations Commodity Trade Statistics Database, 2003.

Trade Regimes

7. The trade regime in the EAC member countries is characterized by the escalating structure of tariffs. A “cascading tariff”5 structure has the lowest rates being imposed on raw materials and capital goods, moderate rates on intermediate goods, and the highest rates on consumer goods. These structures reflect the historical pattern of tariffs in many countries, with high rates being placed on consumer goods partly to restrain demand and collect revenue but also to protect or stimulate domestic producers of final consumer goods over foreign competition. Trade liberalization in recent years has however brought about considerable reductions in the top rates and rationalized the structure of tariff regimes so that the differences have fallen considerably. Table V.3 provides detailed information on the key features of the trade regimes of the EAC member countries.

Table V.3.

Kenya: Features of Trade Regimes of Kenya, Tanzania, and Uganda

article image

Kenya has a select group of products that are granted higher rates than in the tariff structure 0-35.

For Kenya, the alternative minimum duty rates are set as floor rates based on the lowest expected prices.

8. Table V.4 shows that all three countries had progressively reduced their tariffs since the mid-nineties. 6 The most significant changes were in Uganda and to some extent Tanzania and this is manifested by the fall in the maximum rates, the number of tariff bands, and the simple average tariff. In addition, Uganda has narrowed the differences between the top rate on consumer goods and the lower rates on raw materials and capital goods. In contrast, Kenya has not made any progress in liberalizing its tariff schedule, but its simple average tariff has marginally declined as a consequence of modifications in tariff classifications.

Table V.4.

Kenya: EAC Countries: Evolution of Tariff Regimes, 1997-2002

(tariff rates in percent)

article image
Sources: World Trade Organization and United Nations Conference on Trade and Development.

Data for Tanzania are for 2001.

C. The EAC Customs Union

The EAC Common External Tariff

9. The treaty establishing the East African Community (EAC), comprised of Kenya, Tanzania and Uganda, was signed by the three member governments in November 1999. Formally launched in 2001, the EAC treaty provides for the formation of a customs union by 2004. 7 On June 23rd, 2003, the Presidents of Kenya, Tanzania, and Uganda reached an agreement on the CET for the planned customs union. The CET will have three tariff bands. 8 0 percent for meritorious goods, raw materials and capital goods; 10 percent for intermediate goods, and 25 percent for consumer goods.

10. The principle of asymmetry is recognized in the Treaty establishing the East African Community as a core principle underpinning the formation of the EAC customs union. The justification for including the principle of asymmetry in the Treaty is based on the understanding that the three EAC member states are at different levels of economic development and that there is need to address the existing imbalances which could in fact be exacerbated by the customs union. In the negotiations on the CET it was agreed that Tanzania and Uganda will eliminate tariffs on all imports except for an agreed list of commodities9—906 tariff lines for Tanzania and 426 for Uganda—for which the tariff will be reduced gradually to zero, within a period up to five years. In the case of Uganda the items on that list will initially attract a 10 percent tariff and over a five year period this will be uniformly reduced. Tanzania has a more complicated arrangement for tariff reduction, with each product group having a different schedule for reducing tariffs, however no tariff will initially be higher than 25 percent and the reduction to zero will be within five years. In short, the EAC CET will be implemented in two phases: First, all three countries will adopt the three-band structure but Tanzania and Uganda will maintain internal tariffs on a select set of Kenyan imports; second, after five years all internal tariffs will be removed and all Kenyan imports will enter Tanzania and Uganda free of tariffs.

11. A major issue in the negotiations on the CET was reaching agreement on the classification of about 20 percent of the tariff lines which are defined as “sensitive items.” The EAC members claim that these are products that they would like to protect from import competition from the following products:

  • Subsidized exports, mainly agricultural products, from industrialized countries;

  • Second-hand products.

12. The World Bank (2003) indicated that the “sensitive items” included cigarettes, dry cells, fabrics, garments, matches, milk, other cement, packing materials of plastic, palm oil, sugar, tires, used clothes, vehicles (reconditioned cars), vehicles chassis, rice, wheat and wheat flour. These items are equivalent to 361 tariff lines and estimated at about 20 percent of total imports. 1 As of September 2004, after rounds of negotiation, agreement was reached on the classification of sensitive products and the applicable rates of dut, with the exception of jute bags, rice, and wheat. Further, it was agreed that sensitive products could not be “protected” by the maximum rate and therefore required special policy measures. The EAC member states agreed that the sensitive items would attract rates of more than 25 percent and in some instances a mixture of specific duty and ad valorem rates.

13. The new common external tariff will have differential effects on the trade regimes in the member countries. The introduction of the three-band tariff structure will increase tariffs in Uganda and to a lesser degree Tanzania, and reduce tariffs in Kenya. In Table V.5 the number of tariff lines that are likely to increase in Uganda is 3,066, compared to 1,224 in Tanzania and 1,144 in Kenya. In contrast, the EAC CET is likely to lower significantly more tariffs in Kenya (3,216) compared to Tanzania (2,364) and Uganda (1,353). In addition, World Bank (2003) research estimates that with the full implementation of the CET the simple average tariff in the three countries will be 10.9 percent which represents a significant decline for Kenya from a simple average tariff of 16.6 percent and to a lesser degree Tanzania with a simple average tariff of 14.3 percent. However, for Uganda there will be an increase of about 20 percent in the simple average tariff. 2

Table V.5.

Kenya: EAC Countries Estimated Effects of Proposed Tariff Changes

article image
Note: Estimated at 6-digit level of HS-classification.

14. Tanzania and Uganda apply excise duties and other discriminatory charges as a means of protection mainly against Kenyan imports. 3 Tanzania applies excise taxes on 55 items at specific or ad valorem rates of 10-30 percent with peaks of over 50 percent, mostly on Ugandan and Kenyan imports. Also, there are suspended duties on 118 items in the top tariff bracket, with peaks of 35 and 40 percent. Uganda applies discriminatory excise duties at ad valorem rates of 10 percent on 467 items increasing substantially to 75 percent for beverages and 130 percent for tobacco. 4 With the implementation of the EAC CET all discriminatory excise duties (except those applied to mineral water, tobacco, beer and alcoholic beverages) together with suspended duties will be eliminated.

15. In the other areas required for the establishment of an EAC customs union progress has been made but there are still some outstanding issues. The current situation can be summarized as follows:

D. Trade Impact: An Assessment for Kenya

Economic Integration Theory

16. Regional trading arrangements (RTAs) alter the prices of imports from members of the RTA (as tariffs are phased out) relative to imports from the rest of the world. Consequently, demand patterns will change resulting in adjustments in trade and output flows. Will these changes be beneficial to participants in an RTA? Alternatively, will a RTA generate gains from trade? Viner (1950) investigated this question and found that the welfare impact of an RTA is ambiguous. Gains will occur if higher cost domestic production is replaced by cheaper imports from a partner country—trade creation. On the other hand, if partner country production replaces lower cost imports from the rest of the world—trade diversion—there will be losses. Therefore, membership in a RTA will have positive and negative effects on an economy and it will be the net impact that will determine whether there are welfare gains or losses.

17. In assessing the static effects of forming an effective RTA three important principles from the theory of integration must be considered: First, the allocative or efficiency gains of economic integration depend on whether the products produced by members of the RTA are in direct competition or complementary to each other. 7 For there to be competitive economies or efficiency gains in an RTA there must be a considerable degree of overlap in the range of commodities produced by members of the RTA. The creation of an RTA where there exists overlapping production with significant differences in production costs between members can lead to large gains from trade as resources are allocated more efficiently among member countries. Intra-industry trade (e.g., Ford cars for Honda) characterizes most trade between industrial economies and the formation of an RTA is likely to lead to competitive gains. For example, it can be argued that the members of the European Union (EU), US/Canada FTA and the Australia/New Zealand FTA are competitive economies and that there were significant gains from trade. It is questionable whether the members of a large number of RTAs between developing countries can be characterized as competitive economies. Typically, members of developing country RTAs have a narrow range of exports of goods and services, invariably primary commodities that are exported to industrialized countries often under unilateral preferential arrangements. Therefore, there is little scope for efficiency gains.

18. Economies whose structure of production are not competitive tend to be complementary and this can result in both gains and losses from RTAs. Complementarity exists when members of RTAs produce commodities or products that do not compete much with the local production of other RTA members. Traditional integration theory contends that, in the case of complementary economies, economic integration will have the usual trade diversion and trade creation effects; the higher the barriers to trade with non-members, the higher the risk of trade diversion. Intuitively, one can argue that complementarity exists between developed and developing country members in an RTA (i.e., North/South RTAs). Trade between industrial countries and many developing countries is often characterized as trade in homogenous products, e.g., wheat for textiles. In this case each country will have a comparative advantage in the export of a different type of good while all goods will be consumed by all member countries. The proposed regional economic partnership agreements that are part of the Cotonou Agreement between the EU and the member states of the Africa, Caribbean and Pacific (ACP) region might be characterized as RTAs between complementary economies.

19. The intra-EAC trading patterns in Table V.2 indicate that trade linkages are relatively weak. Therefore, one cannot really characterize the economies as either complimentary nor competitive. In the case of the latter, this means there is not a considerable degree of overlap in the range of commodities produced by EAC members.

The Trade Simulation Model

20. Partial equilibrium models are widely used to simulate and measure the effects of changes in trade policy. The models assess the effects of specific changes in tariffs or other trade taxes on trade flows, revenue, prices, and some measures of welfare (consumer surplus) at a given point in time. Typically, a simulation model based on simple Vinerian customs union theory is employed. A simulation of the impact of the EAC CU was conducted utilizing a static, partial equilibrium methodology—SMART8 (See Appendix I)9. Notably, SMART, unlike some partial equilibrium models, assumes that products

imported from different regions are imperfect substitutes among themselves10.

21. The World Integrated Trade Solution (WITS) software developed by the World Bank was used to conduct the simulations. 11 WITS utilizes the UN Statistics Division COMTRADE and the UNCTAD Trade Analysis and Information System (TRAINS) databases providing access to data on trade flows12 and MFN tariff rates at the HS six digit level of disaggregation. World Bank staff and the Kenya Revenue Authority provided information on the tariff preferences offered to COMESA partners and the negotiated common external tariff. The SMART simulations were done using the WITS software.

22. The simulation results produced by SMART indicate that the move from the current MFN tariff rates to the three-band EAC CET is likely to have a positive impact on trade with an increase in trade of US$193.5 million13 with trade creation estimated at US$193.9 million and trade diversion at US$ 0.3 million14. Table V.6 shows the impact on trade and the estimated trade creating and trade diverting trade flows for all products in each tariff band i.e. 0 percent and 25 percent. The results reveal that 81.2 percent (US$157.5 million) of trade creating flows resulting from the move to the new EAC CET are accounted for by products that attract a 0 percent tariff rate. 15 Trade creation has a positive effect on welfare as consumers can purchase cheaper imports than more expensive local goods. However, it means import-competing producers will need to become more competitive or move into new product lines. These sectoral adjustments are the transitional or adjustment costs of lowering trade barriers.

Table V.6.

Kenya: Trade Simulation Results

article image
Source: Fund Staff Estimates

23. The move to the maximum tariff rate of the EAC CET results in trade creation estimated at -US$9.1 million. The model reports the results as negative trade creation but this really reflects lower trade flows resulting from higher tariff rates. In other words, this means that the new EAC CET led to higher tariff rates for some of these product lines and with higher import prices, import flows declined. 16 Further examination of the individual product tariff lines revealed that many products that attracted a 15 percent MFN tariff rate now face the maximum tariff rate. Notably, some of these products—fish, pigs, other black tea, yeasts, pictures and designs and steel products—are produced locally, hence there is a protectionist objective.

24. Another important feature of the results reported in Table V. 6 is the negligible trade diversion resulting from the new EAC CET. An important factor that might be affecting the quantitative results is that the baseline imports from Tanzania and Uganda reported in the official statistics significantly underestimate intra-regional trade because of the prevalence of unrecorded informal cross-border trade. Mkenda (2001) cites surveys that indicated that in the 1994/95 period, unofficial cross-border trade between Kenya and Uganda was about 49 percent of official trade. Between Tanzania and Kenya, cross-border trade as percentage of official trade in the 1995/96 period was about 12 percent and between Tanzania and Uganda, it was about 45 percent.

25. The simulation results provide preliminary evidence that the EAC CU will have positive trade benefits for Kenya as the adoption of the EAC CET will lead to increased flows of cheaper extra-regional imports that are likely to, lower consumer prices with positive welfare effects. Note that in the simulation, the removal of internal tariffs was accompanied by a lowering of MFN tariffs with the adoption of the EAC CET. A World Bank (2000) study concluded that regional integration arrangements (RIAs) between developing countries (South-South RIAs) that provide preferential access to member states but keeps external trade policy with respect to the rest-of-the-world unchanged are likely to lower welfare for the bloc as a whole. High external tariffs encourage trade diversion and provide strong incentives for inefficient firms to expand. Fundamentally, high external barriers negate the benefits from increased competition. Therefore, to ensure that an RTA does not encourage inefficiency, facilitate trade diversion, and ultimately reduce economic welfare, it is essential to lower MFN tariffs as barriers to intra-RTA trade are eliminated. 17 Therefore, Kenya could continue to derive benefits from progressively lowering trade barriers, specifically the EAC CET.

Transitional Costs

26. Despite the potential benefits from liberalization of the trade regime there are costs that would have to be addressed. As noted earlier, trade creation means that the import-competing sectors would face increased competition and would need to make adjustments to improve efficiency and overall competitiveness. Consequently, there may be transitional output and employment losses associated with the EAC CU. Policies would need to be put in place to minimize the dislocations caused by the lowering of tariffs. For import-competing sectors to respond to increased competition from cheaper imports it is vital that Kenya, over the medium term, sustain the implementation of a comprehensive package of macroeconomic and structural reforms to improve efficiency and international competitiveness. This would include:

  • strong governance policies to improve transparency and accountability and eliminate corruption;

  • strengthening the efficiency of the financial system;

  • labor market reforms to increase labor market flexibility;

  • an accelerated program of parastatal reform and privatization to increase efficiency and private sector involvement in the economy; and

  • and prudent fiscal policies to ensure that adequate resources are devoted to infrastructural development and improving the levels of education and health.

A poverty and social impact analysis (PSIA) of trade reforms is planned by the authorities and could provide the basis for programs to address these concerns.

27. The customs union is expected to result in revenue losses. The SMART simulations estimated that the full implementation in Kenya would result in customs revenue losses of US$113.3 million. An earlier analysis by the World Bank (2003) estimated the revenue losses from the proposed three-band structure(0,10,25) of approximately US$150 million for Kenya. 18 The empirical evidence thus suggests there will be short-run revenue losses from the full implementation of the EAC CU and policymakers have to design policy responses to recoup revenue losses. Krause (2003) estimated that in Kenya customs exemptions amount to 22 percent of potential customs revenue, so to compensate for revenue losses, policymakers could streamline exemptions, widening the tax base and increasing revenues.

E. Other Reasons for East African Integration

28. Trade integration is not the only reason why policymakers in Kenya might find regional integration in the East African region a desirable policy. Other factors are described below:

“Widening and deepening” of regional integration

29. From a Kenyan perspective, some commentators see the recently established EAC CU as providing an impetus to the COMESA CU. Although Tanzania is not a member of COMESA19 it is felt that the EAC group led by Kenya could set the EAC CET as the goal for the COMESA CU and be the prime force in the negotiations. A wider COMESA CU is attractive to Kenya as it provides a larger market to encourage the expansion of its manufactured or non-traditional exports to the region.

30. Another important factor might be the 'Economic Partnership Agreements’ (EPAs) that are to be negotiated between the European and sub-Saharan Africa (SSA) countries20. The Cotonu Agreement provides for the negotiation of reciprocal trade agreements between various geographical configurations in sub-Saharan Africa and the EU covering trade in goods and services and some trade-related areas. Currently, the regional groupings identified to negotiate EPAs include COMESA. The EAC has not been identified as a regional grouping for the negotiations. However, if the EAC is able to drive the negotiations for a COMESA CU, it could potentially be an important partner in the negotiations with the EU. Potentially, this is the most important regional agreement Kenya will negotiate as it offers a favorable opportunity for SSA countries to integrate into the global economy and to benefit from deeper integration with a developed region. 21

Trade Facilitation and “Behind the Border Reforms”

31. Small and/or poor developing countries can pursue enhanced trading arrangements (including outside the framework of an RTA) by deepening cooperation in trade facilitation and “behind the border” reforms. An important question is whether more intensive regional cooperation in trade-related areas such as trade facilitation and “behind the border” reforms—these areas include sanitary and phyto-sanitary (SPS) standards, technical standards, investment code, competition law and intellectual property rights—is likely to expand trade and raise economic growth by increasing efficiency as well as private investment (domestic and foreign). Conceptually, adopting and implementing simple, transparent import and export regulations and efficient procedures for customs clearance will reduce transactions costs and enhance efficiency in EAC member countries and improve the environment for trade expansion. “Behind the border” reforms are increasingly an important part of the international trade architecture and of growing importance in the multilateral trade negotiations in the WTO. These reforms place great demands on a country’s human resource and institutional capacity and it seems intuitive that regional approaches will be beneficial for SSA countries with limited human resources and weak administrative capacity.

Public Goods

32. Schiff (2000) argued that regional cooperation on public goods—such as water basins (lakes, rivers), infrastructure (roads, railways, dams), the environment, hydroelectric and other sources of energy, fisheries can generate benefits for member states. In the case of the EAC member states there is a lot scope for cooperation in these areas and support can be received from the World Bank together with other multilateral, regional, and bilateral agencies.

F. Conclusions

33. Kenya, Tanzania and Uganda have undertaken trade policy reforms that have consisted of liberalization of their trade regimes at both the regional and global levels. As they have promoted more open and liberal trade policies the three countries have simultaneously embarked upon a process to integrate their economies through the creation of the East African Community (EAC). The formation of the EAC customs union is an important step in the process of deepening regional integration. Generally, RTAs between competitive and/or complementary economies have resulted in positive static and dynamic benefits for the participating countries. However, many RTAs between developing countries are not between economies that have these characteristics and the results have been disappointing. The trade linkages between the three EAC member states are not strong. However, the establishment of the EAC CU and the introduction of the EAC CET do seem to have potentially positive benefits for Kenya. The results from a SMART trade simulation model suggest that the EAC CET, by lowering tariffs has a positive impact on trade largely from trade creation. Lower tariffs result in lower import prices and increased flows of cheaper imports that improve consumer welfare.

34. The preliminary evidence from the simulations supports the pursuit of more liberal trade policies. However, there are transitional costs that must be addressed to minimize economic dislocation, including revenue losses. Furthermore, trade creation means the import-competing sectors will face increased competition from cheaper imports, and producers will have to improve efficiency and competitiveness. Sustained macroeconomic and structural reforms will be needed to ensure that a favorable enabling environment is created that will facilitate internationally competitive production.

35. There are other factors beyond trade integration that Kenyan policymakers may consider in pursuing closer East African integration. These include: First, the widening and deepening of regional interaction with other countries in the Eastern and Southern African region through COMESA and the negotiation of an EPA between COMESA and the EU, with its centerpiece being a comprehensive regional trade agreement. Second, regional cooperation in trade facilitation and “behind the border” reforms offer potential benefits to Kenya. Improvements in trade facilitation can improve transparency, reduce the costs of doing business and promote trade. Regional cooperation in implementing “behind the border reforms”, which are an increasingly important part of the architecture of the international trading system, can improve efficiency and facilitate trade in goods and services. Finally, regional cooperation in public goods can, among other things, lower the cost of vital infrastructural development, promoting growth and development.

Appendix—The Smart Simulation Model

The simplest version of SMART and its definition of trade creation and trade diversion is presented below.

A. Simplest Version

Assumptions:

  • 1) Partial Equilibrium: no income effects

  • 2) Armington Assumption: HS 6 digit goods imported from different countries are imperfect substitutes, i.e., bananas from Ecuador are an imperfect substitute to bananas from Saint Lucia.

  • 3) Export supplies are perfectly elastic: world prices of each variety (e.g., bananas from Ecuador) are given.

Analytical setup

One possible analytical setup for the demand structure in SMART is to assume a two-stage budgeting procedure (where income is kept exogenous). A better alternative is to assume a quasi-linear an additive utility function that is also additive on a composite numéraire good. More formally:

U=Σgug(mg)+n(1)

where n is the consumption of the composite numéraire good, mgis the consumption of the aggregate import good g (aggregate in the sense that it is a function of imports of good g from different countries); and ug is the sub-utility function of good g. The fact that the utility function is additive ensures that there are not substitution effects across goods g, and the linearity on the composite and numéraire good n ensures that there are no income effects.

Maximization of (1) subject to a budget constraint yields:

mg,c=f(pg,cd;pg,cd),g,c(2)n=y-ΣcΣgpg,cdmg,c

where mgc are imports of good g from country c, pg,cd is the domestic price of imported good g from country c, pg,cd is the domestic price of good g imported from all countries other than c, y is national income. Thus consumption of the composite and numéraire good, n absorbs all income effects. Domestic prices are given by:

DMSDR1S-#2342182-v2-Kenya - 2008 - 1st Review PRGF and 2004 Art IV - Selected Issues Paper and Statistical. DOC June 12, 2009 (4:41 PM)

pg,cd=pg,cw(1+tg,c)(3)

where pg,cw is the world price of good g imported from c, tgc is the tariff imposed on imports of good g imported from c, and is defined as:

tg,c=tgMFN(1-θg,c)(4)

where tgMFN is the Most Favored Nation (MFN) tariff imposed on good g, and θgc is the tariff preference ratio on good g when imported from country c72

Trade creation

Trade creation is defined as the direct increase in imports following a reduction on the tariff imposed on good g from country c. To obtain this, SMART uses the definition of price elasticity of import demand:

εg,c=dmg,c/mg,cdpg,cd/pg,cd<0(5)

Solving (5) for dmgc we obtain the trade creation (TCgc) evaluated at world prices and associated with the tariff reduction on good g when imported from country c:73

TCg,c=pg,cwdmg,c=pg,cwϵg,cmg,cdpg.cdpg,cd(6)

Note that using (3), we have dpg,cd=pg,cwdtg,c. Substituting this and (3) into (6) yields:

TCg,c=pg,cwdmg,c=pg,cwϵg,cmg,cdtg,c(1+tg,c)=ϵg,cmg,cdtg,c(1+tg,c)(7)

Equation (7) defines the extent of trade creation on imports of good g from country c.

Note that in the last equality we simply choose units of all goods so that the world prices are equal to 1. One can then interpret mgc as import value of good g from country c measured at world prices. This normalization of units is undertaken from now on in order to simplify the expressions, so that m g , c represents both imported quantities and value of good g from

country c. As long as world prices are kept exogenous (i.e., export supply functions are perfectly elastic), this normalization has no implications for the derivations above and below.

To obtain the overall level of trade creation across goods or countries one simply needs to sum equation (7) along the relevant dimensions:

TCg=ΣcTCg,cTCc=ΣgTCg,c(8)TC=ΣgΣcTCg,c

Trade diversion

If the tariff reduction on good g from country c is a preferential tariff reduction (i.e., it does not apply to other countries, * c, then imports from country), then imports of good g from country c are further going to increase due to the substitution away from imports of good g from other countries that becomes relatively more expensive. This is the definition of trade diversion in the SMART model.

In order to measure trade diversion, let us use the definition of the elasticity of substitution, (σg, c,≠) across imports of good g from country c and all other countries (≠c):

σg,c,c=d(mg,cmg,c)/mg,cmg,cd(pg,cdpg,cd)/pg,cdpg,cd<0(9)

Note that:

d(pg,cdpg,cd)/pg,cdpg,cd=pg,cwdtg,cpg,cw(1+tg,c)pg,cw(1+tg,c)pg,cw(1+tg,c)=pg,cwdtg,cpg,cw(1+tg,c)=dtg,c(1+tg,c)(10)

Recalling that by definition of trade diversion dmgc = -dmgc, we have:

d(mg,cmg,c)=dmg,cmg,c-mg,cdmg,cmg,c2=dmg,c(mg,c+mg,c)mg,c2(11)

Substituting (11) and (10) into (9) and solving for dmgc yields the expression for trade diversion, TDgc:

TDg,c=dmg,c=mg,cmg,cmg,c+mg,cdtg,c1+tg,cσg,c,c(12)

B. Constraining Trade Diversion

There is one additional problem associated with the measurement of trade diversion. Indeed, by definition of trade diversion it cannot be larger than the original imports of good g from other countries ≠ c, i.e., TDg,c=dmg,c=-dmg,cmg,c A simple way of introducing this constraint is to defined trade diversion as follows:

TDg,c=dmg,c=-dmg,c={mg,cmg,cmg,c+mg,cmg,cdtg,c1+tg,cσg,c,cif-dmg,cmg,cif-dmg,c>mg,c(13)

So the constraint is binding only when it is necessary.

An alternative to the simple constraint in (13) is the one currently used by SMART. It introduces the constraint for all observations independently of whether the constraint is binding or not. This is done by transforming (12), so that TDg,c=dmg,cmg,c,g,c:

TDg,c=dmg,c=-dmg,c=mg,cmg,cdtg,c1+tg,cσg,c,cmg,c+mg,c+[mg,cdtg,c1+tg,cσg,c,c](14)

By adding the term in (14) the term in square brackets to equation (12), SMART constraints trade diversion to be equal to mgc when the term in square brackets (the change in tariffs multiplied by the change in relative prices and the elasticity of substitution) tends to infinity (or minus infinity). Indeed:

lim[mg,cdtg,c1+tg,cσg,c,c]-TDg,c=mg,c(15)

Equation (14) is clearly an underestimation of the trade diversion effect (we add a positive term to the denominator), whenever the term in squared brackets does not tend to infinity (e.g., for small tariff changes). More problematic is the fact that the terms in square brackets cannot tend to infinity unless either imports from c (mgc) or the elasticity of substitution are initially infinitely large. In which there is either no reason to worry about trade diversion or we are in a world with perfectly homogeneous goods in which case the constraint is always binding. Under more reasonable assumptions, the term in squared brackets can only tend to -mg,ctg,c1+tg,cσg,c,c as dt g, c tends to -tg, c when the tariff on good g from country c is eliminated. It is then not clear to which value the trade diversion term tends to, apart from the fact that it is clearly an underestimation of the true trade diversion for most values. For these reasons, we suggest the use of (13) rather than (14) to measure trade diversion.

Again the expression in (13) or (14) could be added across different dimensions (goods, countries or both) to obtain total trade diversion terms as we did for trade creation in equation (8). Finally, the total increase in exports of good g from country c associated with a preferential tariff granted to good g originating in country c is given by the sum of the trade diversion and trade creation terms.

Table 1.

Kenya: Gross Domestic Product by Origin at Constant Prices, 1996-2003

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues.

Includes general government.

Table 2.

Kenya: Gross Domestic Product by Origin at Current Prices, 1996-2003

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues.

Includes general government.

Table 3.

Kenya: Expenditure on Gross Domestic Product at Constant Prices, 1996-2003

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues.
Table 4.

Kenya: Expenditure on Gross Domestic Product at Current Prices, 1996-2003

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues.
Table 5.

Kenya: Gross Domestic Product, GDP Deflator, Population, and Real Per Capita GDP, 1987-2003

article image
Sources: Government of Kenya, Economic Survey, various issues; World Bank, World Development Indicators, various issues; and Fund staff estimates.
Table 6.

Kenya: Gross Fixed Capital Formation at Current Prices, 1996–2003

article image
Sources: Government of Kenya, Statistical Abstract and Economic Survey, various issues.