WPS5625 Policy Research Working Paper 5625 How Integrated Is SADC? Trends in Intra-Regional and Extra-Regional Trade Flows and Policy Alberto Behar Lawrence Edward The World Bank Development Research Group Trade and Integration Team April 2011 Policy Research Working Paper 5625 Abstract Do Southern African Development Community this sense, and contrary to stylized fears, the Southern countries trade enough with each other and with the African Development Community region is quite rest of the world? Although its share of world trade has integrated. Although the Southern African Development fallen, appropriate benchmarking shows that, controlling Community has reduced its tariffs, the structure remains for gross domestic product and other characteristics, complex and could be lowered on intermediates. Other Southern African Development Community countries impediments make it costly and difficult to move goods, have experienced an increase in openness that is but are at levels that are comparable with countries at comparable to other developing countries. Once similar levels of development. Although this may be market size and geography are taken into account, trade surprising, there is still scope for improvement and the between Southern African Development Community disadvantageous geography of the Southern African countries is actually high. Southern African Development Development Community makes it important for other Community countries also trade more products with trade impediments to be reduced. each other than they do with the rest of the world. In This paper is a product of the Trade and Integration Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at abehar@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team How integrated is SADC? Trends in intra-regional and extra-regional trade flows and policy* Alberto Behar and Lawrence Edwards** * Paper is a condensed version of a chapter written for a book as part of a World Bank Regional Investment Climate Assessment of SADC. **respectively Development Economics Group, World Bank and School of Economics, University of Cape Town. 1) INTRODUCTION International trade is viewed as one of the key factors underlying the success of the fastest growing economies yet many countries remain isolated and have failed to achieve this integration. To what extent are the SADC countries afflicted? To answer this question, this paper looks at recent trends in intra-regional trade flows and in the regions trade with the rest of the world. Its analytic objective is to identify the extent to which SADC economies have integrated into the global trading system as well as the regional market. In addition, the paper addresses some of the main institutional and policy barriers to trade in the region. The extent of market integration is evaluated using trade flow data. SADCs aggregate trade performance since 1990 is benchmarked against international counterparts. One aim of this analysis is to identify whether SADC is being marginalized in world trade. The results in section 2 show that SADC has continued to experience a decline in its share of world trade over the past decade and a half. However, the decline also reflects relatively poor economic growth and not necessarily structural impediments to trade. Benchmarked against GDP, SADC has experienced an increase in openness over the past decade that is comparable to other developing countries. Similarly, regressions indicate increased exports by SADC members over time even controlling for GDP growth. This change is equal in magnitude for SADC exporters as other countries. SADC has therefore become more integrated with the world economy over the past decade and a half and the extent of this increase in integration is equivalent to other comparable countries. In section 3, the paper then looks at intra-SADC trade. Intraregional trade rose as a share of total trade during the 1990s, but progress in this regard has slowed recently. The level of and trends in intra-regional trade, however, are not even across countries. South Africa has become an important source of imports for SADC countries with the ending of sanctions in the early 1990s, yet imports by South Africa from the SADC region remain small. Regression results suggest that, ceteris paribus, SADC countries trade with each other more than they do elsewhere. The coefficients indicate they trade with each other more than twice as much as other pairs do. Trade is in this sense regionalized in SADC. We continue with an analysis of product market integration in section 4. High levels of concentration are found; the top 10 products at the 6-digit HS level account for upwards of 70% of intra-SADC trade flows for each country. Most of these products are resource-based, which reflects the comparative advantage of the region. Such high levels of concentration are not unique to intra-SADC trade. In fact, exports to non- SADC members appear to be even more concentrated. Additional measures that compare the proportion of possible goods traded yield similar outcomes. Most SADC countries actually export more products to the region than the rest of the world. The product composition of exports to the rest of SADC also differs from the product composition of exports to the rest of the world. To the extent that product market integration leads to a greater dispersion or diversification 2 of trade, these trends would be indicative of relatively high levels of market integration within the region. The findings are consistent with work for sub-Saharan Africa. They imply that SADCs trade performance is not sub-par, but this does not mean SADC is trading enough. What then is the scope for further increases in trade? Tariffs on imports are a key policy instrument available to government to influence product market integration. In section 5, we show SADC members have made significant progress in reducing barriers to trade. Trade barriers between members have largely been eliminated under the SADC Free Trade Agreement. MFN rates have also fallen. The SADC region now faces a trade policy environment that is more conducive towards promoting intra- and extra-regional trade flows and product market integration. Nevertheless, scope remains for further MFN reform, particularly of tariffs on intermediate inputs. The structure of tariffs also varies substantially across SADC countries, remains complex in many countries and inhibits regional trade flows by necessitating complex rules of origin. Further, widely varying tariff structures will inhibit negotiations on a common external tariff required under the proposed customs union. In section 6, the data indicates that trade is costly and difficult in many SADC countries. However, using a benchmarking exercise that considers SADCs geography and level of development, trade impediments are not uncharacteristically high. Section 7 briefly concludes. 3 2) IS SADC BEING MARGINALIZED IN WORLD TRADE? Sub-Saharan Africas (SSA) share of world trade declined dramatically in the second half of the 20th century (Amjadi and Yeats 1995; Amjadi, Reincke and Yeats 1996; Ng and Yeats 1996). This section analyzes the persistence of these trends in recent years focusing in particular on SADC countries. Figure 1 compares the value (US$) and volume of merchandise exports by SADC with the rest of Sub-Saharan Africa, the ,,world and developing countries excluding SSA. Growth in exports from SACU and the rest of SADC was mediocre during the 1990s relative to the rest of the world and other developing countries, but rose strongly from 2002. Growth in the dollar value of exports was particularly strong and can be attributed to improved terms of trade associated with the commodity price boom. Export performance evaluated in terms of volumes is more mediocre after 2002, particularly for the rest of SSA where oil-rich Nigeria dominates. Figure 1: Merchandise exports Current prices (US $), 1990=100 Constant prices, 1990=100 1200 700 600 1000 500 800 400 600 300 400 200 200 100 0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 1990 1992 1994 1996 1998 2000 2002 2004 2006 Rest of SADC Rest of SSA SACU World Developing excl. SSA Source: Own calculations using World Development Indicators. Export volumes were converted to values using 2000 prices. The sample consists of 99 countries, 37 of which are from SSA (15 from SADC), 41 from other low and middle-income countries and 21 from high-income countries. Volume data for Namibia are not available, but the country is included in the figure based on current prices. Only countries for which data are available in each year are included in the sample to avoid changes in trade values arising from changes in the country composition. As a consequence, while the SADC share of world trade in current US dollars in 2008 (1.6 percent) was marginally higher than its share in 1990 (1.5 percent), in real terms its share declined from 1.3 percent to 0.95 percent over this period (see Figure 2).1 The apparent marginalization of SADC in world exports is more drastic when compared with other developing countries. The share of SADC in developing country real exports fell from 7.1 to 2.9 percent. 1 There is substantial heterogeneity in export performance across countries within the SADC area. Real export growth since 1990 exceeded the world average for five of fourteen countries while Mozambique and Lesotho exceeded the rest of the developing world. 4 Figure 2: SADC merchandise exports as a share of world exports and developing country exports, nominal and real values Share SADC in World exports and Developing country exports 10% 9% 8% Nominal exports, 7% share Developing 6% Real exports, share 5% Developing 4% Nominal exports, share World 3% Real exports, share 2% World 1% 0% 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Source: Own calculations using data from World Bank World Development Indicators. See earlier figure for further details. The sample consists of 99 countries, 37 of which are from SSA (15 from SADC), 41 from other low and middle-income countries and 21 from high-income countries. Volume data for Namibia are not available, but the country is included in the figure based on current prices. Only countries for which data are available in each year are included in the sample to avoid changes in trade values arising from changes in the country composition. However, trade values alone misrepresent the extent to which an economy under or over trades as they do not take into levels of GDP (Rodrik, 1997). Figure 3 therefore plots trends in the ratio of exports to GDP for various regions. There is substantial variation across SADC countries, but the data indicate that SADC is relatively open to trade compared to its international counterparts.2 For example, in 2008, the world ratio of exports was about 20% while those for SACU and the rest of SADC were above 30%. The data also indicate that most SADC countries have become more open during the 1990s. 2 A similar observation is reached when using exports plus imports to GDP as the measure of openness. 5 Figure 3: Merchandise exports as a share of Gross Domestic Product, 2000 GDP weighted GDP weighted Exports/GDP ratio 0.40 0.35 0.30 0.25 0.20 0.15 0.10 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 SACU World Rest of SADC Rest of SSA High-income Other developing Notes: Own calculations using World Bank Development Indicators. Region aggregates are constructed using 2000 GDP values as weights. Zimbabwe is excluded. Gravity model based estimates, which control for geographical and other observed and unobserved country characteristics, corroborate these findings (see Appendix 1 for a brief description of the gravity model specifications and the full results in Table 9). Exports from the average country in the sample were 18.5 log points in higher in the 2001-2005 period than the 1991-1995 period, controlling for GDP growth (and unobserved time invariant characteristics) (column 1). The trend is no different for SADC countries, as revealed by the insignificant coefficients on the SADC-period interactions terms. In addition, SADC countries are found to be no more or less prone to trade relative to the rest of the world, controlling for GDP, distance and a number of other geographical features (see the coefficients on the SADC exporter and importer dummies in column 2). 6 3) HOW INTEGRATED IS TRADE IN THE SADC REGION? Table 1 and Table 2 present the share of SADC trade in regional exports and imports from 1980 to 2003. African trade data are notoriously problematic with different datasets providing different values of trade flows (Yeats 1990). Concerns about data quality are particularly relevant for intra-African trade flows. Nonetheless, various insights emerge from the data. Intra-SADC trade grew significantly from 1980, but has halted in recent years. For example, the share of SADC exports destined for the region more than tripled to 9.9 percent from 1990 to 1995, but then rose very gradually to 12.1 percent in 2008 (based on sample excluding Angola, DRC, Madagascar and Seychelles for which earlier data are not available) (Table 2). These trends are corroborated by the gravity model estimates (column 1 of Table 9 in Appendix 1) which imply no significant change in intra-SADC trade over the 1991 to 2005 period (see the insignificant coefficient on the intra-SADC dummy variable). Dependence on the region for trade divides countries into two groups. Malawi, Mozambique, Zambia, and Zimbabwe depend heavily upon SADC, particularly for imports. These countries source upwards of 50 percent of their imports from other SADC countries and sell more than 20 percent of their exports to the region. The remaining countries in SADC maintain much stronger trade relationships with the rest of the world (ROW). For example, intra-regional trade makes up approximately 10 percent of Mauritian exports and imports. SACU sources only 5.6 percent of its imports from the region. SADC accounts for a much higher percentage (10.5) of SACU exports, which leads to large trade imbalances between SACU and the rest of SADC. Substantial asymmetries in trade flows persist. SACU trade, which is predominantly made up of South African trade flows, dominates intraregional trade flows. Between 60 to 70 percent of SADC exports to the region are sold to SACU (Table 1), while 80 to 90 percent of SADC (excluding SACU countries) imports from the region are purchased from SACU (Table 2). The region is therefore more dependent on South Africa as a source of imports than as a market for exports. 7 Table 1 Share of SADC trade in SADC country imports Proporti Proportio on from n from SACU SACU 1980 1985 1990 1995 1999 2003 2008 2003 2008 Angola 0 0.6 0.8 7.1 10 6.5 99.9 DRC 0.4 1.6 1.1 18.1 31.5 42.8 3.3 Malawi 36.7 53 24.8 49.2 64.4 57.5 58.3 65 46.4 Mauritius 14.5 4.2 9.9 11.3 11.2 13.2 9.9 97 84.8 Mozambique 3.7 5 7.6 55.5 58.6 39.5 38.0 97 94.7 SACU 0.1 1.8 1.8 2.1 1.9 2.7 5.6 0.0 Tanzania 0.7 0.7 1.3 13.9 13.3 15 11.5 66 90.4 Zambia 1.2 10.9 7.9 49.1 65.5 65 59.1 95 74.1 Zimbabwe 8.3 31.7 33.1 51.2 51.2 56.1 76.1 94 91.8 Seychelles 10.7 55.6 Madagascar 9.3 72.1 Intra-SADC 1.6 4.7 5.1 9.9 10.2 12.3 90 80.3a excl. Angola, DRC, Madagascar and Seychelles 10.6 12.3 78.8 Source: Development Network Africa (2007) and Chauvin, S. and G. Gaulier (2002) updated to 2008 using HS6-digit data obtained from UNComtrade. Note: Intra-SACU trade is excluded. 2002 values used for Zimbabwe in 2003. a. 80.3 percent of SADC (excluding SACU countries) imports from the region are sourced from SACU. Table 2: Share of SADC trade in SADC country exports Proporti Proportio on to n to SACU SACU 1980 1985 1990 1995 1999 2003 2008 2003 2008 Angola 0.03 0 0.01 0.03 0.7 1.8 99.9 DRC 0.05 0.03 0.1 6 0.3 6.7 3.3 Malawi 12.4 15.4 1.6 17.2 16.9 20.1 21.8 74 50.4 Mauritius 1.4 0.1 1.2 1.4 1.4 2.1 11.3 76 32.8 Mozambique 1.1 0.3 0.2 32.1 17.41 24.6 17.2 74 65.4 SACU 0.7 2.8 2.5 10.7 11.5 9.7 10.5 Tanzania 5.2 0.1 0.5 1.4 7.4 9.4 17.2 45 55.7 Zambia 0.9 3.1 0.8 3.8 7.8 40.6 20.0 50 54.1 Zimbabwe 1.3 25 30.7 31.7 28 30.5 64.4 79 81.2 Seychelles 1.2 Madagascar 3.1 Intra-SADC 0.9 3.4 3.1 9.9 10 6.1 68 72.8a excl. Angola, DRC, Madagascar and Seychelles 10.6 12.1 62.9 Source: Source: Updated table from Development Network Africa (2007) and Chauvin, S. and G. Gaulier (2002). Note: Intra-SACU trade is excluded. 2002 values used for Zimbabwe in 2003. a. 72.8 percent of SADC (excluding SACU countries) exports to the region are sold to SACU. Further insights on the regionalization of SADC trade are provided by the gravity model. These estimates (Column 3 of Table 9 in Appendix 1) reveal that SADC trade is regionalized: intra-SADC trade is relatively high in relation to what intra-regional incomes and distance would predict. For example, the coefficient on the intra-SADC dummy is significant and at 0.976 suggests that SADC countries trade more than double what would be implied by the gravity model benchmark. Additional specifications indicated that these results are robust to the exclusion of South Africa. The results are consistent with those of earlier research for SSA by Foroutan and Pritchett (1993).3 3 See also Subramanian and Tamirisa (2003) and Yang and Gupta (2008). 8 4) HOW DIVERSIFIED IS SADC TRADE? High barriers between countries reduce not only the volume of a particular product being exported to a particular destination, but also the number of different products being exported there (Melitz, 2003). Therefore, we would expect greater integration to manifest itself in the form of larger varieties of products being traded. To enhance our understanding of SADC integration, this section now assesses the diversity of products being traded; it first reveals the share of exports accounted for by the top 10 export products before presenting measures of the ,,thickness of trade. Figure 4 presents the share of total exports to SADC and the Rest of the World (ROW) made up by the top 10 export products by value to each region in 2008. What is striking is the high level of export concentration. The top 10 export products account for over 60 percent of exports for SADC members outside of SACU. In some cases (Malawi, Mozambique, Zimbabwe) the top 10 products account for over 90 percent of export volumes. The concentration of SADC exports, however, is higher with the Rest of World than with SADC members. In most cases the export concentration ratio is 15 percentage points or more greater for exports to the rest of the world than it is for exports to other SADC members. This is consistent with the aggregate analysis of intra-SADC trade and corroborates the finding that SADC trade is regionalized. Figure 4: Share of top 10 export products in total exports to SADC and Rest of world (ROW), 2008 120 100 80 60 40 20 0 Zambia Zimbabwe Tanzania Madagascar Namibia Mozambique Malawi SA Mauritius Seychelles Exports to SADC Exports to ROW Notes: Own calculations using HS 6-digit 2008 data obtained from UNComtrade. Looking at the actual products exported in Table 3, the top 10 products to the Rest of World and SADC are comprised mainly of primary products, although Mauritius and Malawi are also exporters of clothing and textile products.4 A further observation is 4 Not shown in this table is the relatively high level of apparel exports from Swaziland and Lesotho associated with preferences under the African Growth and Opportunity Act (Portugal-Perez 2008; 9 that the product compositions of trade flows to the region and to the rest of the world differ. There is some overlap, but products that make up a high proportion of exports to the SADC area often make up only a small proportion of the countrys exports to the rest of the world. Export market "Thickness" An alternative quantity-based measure of market integration is a modification of the market "thickness" indicator (Ij) developed by Knetter and Slaughter (2001), which measures the share of all possible products that a country actually exports.5 Lower trade barriers between countries are expected to increase the range of products traded.6 Table 4 presents the total number of products and proportion of all possible HS 6-digit products (in parentheses) exported by each SADC country to the region and to the rest of the world in 2008. The final column present the simple average number of products exported to each SADC member (excluding Botswana, Lesotho, Swaziland and Namibia). Madagascar, for example, exports on average 89 products (2 percent of the 5222 HS6-digit product lines) to each SADC member. However, because the same products are not exported to each SADC countries, the total number of distinct product lines exported by Madagascar to the SADC region is substantially higher at 725 (14 percent of all possible products).7 The table offers the following insights: SADC countries export a relatively high proportion of all possible 6-digit products, despite the concentration levels shown earlier. Almost all the SADC countries export over 30 percent of all 6-digit products, with SA exporting a high 89 percent of possible products. Most of these make up very small values for SADC countries and some of the products exported are probably re-exports. There is substantial heterogeneity in the product composition of each SADC countrys exports across destinations. In other words, SADC countries export different products to different countries. The average number of distinct products exported by SADC countries to other members is low relative to the total number of products exported by the country to the region. It is only South Africa that exports a relatively high proportion (52 percent) of 6 digit products to all other SADC members. Frazer and van Biesenbroeck 2010). Note also the data quality issues, as is revealed by the export of Helicopters, which make up 9 percent of Madagascars exports to the region, in 2008. 5 Let zijk be a categorical variable that is equal to one if country j has some positive value of exports to country k in product i. With N possible products and K-1 destinations (excluding country j) in the sample, the maximum number of bilateral exports by country j equals Nx(K-1). The thickness measure Ij is then calculated as Ij i k j zijk ( N ( J 1)) . 6 The indicator provides no information about the volume of trade. Lower trade barriers that lead only to an increase in the volumes of trade between countries will leave the indicator unaffected. Similarly, lower trade barriers that result in national production becoming more specialized will result in a reduction in the ,,thickness of trade according to this measure. These caveats need to be considered when interpreting the results. 7 Note that the difference between the average number of products and the total number exported to the rest of SADC (e.g. 89 vs. 725 for Madagascar) is an indicator of differences in the product composition of exports to the other SADC members. 10 Table 3: Description and share of top 10 HS 6-digit export products by value to SADC and ROW, 2008 Top 10 exports to SADC by product Madagascar Malawi Mauritius Mozambique Seychelles SA Tanzania Zambia Zimbabwe Petroleum oils and o 20.24 Cotton, not carded 11.39 T-shirts, singlets, 7.24 Electrical energy 53.77 Contnrs fr cmprssd 25.94 Petroleum oils and o 8.79 Gold, nonmonetary, 38.15 Copper ores 22.27 Nickel mattes 13.68 Helicopters of an u 9.13 Other knitted or cro 9.67 Men's or boys' trou 5.09 Petroleum oils and o 6.80 Animal feed prep 25.79 Corn (maize), other 3.92 Wheat or meslin flo 7.77 Copper wire, refine 11.39 Nickel ores and con 12.18 Coniferous wood saw 3.85 Sunflower seeds, wh 8.32 Animal feed prep ex 4.94 Tobacco 4.19 Cigarettes containi 5.50 Trucks, nesoi, dies 2.71 Prec nesoi & semipr 5.33 Tobacco 6.35 Unused postage 5.80 Shrimps and prawns, 3.54 Tobacco 7.11 Looped pile fabrics 3.41 Bran sharps & oth r 1.43 Fish, nesoi, with b 4.68 Structures and part 2.00 Gold, nonmonetary, 3.58 Electric conductors 4.81 Trucks, nesoi, dies 5.06 Boring or sinking m 2.96 Corn (maize), other 6.11 Yarn, carded wool, 3.30 Unused postage, che 1.35 Fish fats & oils (n 4.16 Fertilers contain n 1.64 Soap in forms nesoi 1.77 Cane sugar, raw, so 4.02 Auto regulating ins 4.88 Salt incl tbl/dentr 2.51 Black tea fermdt & 5.94 Men's or boys' shir 2.79 Cotton seeds, wheth 1.34 Flour meal & pellet 3.92 Other slag and ash, 1.41 Oth furn art exc he 1.74 Corn (maize), other 3.42 Cotton yarn> 4.30 Peas, dried shelled 2.09 Cane sugar, raw, so 5.82 Woven cotton fabric 2.75 Bananas and plantai 1.27 Crustcns nesoi lve/ 2.10 Wheat (other than d 1.38 Portland cement exc 1.74 Wheat or meslin flo 2.44 Liq dielect transfr 3.69 Men's or boys' shir 2.02 Natural rubber in p 3.81 Art for conveying o 2.63 Tobacco, partly or 1.19 Water, mineral 1.89 Electrical energy 1.28 Urea, whether or no 1.55 Refined copper cath 2.44 Men's or boys' suit 3.63 Jerseys, pullovers, 2.02 Men's or boys' othe 3.05 Men's or boys' shir 1.96 Petroleum oils and o 1.13 Motorboats, other t 1.74 Manganese 1.13 Palm oil, refined b 1.45 Portland cement 2.42 Cut flowers/buds dr 3.36 Parts for boring or 1.94 Plywood, veneer pan 2.69 Wov cot fab, dye pl 1.58 Shrimps and prawns 0.73 Watermelons, fresh 1.58 Bituminous coal, no 1.02 W/g blouses shirts 1.11 Plates sheets strp 2.19 Esters of acetic ac 2.54 8941. 1089. Trade Value (US$ m) 51.7 191.9 234.5 421.0 2.9 4 367.4 1016.0 5 share of top 10 50.32 63.91 35.68 73.20 77.32 25.27 64.19 61.73 59.12 share top 10 in 2003 to SADC 65.2 61.60 70.20 19.50 75.50 73.20 68.20 Top 10 exports to ROW by product Madagascar Malawi Mauritius Mozambique Seychelles SA Tanzania Zambia Zimbabwe Men's or boys' trou 15.6 Tobacco 60.3 T-shirts, singlets, 16.4 Unwrought aluminum 71.4 Tunas/skipjack/boni 37.9 Bituminous coal, no 6.8 Gold, nonmonetary, 22.0 Refined copper cath 51.1 Cut flowers and flo 24.6 Shrimps and prawns, 6.7 Tobacco, partly or 22.8 Cane sugar, raw, so 15.8 Tobacco, partly or 7.9 Petroleum oils 37.0 latinum, unwrought 5.6 Precious metal ores 11.3 Plates sheets strp 23.7 Tobacco, partly or 13.3 Women's or girls' t 6.6 Cane sugar, raw, so 5.8 Tunas/skipjack/boni 11.6 Shrimps and prawns, 2.9 Fish nesoi, salted 18.2 Ferrochromium 5.2 Coffee, not roasted 6.3 Copper ores 11.4 Cotton, not carded 11.6 Petroleum oils and o 5.0 Black tea fermdt & 3.7 Men's or boys' shir 6.3 Cotton, not carded 2.4 Parts of airplanes 1.4 Filter/purify machi 4.6 Tobacco, partly or 4.8 Cobalt and articles 7.2 Ferrochromium 5.9 Jerseys, pullovers, 4.1 Peas, dried shelled 1.4 Transmission appr i 4.3 Sesame seeds 1.5 Fish fats & oils (n 1.1 Platinum metal, sem 4.1 Fish fillets & oth 4.6 Ash and residues ne 0.7 Collectors items of 4.8 Airplane & ot a/c, 3.8 Sweaters, pullovers 1.0 Men's or boys' trou 3.6 Titanium ores and c 1.4 Surveying instrment 0.7 Pass veh spk-ig int 4.0 Fish fillets, froze 3.8 Prec nesoi & semipr 0.7 Peas 4.1 Women's or girls' t 3.8 Nuts nesoi, fresh o 0.6 Fish, nesoi, with b 2.9 Petroleum oils and o 1.1 Instr & appl f medi 0.5 Agglomerated iron 3.6 Cotton, not carded 2.2 Copper wire, refine 0.6 Peel, citrus or mel 3.0 Jerseys, pullovers, 3.7 Sweaters, pullovers 0.6 Men's or boys' shir 2.6 Cashew nuts, fresh 0.7 Breathing appliance 0.4 Manganese ores 2.9 Black tea fermdt & 2.0 Cobalt ores and con 0.6 Citrus fruits, inc 2.6 Wooden bedroom Rhodium, W/g blouses, shirts 3.4 furniture 0.2 T-shirts, singlets 2.4 Nonconiferous wood 0.7 Fish, nesoi, with b 0.3 unwrought 2.9 Peas, dried shelled 1.9 Precious metal ores 0.5 Elect appr f prtct 2.3 Vanilla beans 3.1 Coniferous wood 0.2 Diamonds 2.0 Vessels,nesoi,for t 0.7 Flour meal & pellet 0.2 Diam ex ind unwkd 2.4 Wheat (other than d 1.7 Vegetables mixtures 0.5 Tobacco 2.2 1841. Trade Value (US$ bill) 1599.6 687.1 7 2033 241.5 64400 1769.5 4054.9 603.4 share of top 10 55.8 96.7 68.0 90.8 97.8 42.3 60.7 96.8 74.2 share top 10 in 2003 to ROW 90.9 69.2 92.6 44.5 83.9 86.6 93.7 Source: Own calculations using SADC Trade Database. Data are classified at the 6-digit HS level. Tanzania data is for 2007. 11 Many SADC countries export a more diverse range of products to the SADC region than to the rest of the world, which is indicative of the regionalization of SADC trade. For example, Malawi exported 1008 distinct HS 6-digit products to the rest of SADC, but only 435 to the rest of the world. Mauritius, Namibia, Mozambique, South Africa, Zambia and Zimbabwe, similarly export a more diverse range of products to the rest of SADC. Table 4: Export-thickness measures for SADC countries, 2008 Total products to: Average to each Exporter World SADC ROW SADC country Madagascar 1937 (37) 725 (14) 1821 (35) 89 (2) Malawi 1122 (21) 1008 (19) 435 (8) 171 (3) Mauritius 2564 (49) 2057 (39) 1831 (35) 216 (4) Mozambique 1507 (29) 1137 (22) 952 (18) 148 (3) Namibia 3448 (66) 3272 (63) 1703 (33) 611 (12) Seychelles 511 (10) 182 (3) 440 (8) 19 (0) South Africa 4667 (89) 4477 (86) 4286 (82) 2737 (52) Tanzania 2183 (42) 1404 (27) 1672 (32) 214 (4) Zambia 2216 (42) 2127 (41) 538 (10) 353 (7) Zimbabwe 1768 (34) 1545 (30) 760 (15) 290 (6) SADC combined 4823 (92) 4704 (90) 4482 (86) 2965 (57) Notes: Own calculations using HS 6-digit level data obtained from UNcomtrade. Values for Tanzania are for 2007. Note that the SADC region on the import side includes the SACU members, except for SA who does not declare exports to the other SACU countries. The total number of HS2002 codes in UNcomtrade is 5222. Data is for gross exports and therefore contains re-exports To summarize, intra-SADC trade is low, but this is partly a consequence of low levels of economic development. Once we condition on income levels, SADC countries have experienced an increase in openness comparable to other developing countries. Intra-SADC trade is also found to be relatively high and diversified. Nevertheless, there are some worrying trends. Growth in intra-regional trade has slowed in recent years. Exports from the region continue to decline as a share of world trade. Therefore, in the following sections we look at the role of tariffs and other factors that impede further growth in SADC trade. 12 5) TARIFF BARRIERS TO TRADE Tariffs remain a powerful instrument through which government can directly influence international trade and product market integration even though they are not necessarily the most important barrier to economic integration (Anderson and van Wincoop, 2004). Tariffs restrict imports and introduce a wedge between domestic and international prices. It is less recognized that tariffs are a tax on exports.8 Tariffs on intermediate inputs raise production costs and adversely affect the ability of exporters to compete internationally. Additionally, by raising the relative profitability of supplying the local market, scarce resources are drawn away from export competing sectors. Finally, lower imports and exports can cause the currency to appreciate, creating further adverse incentives to produce for the export market. Tariff developments in SADC countries have taken two forms: Regional integration and multilateral/unilateral reform. On the regional front, SADC countries have actively participated in regional integration schemes. Internal tariff barriers were largely eliminated by the time SADCs Free Trade Agreement was launched in 2009 (http://www.sadc.int).9 The formation of the FTA is considered the first step in a much grander integration program including a Customs Union supposedly by 2010, a Common Market (CM) by 2015, a Monetary Union (MU) by 2016 and a Single Currency by 2018 (http://www.sadc.int). Yet sub Saharan Africa is characterized by a plethora of overlapping regional integration arrangements, each with their own proposed tariff schedules and rules of origin. Of SADC members, only Mozambique is not a member of another arrangement. Multiple memberships by SADC countries in existing or proposed customs unions (Tanzania in EAC, COMESA which includes all SADC members except for South Africa, Botswana and Mozambique) is inconsistent with the proposed formation of a SADC Customs Union. This dilemma of multiple memberships also extends to other areas such as infrastructure, where different harmonization options and strategies are being pursued (Kritzinger-van Niekerk and Moreira 2002). The second type of tariff reform is multilateral or unilateral liberalization. On this front, SADC countries have made considerable progress in reducing import barriers since the early 1990s. Table 5 presents summary statistics of the 2008 MFN tariff rates applied by each SADC country. Simple average tariffs range from 2.9 percent in Mauritius to 25.5 percent for Zimbabwe.10 For most countries, average MFN rates 8 This is the well known Lerner Symmetry theorem. Edwards and Lawrence (2008) find that reductions in the export bias associated with tariff liberalisation accounted for approximately 40 percent of South Africa's export growth after 1994. Similarly, Edwards (2010) finds that liberalisation of tariffs on intermediate inputs is associated with improved manufacturing export performance in Africa. 9 South Africa, for example, reduced most tariffs on SADC imports to zero in 2000 (Edwards 2005), while other countries phased their tariffs down at a slower rate (http://www.sadc.int/fta). 10 The average rate for Zimbabwe appears to be particularly high even compared to earlier period. For example, the WTO World Tariff Report for 2007 estimates average protection in Zimbabwe to equal to 14 percent. The 2006 WTO World Tariff Profiles indicate a simple average of 16 percent for Zimbabwe for 2003. This highlights the difficulty in estimating average protection for countries. Many of the countries, SACU in particular, used non-ad valorem rates such as specific rates, mixed rates and compound rates. SACU also used formula duties which we find a reservation price. If import prices fell 13 range from 7 to 14 percent, which situates them in the range for low-income and upper-middle-income countries. These current tariff levels in SADC countries are considerably lower than they were during the early 1990s, despite limited offers made in the Uruguay round to reduce bound rates (Wang and Winters, 1998). As shown in Table 6, the simple average MFN tariff rate applied by SADC countries fell from 18.8 percent in 1997 to 10.2 percent in 2007.11 The average decline in tariffs is therefore comparable to other developing countries (Edwards, 2005). Declines in protection were particularly high in Mauritius, Seychelles, Malawi and Tanzania, but these reductions came off a high base: initial tariffs rates in these countries exceeded 20 percent in 1997. Madagascar appears to be the exception with average tariffs rising from 6.9 percent in 1995 to 12.4 percent in 2007. It is not clear from the data whether this reflects the replacement of non-ad valorem tariffs with ad-valorem rates or actual increases in tariff protection. Also of interest are variations in the degree of liberalization across different sectors. Table 6 shows average protection declined for all end-use categories (consumer goods, intermediate goods and capital goods) from 1997 to 2007, with relatively strong decreases in tariffs on consumer goods. This is suggestive of a decline in effective protection in SADC countries from the mid-1990s. Nevertheless, tariff escalation remains high. In 2007, the average tariff consumption goods at 19 percent was 3.5 times the average tariff on capital goods and more than double the tariff on intermediate goods. Such escalation of the tariff schedule suggests that effective protection rates on consumer goods are substantially higher than 19 percent. It is only Mauritius that, on average, imposes tariffs of less than 10 percent on consumption goods. Looking beyond average tariff rates, various indicators of complexity in Table 5 reveal enormous differences across the SADC members. For example, Mauritius, Zimbabwe and SACU members have over 150 tariff bands and are followed by Zambia and Seychelles with between 30 and 50. The remaining members impose less than 10 bands. The DRC, Madagascar, Malawi, Mozambique, Tanzania, Zambia and Zimbabwe impose tariffs in excess of 15 percent on 33 to 41 percent of all tariff lines. In contrast, less than 10 percent of tariff rates exceed 15 percent in Angola and Mauritius. below this price additional tariffs were levied (Edwards, 2005). Calculating the ad valorem equivalent of these non-ad valorem rates is sensitive to international prices. 11 The simple average is presented, as the import-weighted average results in a downward bias in the average tariff rate, as numerous highly protected products are not imported. Anderson and Neary (1994) derive alternative "trade restrictiveness" indicators that better reflect the welfare costs of protection (the dead weight loss is proportional to the square of tariff rates) and use import weights adjusted for the distortionary effect of tariffs. However, even these measures of protection require positive import values. Finally, the data does not include ad valorem equivalents of the numerous non- ad valorem rates applied by these countries, particularly during the 1990s. 14 Table 5: Structure of MFN tariffs applied by SADC economies, 2008 South SADC Angola Botswana DRC Lesotho Madagascar Malawi Mauritius Mozambique Namibia Seychelles Africa Swaziland Tanzania Zambia Zimbabwe average year 2008 2008 2008 2008 2008 2008 2008 2008 2008 2007 2008 2008 2008 2008 2008 Number of tariff lines 5201 6671 5794 13348 6362 5397 12516 5203 6671 5122 6671 6671 5260 5984 5899 6851 Complexity Number of bands 7 157 4 157 5 6 283 6 157 51 169 157 14 31 372 105 Duty free lines (% total) 0 59.5 0 59.5 1.9 9.8 87.7 2.9 59.5 87.2 59.5 59.5 37.2 19.3 6.2 36.6 Non-ad valorem (% lines) 0 2.3 0 2.6 0 0 2.8 0 2.3 0.6 2.3 2.3 0.2 2.1 6.8 1.6 Binding coverage (%) 100 96.6 100 100 29.7 31.2 17.9 13.6 96.6 96.6 96.6 13.4 16.8 21.2 59.3 All products 7.3 7.8 12 7.8 12.5 13 2.9 10.1 7.8 8.2 7.8 7.8 12.6 13.8 25.5 10.5 Average Non-agriculture 6.9 7.6 11.9 7.5 12.1 12.6 2.7 9.5 7.6 6.4 7.6 7.6 11.5 13 25.5 10 rates Agriculture 10 9.4 12.8 9.4 14.7 15.5 4.2 13.8 9.4 19.7 9.3 9.4 19.9 19.3 25.4 13.48 Maximum rate 30 346 30 141 20 25 286 20 346 786 >1000 346 113 66 >1000 197 Domestic spikes (3*average) (% lines) 2.5 9 0 9 0 0 11.7 0 9 10.8 9 9 0.7 0.1 5.6 5.1 Dispersion International (>15%) (% lines) 10 21 35.2 21 38.3 36.9 5.5 33.5 21 10.8 21 21 40.7 33.2 35 25.6 Coefficient of variation 92 154 51 138 50 73 333 72 154 422 206 154 95 73 215 152 Source: WTO World Tariff Profiles 2009. 15 Table 6: Simple average applied MFN tariff, percent 1997 2001 2007 Change 97-07 % Angola 8.81 7.2 -1.48a Madagascar 6.94 4.61 12.4 5.11 Malawi 25.3 13.1 13.3 -9.58 Mauritius 28.7 18.4 3.15 -19.85 Mozambique 15.7 13.8 10.3 -4.67 Seychelles 28.3 7.12 -16.51a SACU 11.3 8 7.74 -3.20 Tanzania 24.3 16.3 12.6 -9.41 Zambia 14.1 12.6 13.7 -0.35 Zimbabwe 23.8 19.6 14.1 -7.84 Pooled simple average 18.8 14.4 10.2 -7.24 Pooled import-weighted 8.42 6.95 6.45 -2.00 Tariffs by End-Use Consumption goods 31.3 26.3 19.7 -8.9 Intermediate inputs 15.2 11.0 8.7 -5.7 Capital goods 12.4 8.1 6.2 -5.5 Source: Team calculations using TRAINS data at HS 6-digit level; SACU tariffs from 1997 are obtained from Edwards (2005) Notes: a. Change is based on the 2001-2007 period. *1995 tariff used in Madagascar for 1997 period *2002 tariff used for 2001 period for Mauritius, Zambia and Angola *2006 tariff used for 2007 for Angola and Malawi. 2008 tariff used for Zambia The end-use classification is based on the BEC classification obtained UN Statistics. Passenger vehicles are excluded as they are both a capital and consumption good. The percentage change in the tariff inclusive border price is calculated as (t1 t0)/(1 + t0), where t1 and t0 refer to tariff rates in the final and initial periods, respectively. These rates do not include ad valorem equivalents and are therefore not directly comparable to those obtained from the WTO World Tariff Profiles 2009 and used in the prior table. In addition to regional or unilateral measures to reduce tariffs, many SADC countries are members of the WTO and benefit from preferential access to developed markets through the Generalized System of Preferences, Everything But Arms and the Africa Growth and Opportunity Act (AGOA). For example, preferential access for apparel is important for Lesotho, Botswana, Namibia, Madagascar and Malawi, who have all seen a rise in exports as a result of AGOA (de Melo and Portugal-Perez, 2009). In sum, the SADC FTA has largely eliminated within SADC tariffs, but many countries remain members of multiple overlapping and sometimes inconsistent agreements on the continent. MFN tariffs fell from the mid-1990s, although the degree of liberalization varies across members and effective protection remains high on consumer goods. The structure of tariffs is also complex in some SADC countries and will inhibit product market integration, despite the formation of a free trade area. 16 6) OTHER TRADE IMPEDIMENTS It is by now accepted that institutional, infrastructure and regulatory burdens present obstacles to the movement of goods across borders. Obstacles include poor infrastructure (Limão and Venables, 2001), market regulations that restrict competition in transport (Teravanithorn and Raballand, 2008) and weak micro-level institutions, including port efficiency, customs environment, regulatory environment and policies affecting cost of entry (Johnson and others, 2007; Wilson and others, 2005; Njinkeu and others 2008). These micro-level institutional effects are often greater impediments to African trade than tariff barriers (Portugal-Perez and Wilson 2009). For example, it costs more than twice as much to clear a standard 20-foot container for exports or imports in SSA and SADC countries as in the East Asia & Pacific (Table 7). Costs are particularly high in Zimbabwe, Botswana, Zambia and Congo (DR). The time taken to export and import is also high in SSA and SADC countries compared to other regions: more than three times that of the OECD and twice that of Latin America & Caribbean (Table 7). Table 7: Time delays and trade costs Cost to Cost to Documents Time to export Documents Time to import (US$ to export export (US$ per to import import per Region or Economy (number) (days) container) (number) (days) container) East Asia & Pacific 6.7 23.1 909.3 7.1 24.3 952.8 Eastern Europe & Central Asia 6.5 26.8 1581.8 7.8 28.4 1773.5 Latin America & Caribbean 6.8 18.6 1243.6 7.3 20.9 1481 Middle East & North Africa 6.4 22.5 1034.8 7.4 25.9 1221.7 OECD 4.3 10.5 1089.7 4.9 11 1145.9 South Asia 8.5 32.4 1364.1 9 32.2 1509.1 Sub-Saharan Africa 7.8 33.6 1941.8 8.8 39.4 2365.4 SADC 7.4 35.1 1903.7 8.8 42.4 2348.3 Angola 11 65 2250 8 59 3240 Botswana 6 30 2810 9 41 3264 Congo, Dem. Rep. 8 44 2607 9 63 2483 Lesotho 6 44 1549 8 49 1715 Madagascar 4 21 1279 9 26 1660 Malawi 11 41 1713 10 51 2570 Mauritius 5 14 737 6 14 689 Mozambique 7 23 1100 10 30 1475 Namibia 11 29 1686 9 24 1813 South Africa 8 30 1531 9 35 1807 Swaziland 9 21 2184 11 33 2249 Tanzania 5 24 1262 7 31 1475 Zambia 6 53 2664 9 64 3335 Zimbabwe 7 53 3280 9 73 5101 Source: World Bank Doing Business Survey. http://www.doingbusiness.org/ExploreTopics/TradingAcrossBorders/ [Accessed 10 June 2010] Yet simple unconditional cross-country comparisons such as these do not take into account the interdependent relationship between geography and trade costs. African countries are often far from developed markets and many countries are landlocked, all of which raise internal and external transport costs (Behar and Venables, 17 forthcoming).12 Rather than indicating policymakers are doing a bad job, it could show that nature makes their job harder than elsewhere. Furthermore, various composite indicators of the trade environment are closely correlated with the level of development. This is most clearly reflected in Figure 5 which plots the 2009 World Economic Forum Enabling Trade Index (ETI) against log GDP per capita ­ a proxy for development. The fitted (quadratic) line is also included. Similar plots based on logistics quality and time delays are presented in Appendix 2. Analogous to our benchmarking of trade flows earlier in this paper, we benchmark SADCs trade environment against other countries with a similar level of development. For example, Figure 5 reveals high ETI values in many SADC countries relative to their GDP per capita. This suggests that these countries are not unusually constrained by their environment relative to their peers. Figure 5: Overall Enabling Trade Index against log GDP per capita Overall Enabling Trade Index 6 5 Mauritius Botswana 4 Namibia South Africa Malawi Madagascar Zambia Lesotho Tanzania Mozambique 3 Zimbabwe 4 6 8 10 12 ln(GDP per Capita), 2005 PPP Source: Team calculations using the World Economic Forum Enabling Trade Indices (ETI) obtained from the WEF Global Enabling Trade Report (2009) and GDP data from the World Bank World Development indicators. Notes: ETI indicators range from 1 (worst outcome) to 7 (best outcome). Values for Botswana are obtained from the 2010 edition Table 8 presents the results of a benchmarking exercise to identify the relative performance of SADC in terms of its institutional and other obstacles to trade. Composite indicators of the trade environment are regressed on GDP per capita and its square as well as population, area and a dummy for whether the country is landlocked.13 A dummy for the SADC region is included to estimate the extent to 12 The gravity model estimates presented in column 4 of Table 9 in the appendix reveal that SADC countries are disproportionately disadvantaged by their distance from markets. 13 Note that this equation does not account for the endogeneity of trade-related institutional environment and level of development, although lagged GDP levels (2005 values) are used in the regression. High trade costs, for example, may contribute towards low levels of development. The equation is therefore not to be interpreted as a causal relationship. Rather, it is used a simple benchmarking exercise. 18 which institutional and other obstacles to trade in this region deviate from expected values based on income, population and geography. The composite indicators are drawn from the World Economic Forum (WEF) Enabling Trade Index (ETI), the World Bank Logistics Performance Index (LPI) and an index constructed from the World Bank Trading on Time variables presented in Table 7.14 Higher values reflect fewer obstacles to international trade. Table 8: Determinants of Trade Cost indicators ln Ln SADC (GDP/Capita) (GDP/Capita)2 Ln(area) ln(population) Landlocked dummy N R2 Overall index -0.168 0.028 -0.265*** 0.117*** -0.918*** -0.085 187 0.605 Trading on Import index 0.032 0.017 -0.245*** 0.103** -0.990*** -0.149 187 0.603 Time index Export index -0.368 0.039* -0.286*** 0.131*** -0.847*** -0.022 187 0.573 Overall index -1.144** 0.093*** -0.115** 0.039 -0.071 0.214** 135 0.757 Market access 0.514 -0.028 -0.037 -0.082 0.157 0.540** 135 0.131 Border administration -1.476** 0.123*** -0.204** 0.118** -0.319** 0.224** 135 0.745 Efficiency of customs administration -1.061 0.103** -0.250** 0.188** 0.350* 0.365** 135 0.595 Efficiency of import-export World procedures -0.358 0.049 -0.369*** 0.199*** -1.112*** -0.034 135 0.67 Economic Transparency of border Forum administration -3.013*** 0.219*** 0.007 -0.033 -0.196 0.340*** 135 0.768 Enabling Transport & communications Trade infrastructure -1.695*** 0.139*** -0.132** 0.146*** 0.017 -0.099 135 0.855 Indices) Availability and quality of transport infrastructure -0.789* 0.082** -0.024 0.069 0.014 0.152 135 0.687 Availability and quality of transport services -1.543** 0.121*** -0.227** 0.266*** 0.146 -0.195 135 0.725 Availability and use of ICTs -2.736*** 0.214*** -0.150** 0.107* -0.106 -0.252* 135 0.879 Business environment -1.910** 0.135*** -0.089 -0.027 -0.135 0.186 135 0.599 LPI Score -1.192*** 0.089*** -0.104*** 0.150*** 0.006 -0.017 161 0.806 Customs -1.819*** 0.126*** -0.103** 0.123*** 0.004 0.04 161 0.751 World Bank Infrastructure -1.707*** 0.126*** -0.101** 0.178*** 0.002 -0.045 161 0.828 Logistics International Shipments -0.627** 0.049*** -0.137*** 0.137*** 0.054 0.067 161 0.567 Performance Logistics quality and Index competence -1.237*** 0.094*** -0.103** 0.172*** -0.025 0.023 161 0.773 Tracking and tracing -1.278*** 0.095*** -0.125** 0.176*** 0.002 -0.005 161 0.722 Timeliness -0.663** 0.056*** -0.051 0.123*** -0.009 -0.196** 161 0.657 Notes: Estimates are robust to heteroskedasticity. The WEF based dependent variables range from 1 (worst outcome) to 7 (best outcome). PPP GDP per capita (2005 prices) and population are obtained from World Development Indicators. Internal distances are obtained from CEPII. The World Bank Doing Business (Trading on Time) indices are constructed as follows: The underlying variables (export and import costs, time and documents) are first standardized (mean 1 and variance 1). The simple average of these is then converted to a 1 to 7 scale (1 corresponds to worst outcome, 7 to best outcome). The underlying trading times, documents and costs are the most recent values available from the World Bank Doing Business Indicators (accessed 1 June 2010). The World Economic Forum Enabling Trade Indices are obtained from the WEF Global Enabling Trade Report (2009). Values for Botswana are obtained from the 2010 edition. Area measured using internal distance of country, which is calculated as d ii 0.67 area / (see Head and Mayer (2002). * p<.1; ** p<.05; *** p<.001 In Table 8, the explanatory variables are in columns and each row presents a regression for a different measure of trade costs. The simple regression fits the data reasonably well with 60 to over 80 percent of the variation in the composite indicators 14 The ETI and LPI are indicators of the various institutional, infrastructural and regulatory or policy constraints to the free flow of goods over borders. The WEF Enabling Trade Index, for example, is an index of barriers to trade related to market access, border administration, transport and communications infrastructure, and the business environment. The Logistics Performance Index is a summary of various areas of the logistics environment relating to customs clearance, trade and transport-related infrastructure, logistics services, etc. See WEF Global Enabling Trade Report (2009) and Arvis and others (2010) for further details on each index. 19 explained by the variables. In general, obstacles to trade are higher in countries that are landlocked, poor, vast and unpopulated. The SADC dummies indicate that SADC countries do not face unusually severe obstacles to trade (conditional on geography, population and income) relative to the rest of the world. The obstacles to trade are high, but these reflect particular geographical constraints and correspond closely with their level of development. In fact, the results based on the ETI reveal that the overall trading environment and market access and border administration in particular (but not transport and communications infrastructure) are on average better in the SADC region than the rest of the world, conditional on GDP, population and geography. SADC performs poorly in terms of the timeliness with which its shipments reach the consignee. Regressions including individual country dummies reveal relatively high obstacles associated with logistics in Angola, Botswana, Namibia, Mauritius and Zambia (see Table 10 and Table 11 in Appendix 2). Further, the SACU members, Angola, Zambia and Malawi perform poorly relative to their peers in terms of the required documents, time and cost of exporting and importing. In contrast, Madagascar, Mauritius, Mozambique and Tanzania perform well relative to their income levels. This benchmarking analysis suggests that trade is not particularly constrained in SADC compared to other countries at similar levels of development, although there is variation at the country level. This does not imply that further investments in reducing obstacles to trade will not enhance trade flows from SADC countries. However, investments in these areas will also require an analysis of the comparative costs and (trade) benefits relative to improvements in other institutions rather than those specific to the trade environment. The results are likely to vary by country, which motivates country-level analysis that ideally uses firm level data. Furthermore, trade-related reforms require consideration of the complementarities between the various policy constraints. Building multilane highways will not raise trade if trucks must wait at the border. Port improvements would have limited impact if the problem is getting goods to the coast. For example, Freund and Rocha (2010) highlight the importance of getting goods through transit countries in Africa while Behar, Manners and Nelson (2009) find that your neighbors logistics quality positively influences your own exports. Regional policy co-ordination on reducing obstacles to trade is particularly important for SADC countries, many of which are landlocked. 20 7) CONCLUSION This paper has identified the extent to which SADC economies have integrated into the global trading system, focusing on trade policy reform and intra-regional and extra-regional trade flows. The evidence points towards an increase in integration for SADC countries since the early 1990s. MFN tariffs have been reduced, intra-regional trade flows have increased and trade has risen as a share of GDP. Gravity model estimates confirm the finding of increased integration, as measured by trade to GDP ratios. Further, SADC trade is found to be regionalized; intra-regional trade flows are high relative to predictions. Finally, while obstacles to trade are high in SADC countries, these levels are consistent with their low income levels and adverse geography. The implications are that SADCs trade performance is not particularly bad and its trade policy is not necessarily deficient. This leads one to question whether trade- specific reforms should be a priority. However, many regions continue to implement further reforms so SADC must keep pace. Furthermore, it can be argued that SADC needs to trade more than normal, which requires a trade environment that exceeds benchmarks and doesnt just keep pace with them. Given its unfortunate geography, it is especially important for additional technological and institutional impediments to be minimized. References Amjadi, Azita, and Alexander Yeats. 1995. "Have Transport Costs Contributed to the Relative Decline of Sub-Saharan African Exports? Some Preliminary Empirical Evidence." Policy Research Working Paper 1559, International Trade Division, World Bank, Washington, DC. Amjadi, Azita, Ulrich Reincke, and Alexander Yeats. 1996. "Did External Barriers Cause the Marginalization of Sub-Saharan Africa in World Trade?" Policy Research Working Paper 1586, International Trade Division, World Bank, Washington, DC. Anderson, James and Peter Neary (1994), "Measuring the restrictiveness of trade policy", World Bank Economic Review, 8, 151-169. Anderson, James E., and Eric Van Wincoop, 2004, "Trade Costs", Journal of Economic Literature, 42(3), 691-751. Arvis, Jean-François, Monica Alina Mustra, Lauri Ojala, Ben Shepherd, Daniel Saslavsky. 2010. Connecting to Compete 2010: Trade Logistics in the Global Economy, World Bank. Behar, A. and P Manners. 2010. "Distance to growing markets and sub Saharan African exports", mimeo. African Development Review Behar, A, P Manners and B nelson (2009). "Logistics and Exports", Oxford Department of Economics Discussion Paper Behar, A. and A Venables .forthcoming. Transport costs and International Trade, in Handbook of Transport Economics, eds André de Palma, Robin Lindsey, Emile Quinet & Roger Vickerman. Bigsten, Arne., Paul Collier, Stefan Dercon, Marcel Fafchamps, Bernard Gauthier, Jan Willem Gunning, Abena Oduro, Remco Oostedorp, Catherine Pattillo, Mans Soderbom, Francis Teal and Albert Zeufeck. 2004. "Do African manufacturing firms learn from exporting," Journal of Development Studies 4(3): 115-141. Chauvin, S. and G. Gaulier (2002) "Regional Trade Integration in Southern Africa" CEPII Working Paper No 2002-12. Clarke, George. 2005. "Beyond Tariffs and Quotas: Why Dont African Manufacturers Export More?" Policy Research Working Paper 3617, World Bank, Washington, DC. Collier, Paul. 2000. "Africas Comparative Advantage." In H. Jalilian, M. Tribe, and J. Weiss (eds), Industrial Development and Policy in Africa. Cheltenham, UK: Edward Elgar. Development Network Southern Africa (2007) "SADC Customs Union Study: Evaluation of an Appropriate Model for a SADC Customs Union" Report for the SADC secretariate. 21 de Melo, Jaime and Alberto Portugal-Pérez. 2009. "Rules of Origin, Preferences and Diversification in Apparel: African Exports to the US and to the EU," CEPR Discussion Papers 7072 Djankov, Simeon, Caroline Freund, and Cong Pham. 2010. "Trading on Time." Review of Economics and Statistics 92:1, 166-173. Edwards, L. 2005. Has South Africa liberalised its trade. South African Journal of Economics, 73, 4: 754-775. Edwards, L. (2010) "Protectionist Policies and Manufacturing Trade Flows in Africa" in Yifu Lin, J. and Boris Pleskovic (eds.) Annual World Bank Conference on Development Economics 2009, World Bank Publications. Edwards, L. and Lawrence, R. 2008. South African Trade Policy Matters: Trade Performance & Trade Policy, Economics of Transition, 16(4): 585-608. Foroutan, Faezeh, and Lant Pritchett. 1993. "Intra­Sub-Saharan African Trade: Is It Too Little?" Journal of African Economies 2 (1): 74­105. Frazer, Garth and Johannes Van Biesbroeck, 2007 "Trade Growth Under the African Growth and Opportunity Act" Review of Economics and Statistics, Vol. 92, No. 1, Pages 128-144. Freund, Caroline & Rocha, Nadia, 2010. "What constrains Africa's exports ?," Policy Research Working Paper Series 5184, The World Bank. Johnson, Simon, Jonathan Ostry, and Arvind Subramanian. 2007. "The Prospects for Sustained Growth in Africa: Benchmarking the Constraints." IMF Working Paper WP/07/52, International Monetary Fund, Washington, DC. Keith Head & Thierry Mayer, 2002. "Illusory Border Effects: Distance Mismeasurement Inflates Estimates of Home Bias in Trade," Working Papers 2002-01, CEPII research center. Knetter, M. and M. Slaughter: "Measuring Product-Market Integration", in Linda Goldberg (ed.), Topics in Empirical International Economics: A Festschrift in Honor of Robert E. Lipsey, NBER Conference Volume, 2001. Kritzinger-van Niekerk, L. and Emmanuel Pinto Moreira. 2002. Regional integration in Southern Africa: Overview of recent developments. Regional Integration and Cooperation Africa Region Discussion Paper, World Bank. Limão, Nuno, and Anthony J. Venables. 2001. "Infrastructure, Geographical Disadvantage, Transport Costs, and Trade." World Bank Economic Review 15 (3): 451­79. Longo, Robert & Sekkat, Khalid, 2004. "Economic Obstacles to Expanding Intra-African Trade," World Development, Elsevier, vol. 32(8), pages 1309-1321. Melitz, Marc J. 2003. "The Impact Of Trade On Intra-Industry Reallocations And Aggregate Industry Productivity," Econometrica, 71 (6): 1695-1725. Ng, Francis, and Alexander Yeats. 1996. "Open Economies Work Better: Did Africas Protectionist Policies Cause Its Marginalization in World Trade?" Policy Research Working Paper 1636, World Bank, Washington, DC. Njinkeu, Dominique, S. Wilson, John and Powo Fosso, Bruno, (2008), Expanding Trade within Africa: The Impact of Trade Facilitation, No 4790, Policy Research Working Paper Series, The World Bank. Portugal-Perez, A. (2008). "The costs of rules of origin in apparel: African preferential exports to the United States and the European Union. Policy Issues in International Trade and Commodities Study Series No. 39, United Nations Conference on Trade and Development. Portugal-Perez, Alberto and Wilson, John S., 2009. "Why trade facilitation matters to Africa," World Trade Review, Cambridge University Press, vol. 8(03), pages 379-416, July. Rodrik, Dani. 1997. "Trade Policy and Economic Performance in Sub-Saharan Africa." National Bureau of Economic Research Working paper no. 6562. Cambridge, M.A. Santos Silva, J. M. C. and Silvana Tenreyro. 2006. "The Log of Gravity," The Review of Economics and Statistics, MIT Press, vol. 88(4), pages 641-658, 09. Subramanian, Arvind, and Natalia T. Tamirisa. 2003. "Is Africa Integrated in the Global Economy?" IMF Staff Papers 50 (3): 352­72. Teravaninthorn, S and G Raballand, 2008, Transport Prices and Costs in Africa: A Review of the Main International Corridors, The World Bank, Washington. Wang, Z. and Winters, L. A. 1998. "Africas role in multilateral trade negotiations: Past and Future," Journal of African Economies 7 (Supplement 1): 1-33. Wilson, John S., Catherine Mann, and Tsunehiro Otsuki. 2005. "Assessing the Benefits of Trade Facilitation: A Global Perspective." World Economy 28 (6): 841­71. World Economic Forum (2009). Global Enabling Trade Report 2009. Geneva: The World Economic Forum. World Trade Organisation (2009) World Tariff Profiles 2009, World Trade Organisation, Geneva. 22 Yang, Yongzheng and Sanjeev Gupta. (2008) "Regional Trade Arrangements in Africa: Past Performance and the Way Forward", African Development Review, 19(3): 399-431. Yeats, Alexander. 1990. "On the Accuracy of Economic Observations: Do Sub-Saharan Trade Statistics Mean Anything?" The World Bank Economic Review 4 (2): 135-156. Appendix 1: Gravity model estimates Table 9 presents various estimates based on the gravity model of Behar & Manners (2010). To evaluate the change in intra- and extra-regional trade flows of SADC countries over time, the gravity model is embedded in a panel setting where the time periods are specified as 1991-95, 1996-2000 and 2001-05. Each period reflects the average of (strictly positive) annual values and reflects a value of zero if no annual flows are recorded that period. With this data, we can include dummies for the time period to depict evolution over time. We can also include dummies to indicate whether the exporter, importer or both are SADC countries. The results here are based on IMF DOTS export data. However, the specification is also estimated using data from UNComtrade (using a combination of measures reported by the exporter and importer). While there are major discrepancies in trade flows between the databases, the econometric results are generally consistent for both sets of data. The equation is estimated using OLS and zero values are ignored. Alternative gravity model estimates based on a cross-country database for 2007 & 08 and estimated using a Poisson Pseudo Maximum Likelihood (PPML) estimator proposed by Santos Silva and Tenreyro (2006) corroborate our findings regarding the regionalization of SADC trade. Table 9: Standard gravity estimates Dependent variable: ln (1) (2) (3) (4) exports ln GDP exporter 0.777*** 1.233*** 0.773*** 0.774*** ln GDP importer 0.951*** 0.915*** 0.920*** 0.917*** ln distance -1.480*** -1.774*** -1.771*** (SADC exporter) x (ln -0.433*** distance) 1996-2000 dummy 0.127*** -0.0708*** 0.0312 0.0318 2001-2005 dummy 0.185*** -0.199*** -0.00213 -0.00157 Share a border 1.068*** 0.729*** 0.739*** Formerly same country 1.258*** 1.192*** 1.177*** SADC exporter 0.124 SADC importer -0.0896 SADC pair 1.236*** 0.976*** SADC pair * 1996-2000 0.0126 SADC pair * 2001-2005 -0.0017 Constant -27.28*** -24.38*** -13.11*** -9.235** Observations 53929 53929 53929 53929 Importer, Importer, Fixed effects? Pair None exporter exporter Notes: * p<0.05 ** p <0.01 *** p<0.001 23 Appendix 2: Obstacles to SADC trade Figure 6: Sub-indicators of Enabling Trade Index against log GDP per capita Transport and communications infrastructure Market access 1 corresponds to worst outcome, 7 to best outcome 6 6 Madagascar Lesotho Malawi Mauritius 5 5 Mozambique Zambia Botswana Namibia Tanzania 4 4 South Africa Zimbabwe South Africa Mauritius Botswana Namibia 3 3 Zimbabwe MalawiZambia Madagascar Mozambique Lesotho Tanzania 2 2 4 6 8 10 12 4 6 8 10 12 ln(GDP per Capita), 2005 PPP ln(GDP per Capita), 2005 PPP Border administration Business environment 1 corresponds to worst outcome, 7 to best outcome 7 7 6 6 5 5 Mauritius Mauritius Botswana Malawi South Africa Namibia Botswana Zambia 4 Tanzania South Africa 4 Namibia Madagascar Madagascar Mozambique Tanzania Lesotho Mozambique Malawi Zambia 3 Lesotho Zimbabwe Zimbabwe 3 2 4 6 8 10 12 4 6 8 10 12 ln(GDP per Capita), 2005 PPP ln(GDP per Capita), 2005 PPP Source: World Economic Forum Figure 7: Trading on Time and Logistics Performance Indices against log GDP per capita Doing Business Trading on Time Index Logistics Performance Index 1 corresponds to worst outcome, 5 to best outcome 7 5 Mauritius 6 4 Madagascar Tanzania Seychelles Mozambique South Africa South Africa 5 Lesotho Namibia Swaziland Botswana 3 Congo MalawiZambia Congo Mauritius Madagascar Tanzania 4 Angola Botswana Mozambique Zambia Angola Zimbabwe Namibia 2 3 2 1 4 6 8 10 12 4 6 8 10 12 ln(GDP per Capita), 2005 PPP ln(GDP per Capita), 2005 PPP Note: The World Bank Doing Business (Trading on Time) indices are constructed as follows: The underlying variables (export and import costs, time and documents) are first standardized (mean 1 and variance 1). The simple average of these is then converted to a 1 to 7 scale (1 corresponds to worst outcome, 7 to best outcome). The underlying trading times, documents and costs are the most recent values available from the World Bank Doing Business Indicators (accessed 1 June 2010) and the Logistic Performance Index (2009). 24 Table 10: Conditional estimates of trade and infrastructure constraints to trade in SADC countries World Bank Doing Business indicators (Trading on Time data) (1 corresponds World Economic Forum Enabling Trade Indices (1 to worst outcome, 7 to best outcome) corresponds to worst outcome, 7 to best outcome) Transport & WB Trading WB Trading communicati WB Trading on Time on Time WEF Border ons Business on Time index, index, Enabling Market administr infrastructur environm index imports exports trade index access ation e ent ln(GDP/Capita) -0.073 -0.004 -0.143 -1.336** 0.631 -1.607** -1.706** -2.651*** ln(GDP/Capita)^2 0.023 0.019 0.027 0.103*** -0.034 0.131** 0.140*** 0.176*** ln(internal distance) -0.250*** -0.228*** -0.271*** -0.109** -0.026 -0.199** -0.139** -0.076 ln(population) 0.110** 0.098** 0.122*** 0.039 -0.079 0.116** 0.151*** -0.031 Landlocked -0.952*** -1.018*** -0.886*** -0.079 0.179 -0.327** 0.02 -0.179 Angola -1.477*** -1.185*** -1.769*** Botswana 0.043 -0.209 0.295 0.220** 0.279 0.313 -0.270** 0.555*** Congo D.R -0.458* -0.512 -0.404 Lesotho 0.719*** 0.659*** 0.779*** 0.202** 0.980*** 0.144 -0.089 -0.224 Madagascar 0.634*** 0.213* 1.055*** 0.477*** 1.627*** 0.355** -0.069 0.007 Malawi 0.084 0.196 -0.028 0.510*** 0.992*** 0.394** -0.043 0.678** Mauritius 0.079 0.075 0.083 0.278** 0.655** 0.267* -0.18 0.365** Mozambique 0.386** 0.152 0.621*** 0.239* 1.086*** 0.353** -0.041 -0.433* Namibia -0.398** -0.071 -0.725*** 0.299** 0.401* 0.119 0.271** 0.428** Seychelles -0.531*** -0.394*** -0.668*** South Africa -0.666*** -0.710*** -0.622*** -0.036 -0.183* 0.123 -0.065 -0.039 Swaziland 0.069 0.029 0.109 Tanzania 0.557*** 0.371*** 0.743*** 0.206** 0.544*** 0.184* -0.317*** 0.402** Zambia 0.135 -0.126 0.396** 0.453*** 0.689*** 0.405*** 0.015 0.696*** Zimbabwe -0.191 -0.594 0.213 -0.439 -0.008 -0.215 -0.098 -1.436** N 187 187 187 135 135 135 135 135 R2 0.642 0.627 0.633 0.767 0.188 0.747 0.857 0.635 DSADC -0.085 -0.149 -0.022 0.214** 0.540** 0.224** -0.099 0.186 Notes: See notes to earlier tables for data sources and construction of variables. * p<.1; ** p<.05; *** p<.001 Table 11: Conditional estimates of Logistics Performance in SADC countries World Bank Logistics Performance Index (1 corresponds to worst outcome, 5 to best outcome) Logistics Infrastructu International quality and Tracking LPI Score Customs re Shipments competence and tracing Timeliness ln(GDP/Capita) PPP -1.133*** -1.801*** -1.674*** -0.537* -1.084** -1.414*** -0.491 ln(GDP/Capita)^2 0.086*** 0.125*** 0.124*** 0.045** 0.085*** 0.103*** 0.046** ln(internal distance) -0.095** -0.086** -0.094** -0.117** -0.102** -0.122** -0.047 ln(population) 0.138*** 0.108*** 0.167*** 0.122*** 0.161*** 0.168*** 0.112*** Landlocked 0.022 0.012 0.018 0.095 -0.003 -0.002 -0.002 Angola -0.326*** -0.473*** -0.562*** -0.230*** -0.441*** -0.052 -0.243*** Botswana -0.461*** -0.430*** -0.458*** -0.872*** -0.347*** -0.152* -0.423*** Congo D.R 0.084 0.078 -0.077 0.04 0.526** -0.33 0.206 Lesotho Madagascar 0.253*** 0.198*** 0.589*** 0.587*** 0.159** 0.039 -0.091 Malawi Mauritius -0.175*** 0.094 -0.333*** 0.319*** -0.350*** -0.368*** -0.532*** Mozambique -0.124* -0.244*** -0.022 0.311*** -0.039 -0.213* -0.567*** Namibia -0.406*** -0.456*** -0.371*** -0.263*** -0.238** -0.348*** -0.761*** Seychelles South Africa 0.445*** 0.574*** 0.597*** 0.321*** 0.633*** 0.656*** -0.064 Swaziland Tanzania 0.097** 0.217*** -0.157** 0.223*** 0.019 -0.015 0.234*** Zambia -0.093** 0.077 -0.158** -0.114 -0.171** -0.037 -0.124 Zimbabwe N 161 161 161 161 161 161 161 R2 0.825 0.774 0.848 0.615 0.796 0.741 0.68 DSADC -0.017 0.04 -0.045 0.067 0.023 -0.005 -0.196** Notes: The Logistics Performance Indicator data are obtained from Logistics performance survey data (2009). * p<.1; ** p<.05; *** p<.001 25